#qi-hardware IRC log for Thursday, 2011-06-09

rjeffriesok found the tablet http://www.radioshack.com/product/index.jsp?productId=1152879100:00
rjeffriesVelocity® micro Cruz T301 7" 2GB Tablet $199 USD00:01
rjeffriesactually $160 on Amazon00:02
wpwrakpeople seem to be happy with it. from the specs, i can't tell which CPU it uses. even the 4720 could drive an 800x600 screen.00:03
wpwrakhmm, i wonder how many people on the list find jane's mails readable ... at least with a plain text MUA, it's all just alphabet soup00:06
rjeffriesI get Jane's email via the qi-hardware mailer and can read them ok00:10
wpwrakrjeffries: maybe you're reading them as HTML ?00:11
rjeffriesmaybe so. but qi-hardware mail is not "pretty" my mail client is Google mail 00:12
DocScrutinizer51I usually bounce html-only mail00:13
wpwrakrjeffries: so you're using a web browser. yes, that would be HTML :)00:13
rjeffriesthe Ben sales of qrt 1,200 Wolfgang mentioned earlier implies that Ben sales run rate now is very low00:14
DocScrutinizer51there've been times where I considered bouncing multipart/alternative00:15
rjeffriesthe Pipermail system would strip out HTML me thinks00:15
wpwrakDocScrutinizer: ideally, you'd just bounce all mails. that way, you're safe :)00:15
DocScrutinizer51probably the best00:16
wpwrakrjeffries: yes, sales are quite slow00:16
rjeffrieslow and slow00:16
DocScrutinizer51I wonder who's cooking for those guys00:17
wpwrakDocScrutinizer: i've heard wolfgang's rich wife/girlfriend feeds him :)00:18
DocScrutinizer51nobody's cooking for me, and my hobby is way less expensive00:18
rjeffriesif you start with a working Android tablet (has linux underneath) what would be the path to getting OpenWet or Jlime limping along?00:19
wpwrakDocScrutinizer: see. fancy hobbies give you an air of greatness. this attracts females. as an act of social bonding, they cook for you. that's how it works :)00:19
rjeffriess/OpenWet/OpenWrt  lol00:20
rjeffrieshow is that working for you, wpwrak? ;)00:20
DocScrutinizer51not applicable here, I'd not even want that00:20
wpwrakrjeffries: step 1: find the sources. check if they're complete. see if there are any binary-only drivers in there. if yes, figure out how to replace them or how to live without them.00:20
DocScrutinizer51my GF ha to be as geeky as I'm00:21
wpwrakrjeffries: still working on getting the qi-hw news extolling my exploits to places the local females visit. maybe i should just print the community news and hang a banner from the building. i have the faculty of psychology right across the street, so there would be no shortage of test subjects :)00:24
DocScrutinizer51I definitely need to apply for the mcgyver for mars mission 100:27
DocScrutinizer51but then, the RTT :-/00:29
wpwrakdarn, i was hoping we could pack some bankers and patent mongers into that one-way mission to mars00:30
DocScrutinizer51might correct some idiosyncrasies I adopted thru the years00:30
wpwrakDocScrutinizer: planning to work on your social skills on that flight ? :)00:39
rjeffrieswpwrak should appreciate this vga article 01:51
rjeffrieswpwrak interesting factoids re panelizing pcbs02:01
wpwrak(vga) ah, interesting. the concept is old news, but i didn' know the vga input impedance. that's useful, thanks02:06
qi-botThe build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.trunk-full_system-06072011-2251/02:16
wpwrakkristianpaul: ey, i need you for another feedback from south america. jane's mails - what do they look like for you ? (and what MUA do you use)04:13
kristianpauli use mutt04:15
kristianpaulwpwrak: jane's mail about usb host or cpu ?04:16
wpwrakkristianpaul: the formatting. do they look readable to you ?04:17
kristianpaulhorrible no :(04:17
wpwrak(in general)04:17
wpwrakhehe ;-)04:17
kristianpaulthat make me hard to read the thread actually..04:17
kristianpaulwich is interesting, but i got confused reading it..04:18
DocScrutinizer\o/ GPS partially going down possibly05:33
kyakhail to GLONASS? :)05:37
kristianpaultoo soon05:45
Action: kristianpaul hides05:45
rohhihi... looks like a moko http://cgi.ebay.de/29057002407506:30
zedstarben in schools in nepal http://twitpic.com/58y34607:25
wolfspraulahh, Nepal! almost forgot about it07:28
wolfspraulthe trip was in May, right? How did it go?07:28
wolfspraulzedstar: nice picture! Is it cc licensed so I can upload on qi wiki and mention in the news?07:32
wolfsprauldo you have a whole blog post about your trip, or planning to write one?07:32
tuxbrainzedstar: retwitted :)08:04
wolfspraultuxbrain: hey you actively use Twitter for communicating with your customers and fans now?08:06
wolfspraulyou are so modern :-) It's only 2011... I'm still not using it! (nothing to be proud of, I know...)08:06
wolfspraultwitter.com/tuxbrain - indeed, there you are08:08
wolfspraul98 followers :-)08:08
wolfspraulgood luck, start a revolution maybe! are those people in Madrid still camping out?08:08
wolfspraulyou can start a revolution in spain, and become the tux spokesman08:08
tuxbrainI have 1500 followers on identica, a hundred aprox in twitter as tuxbrain, and personal twitter as dsamblas that has started a week ago to not mix political thoughs with tuxbrain we have to take advantage of ti08:09
tuxbrainof it08:09
wolfspraulyou start the revolution as dsamblas?08:09
wolfspraulhow well do those twitter & identica accounts serve you in communicating with people?08:10
wolfspraulcan you drive some sales through it? get good feedback? keep people informed about what you do?08:10
wolfspraulor you just talk to yourself mostly?08:10
tuxbrainwolfspraul: I have been using it actively for a week now08:10
tuxbrainmaybe two08:10
wolfspraulok too early to tell. keep me posted08:10
wolfspraulhey there is really revolution stuff under twitter.com/dsamblas!08:11
wolfspraulI wasn't too far from the truth :-)08:12
tuxbrainheheeh :)08:12
wolfspraulwhen do I see you camping out at some plaza in Barcelona?08:12
wolfspraulon TV...08:12
wolfspraulyour daughter can bring the bread, should make for some good pics08:12
tuxbrainI have a little camp in my little town just in front of my house08:12
tuxbrainyes I think I appear in some pics of his facebook, as far as I can I try to help them08:13
wolfspraulyou should come to China, with your beard and all we can sell you as "Spanish revolutionaire" easily08:13
wolfspraulgive some speeches etc.08:13
wolfspraulonly we need to make sure you leave the country fast enough before anyone gets worried08:14
wolfspraulnot that you suddenly have 20 million Chinese followers...08:14
tuxbrainheheeh how confortable are chinese prision beds? 08:14
wolfspraulor 20008:14
wolfspraulnah is OK first, you can come as a communist revolutionair08:14
wolfspraullook a bit like Marx08:14
wolfspraulyou can start your speeches at party schools08:14
wolfspraul"the famous revolutionaire from Spain"08:14
wolfspraulthen you can move to private rented locations, finally stadiums08:15
zedstarwolfspraul: hi.....sujan is there now with the devices....spending quite a bit of time in the schools....is going well08:15
wolfspraulyou wouldn't need to worry because you come in the footsteps of Marx, from Europe etc. with bears and all - all authentic08:15
zedstarwolfspraul: they were able to buy cheaper speakers there to connect for audio/video english language learning so working nicely in class room08:15
wolfspraulbring some poster-size pictures with you and some comrades at your camp, that should be enough to get started08:16
wolfspraulzedstar: wow08:16
wolfsprauldo you plan to blog about it?08:17
wolfspraulzedstar: is that picture cc-by/public domain?08:17
wolfspraultuxbrain: well that is really cool stuff with your camp, let's see what happens next.08:17
zedstarwolfspraul: yes to both....i guess when sujan comes back he will do proper blog posts etc08:17
wolfspraulok great, thanks [cc]08:17
zedstarwolfspraul: i was there also at a conference on computing for developing countries and demo'ed the ben and concepts behind it...went down very well...most had no idea of concept of open hardware08:19
tuxbrainzedstar: great :) that's we need difussion of the concept great work really08:19
zedstarthese guys have no computing in classrooms and no internet....problems with electricity.....so they love having content on these kind of devices offline08:21
tuxbrainwolfspraul: well the real revolutionare is my wife :) but he is too lazy to use social networks so I'm some kind of his community manager :P08:21
tuxbrainhe -> she08:21
wolfsprauleven better!08:23
wolfspraulseriously we should get both of you to China to preach the revolution08:23
wolfspraulyou would have millions of fans quick08:23
wolfspraulcan sell them Nanos later :-)08:23
wolfspraulzedstar: I would love to find out which exact applications they use and how, so we can optimize those apps first08:24
wolfspraulI am most worried that we have a lot of apps only in 'nice to look at' state, but when you use them in reality there are too many details in your face to make it a fun experience08:24
zedstari like the idea that these people can feedback into the design or a next device or even make it themselves rather than look to exploit them with version 2 high-end tech08:25
zedstarvideo playing works nicely....even without speakers is decent sound08:25
zedstarbeing able to project content would be a killer feature for this environment08:32
wolfspraulzedstar: they have projectors? what kind of input can the projector handle?08:34
wolfspraulwhich one is better - vga or tv-out?08:34
tuxbrainzedstar: UBB VGA maybe?08:34
wolfspraulzedstar: yeah in case you missed the news when you were traveling... http://en.qi-hardware.com/wiki/Community_news_2011-06-0108:35
zedstarthey dont have projectors so would need to work out the best way....getting hold of used monitors might work if could do vga08:36
zedstartuxbrain yeh that wud be cool i thnk08:38
kyaksometimes it's useful that your computer kicks your ass. After soldering of some broken capacitor i re-applied thermal grease and installed additional cooler (things i wanted to do long ago, but didn't have enough motivation). Now the CPU temperature is around 67°C at LA 0.9 - the kind of temperature used to be during idle.09:02
kyaki hope it won't bring me down again.. I read that the defect rate of Shuttle K45(SE) can reach 40-50%09:03
kyaksome topical web sites are even selling a "Shuttle K45 Capacitor Kit"09:04
wpwrakwolfspraul: (revolution) clandestine communication with fellow conspirators with WPAN ... ;-)10:37
wpwrakkyak: (capacitor) apparently, mainboard makers are now reducing the number of capacitors, cutting into safety margins, because prices have risen due to china's rare earth metal export limitation. so expect new pcs tp be less reliable than old ones.10:44
tuxbrainwpwrak: I must go but i want to let you know fab units are already here :), I have flashed one usb the spectrum seems to work ok with your fix but I try to run make usb and it fails10:45
tuxbrainI think is our mosi friend again10:45
tuxbrainI have to go now more later 10:46
tuxbrainthey look soo cooool10:46
wpwraktuxbrain: wheee !! can you please take a picture of the top of the board ? that way, i can check if there's anything obviously wrong10:47
wpwraktuxbrain: (mosi) hmm, if the fab boards do that too, maybe it's something else then10:48
wpwraktuxbrain: thanks !11:01
wpwraklooks good11:02
wpwraktuxbrain: next step: see what "make usb" wrote in _log11:03
kristianpaul(clandestine wpan) yeah :-)11:13
kyakwpwrak: good thing i'm not planning to buy some new hardware in the new time...11:18
kyak*near time11:19
wolfspraulwpwrak: oh my god. remote visual production testing. that will be fun :-)11:48
wolfspraulremote AOI, semi-automated through Werner11:49
wolfspraulI cross my fingers that this will result in working boards. I finally can buy something from Tuxbrain! :-)11:52
wpwrakwolfspraul: (testing) erm, we have something a little better ;-) http://downloads.qi-hardware.com/people/werner/wpan/prod/index.html12:17
wpwrakwolfspraul: the visual part is just for first impressions and to figure out any problems the automated tests discover12:18
wolfspraulyes I know that we have that, but that's just a theory now. Is it used in the real run/real boards?12:18
wolfspraulI'm not worried about the quality of the tools you provide, I'm worried about how the tools are being applied :-)12:19
wolfspraulI have seen many a bad example in this area in China, ahem.12:19
wolfspraulno tool can be great enough to overcome whatever a human in front of the tool is able to do to it :-)12:19
wpwraktuxbrain: you need to update your ben-wpan repository. the test script you've been running is old. (doesn't have the LED test)12:20
wpwrakwolfspraul: i've used the test process on my prototypes. now it's being applied to the production boards. nothing unusual there :)12:21
wpwrakwolfspraul: the tools just run and do their thing. very little operator input ;-)12:21
wpwrakwolfspraul: there are only two tests where operator assessment is required: the LED test and the spectrum scan. in the latter, the tool determines whether the result is within limits or not, but the operator has to decide whether the configuration was valid. so the operator has to confirm the program's result. (but can't override it ;-)12:24
wpwrak(override) well, except by aborting the test. but you can't turn a fail into a pass. unlike a certain gta02 NOR test :)12:25
kyakwith such firm approach to QA, i have no doubt in atben/atusb i'm going to order. This is a great job!12:28
wpwrakkyak: heh, thanks :)12:32
kyakwpwrak: this might be an unappropriate question, but are you going to be paid for the job you are doing on atben/atusb, or is it all pure enthusiasm?12:40
wpwrakkyak: sadly, it's all pure enthusiasm. qi-hw is a rather penniless venture ...12:41
kyakindeed, it's pretty sad12:41
larscwhere are the multi-million-dollar investors when you need them?12:43
wpwraklarsc: yeah, i keep on asking wolfgang where he's hiding them, but i think he just doesn't want to share12:44
kyakdo you guys play golf? :_12:46
wpwrakkyak: i guess wii golf doesn'12:48
wpwrakt count ? :)12:48
dvdkmay i ask a few stupid git questions?12:48
jluisperl golf?12:49
qnrqdvdk: shoot!12:49
kyakwpwrak: it counts if you have some multi-millionaire joined in :)12:49
dvdkpreviously did a commit to the 'octave' package put not pushed it yet.  is it possible to move this commit and all the other local changes to 'octave' over to a local branch, so i can ignore octave for now, push all the other changes and take up octave, once i have more time for it?12:49
qnrqdvdk: you could cherry-pick them over to a new branch and remove them from your working branch12:50
dvdki.e. all my current octave changes break the build so i don't want to push them just yet, but continue with other work  still12:50
qnrqthen whenever you want to push your stuff you just merge your 2nd branch with the working one12:51
dvdkqnrq: sounds reasonable.  what again is the command to list all unpushed commits?12:51
dvdkgit log?12:51
qnrqgit log lists all commits, not only your unpushed12:51
qnrqdvdk: gitx gives a pretty nice overview, it's a gui tool :)12:51
larscgit log remote-head~your-head12:52
dvdkgitx: command not found12:52
dvdkyou mean gitk? gitg?12:52
qnrqdvdk: no, I meant gitx :) http://gitx.frim.nl/12:52
dvdklooks like it's not packaged for ubuntu yet12:53
qnrqyou could just compile it :)12:53
larscuhm, i meant12:53
larscgit log remote-head..your-head12:53
larscthat will basically list all commits not pushed12:54
dvdkgit branch octave12:55
dvdkgit checkout octave12:55
dvdkso now i'm on the new branch, I gues12:55
xiangfu`dvdk: another ways :), just checkout a new branch for same all your commits.(git co -b new-octave)12:55
xiangfu`dvdk: then checkout to the old branch git checkout old-octave. then (git rebase -i origin/master or origin/trunk)12:55
xiangfu`dvdk: just remove the line that you don't want keep , save and exit :D12:56
dvdkbut the new branch already contains all my committed changes.  how do i remove those changes (currently a single commit) from the 'trunk' branch12:56
dvdkxiangfu: rofl12:56
larscrebase -i12:56
xiangfu`git rebase -i origin/trunk12:56
xiangfu`yes. rebase -i12:56
kyakyeah, again, there are so many way to do the same thing in git..12:56
xiangfu`kyak: :)12:57
dvdkcool, with 'git checkout trunk' most of my commits are now gone12:58
wpwrakgit rebase -i HEAD~5  (or however far you want to go back) is such a useful tool. also note how it gets the interation with the user right ;-)12:59
dvdkso with 'git rebase -i' i  now just remove the 'pick' line for the unwanted commit and aro done?13:00
dvdkcool, seems to have worked13:00
xiangfu`GitX is a git GUI made for Mac OS X13:02
xiangfu`seems not for Linux.13:02
xiangfu`it should all iGit :D, then people will just understand it's for Mac OS 13:03
larscigit igit ;)13:03
wpwraklarsc: sounds appropriate, for mac ;-)13:05
larscwpwrak: exactly my thoughts13:05
kyakphew, i thought it would be a german for "iPoop" :)13:10
dvdkouch, my dns server (dnscache.berlin.ccc.de) is down, so i can't access qi-hardware git.13:15
dvdkcan anybody tell me an alternative open dns server?13:15
dvdkit's a little difficult to google, without having google's ip addres13:16
wpwrakdvdk: is your friend13:16
dvdkwho provides this service?13:16
wpwrakdvdk: google :)13:16
dvdkgoogle :)13:16
dvdkso they get even more of my data :/13:16
dvdkyeah internernet, i'm back13:17
wpwrakdvdk: yeah. and in return, they give you clean DNS responses. no redirects, no wildcards for typos, and all those other lovely things other do13:17
dvdkwpwrak: that's why i used ccc's dns server (but it's slow).  also most german universities seem to have dropped public dns :(13:17
tuxbrainwpwrak: I have pulled the git and ... it pass :)13:34
wpwrakwheee !! :)13:35
tuxbrainI have detected two strange things, a)first is our friend mosi, I have to unplug/plug the atusb to make the gpio test pass, making the test running twice without unplug the atusb is a sure fail on gpio13:39
tuxbrainb) the cristal frecuency somtime pass sometimes not13:39
wpwraktuxbrain: hmm. a) sounds like a reset issue. lemme check that ...13:40
wpwrakb) what does  ntpdc -c kerninfo  say ?13:41
wpwrakalso, b) "sometimes" with the same board or with some boards yet, with others no ?13:44
tuxbrainnow I have to really setup an "allways the same position" and of cours flash the the rest13:46
wpwrakyes, always the same position is important if you're going to do hundreds of devices ;-))13:46
tuxbrainmutrox@mutrox-ekija:~$ ntpdc -c kerninfo13:47
tuxbrainpll offset:           0.007595 s13:47
tuxbrainpll frequency:        5.720 ppm13:47
tuxbrainmaximum error:        0.51234 s13:47
wpwrakhave you tested an atben yet ?13:47
tuxbrainestimated error:      0.024781 s13:47
tuxbrainstatus:               2001  pll nano13:47
tuxbrainpll time constant:    1013:47
tuxbrainprecision:            1e-09 s13:47
tuxbrainfrequency tolerance:  500 ppm13:47
tuxbrainis with the same board , is the only one I have flashed yet13:47
tuxbrainatben(not yet)13:47
wpwrakNTP looks good. crystal is fairly accurate and the NTP daemon is happy, too13:48
tuxbrainso what can be the reason of the cristal fail?13:48
wpwrakcould be that your system is generally more "noisy" than mine13:49
wpwrakwhat happens if you   for n in 0 1 2 3 4 5 6 7 8 9; do atrf-xtal  10000; done  ?13:49
wpwrak(will take ~90 seconds)13:50
tuxbrainsyntax error near unexpected token `do13:50
wpwrakuh ? :)13:51
tuxbrainsorry bad c&p, doing it now13:51
wpwrak"or n in ..." ? ;-)13:51
tuxbrain+0.24 ppm13:52
tuxbrain-4.55 ppm13:52
tuxbrain-5.10 ppm13:52
tuxbrain+5.55 ppm13:52
tuxbrain+11.25 ppm13:52
tuxbrain+6.75 ppm13:52
tuxbrain-7.64 ppm13:52
tuxbrain-3.90 ppm13:52
tuxbrain+4.78 ppm13:52
tuxbrain-4.41 ppm13:52
wpwrakwow. all over the place13:52
tuxbrainthat's good or bad? 13:53
wpwrakis your machine doing something heavy in the background ? kernel compilations, video playback, ... ?13:53
wpwrakit's bad but it may not be the atusb's fault. you may want to compare with my prototype13:54
tuxbrainwell is a limited machine, is an arm(I need that to be able to perform make a mobile test platforms) mmmm let me check with my desktop computer13:55
tuxbrain+4.81 ppm13:58
tuxbrain+4.89 ppm13:58
tuxbrain+4.85 ppm13:58
tuxbrain+4.88 ppm13:58
tuxbrain+4.87 ppm13:58
tuxbrain+4.91 ppm13:58
tuxbrain+4.93 ppm13:58
tuxbrain+4.98 ppm13:58
tuxbrain+4.94 ppm13:58
tuxbrain+5.01 ppm13:58
tuxbrainyes I don't know what that results means but are quite different from on machine to another13:58
wpwrakthat looks a lot more like what i'm used to ;-)13:58
dvdkwpwrak: tuxbrain: you'r using a USRP to do these tests?13:59
wpwrakabout +/- 0.1 ppm variation is what i get on my desktop too. well, a bit less, but then it's (still) a relatively powerful machine13:59
tuxbrainno no external mesurement just wpwrak tools14:00
dvdkbecause if tuxbrain's ARM machine is an efika-mx i'd say it doesn't have the kind of usb controller that can do iso transfers at full usb2.0 rate.14:00
tuxbraindvdk: yes it is14:00
dvdkalso have one of these to play with at work.  but the usb controller (or kernel driver) is really much weaker than even an intel atom board.14:01
tuxbrainyes seems I have to setup my toshiba laptop to perform the test14:01
wpwrakdvdk: alas, tuxbrain lacks fancy tools like USRP and such. but i've written a set of test tools for atben/atusb that should make up for most of it.14:01
wpwrakdvdk: what we're seeing here is simply the kind of problems you'd expect when trying things for the first time on a system different from the author's14:02
dvdkwpwrak: and these tools only use an atusb for testing?  what is the atusb doing that maxes out a 800 MHz iMX CPU?14:02
dvdkdon't let yourself be disturbed by me.  everybody is waiting for atben/atusb :)14:02
dvdkyeah, finally octave compiles again.14:03
tuxbraindvdk: I plan to do a more "profesional" performace test in the near future in a lab with a lot of RF fancy tools, just need to find the time14:03
wpwrakdvdk: here's a description of the test process: http://downloads.qi-hardware.com/people/werner/wpan/prod/14:03
dvdkwpwrak: cool test diagramm.  i guess with that you could even train monkeys to do the test :)14:04
tuxbraindvdk: but for now just knowing than the atben/atusb than leave tuxbrain works is enough for me :)14:04
wpwrakdvdk: the atusb clock frequency test measures the atusb crystal clock in relation to the system clock of the host. it's not particularly taxing in terms of system load, but it is sensitive to system clock variation and scheduling delays. so if the system is very "noisy", the measurements will suffer.14:04
tuxbraindvdk: I'm the empiric demonstration of your afirmation14:04
wpwrakdvdk: (diagram) yeah, i should do diagrams for the sub-parts, too. didn't have enough monkeys for proper behavioral studies, though :)14:05
tuxbrainyes definively the cristal is fault of the efika, the test runs smoth on a Dual core quad :P14:10
tuxbrainlest the atben...14:10
wpwraktuxbrain: thinking of it, the timing variations could also come from a low HZ setting in the kernel. hmm, now, how to retrieve that one ...14:12
wpwraktuxbrain: the MOSI problem is gone for good now ?14:16
qi-bot[commit] David Kühling: alex4: move save-files to ~/.alex4.* http://qi-hw.com/p/openwrt-packages/7d8785d14:17
qi-bot[commit] David Kühling: octave: fix fortran compile flags, enable parallel compilation http://qi-hw.com/p/openwrt-packages/c33acf014:17
tuxbrainyes the mosi thing only happens if I run the test twice without unplug the atusb first14:18
wpwraktuxbrain: ah, but then it still happens all the time ?14:20
tuxbrainif I run the test twice yes14:20
tuxbrainbut also never if I unplug before to do the test14:20
wpwraktuxbrain: hmm. i still don't like it14:21
wpwraktuxbrain: after running the test, can you please do  atrf-gpio -p HHHHoHxH.HxHxHxHh.ohzoHoHH14:21
wpwraktuxbrain: that should produce the usual complaint14:21
wpwraktuxbrain: then, atrf-reset -a14:21
tuxbrainyest it does14:22
wpwraktuxbrain: and then again  atrf-gpio -p HHHHoHxH.HxHxHxHh.ohzoHoHH14:22
tuxbrainstill complains14:22
wpwrakthis means that a power-on reset resets something a software-controlled reset doesn't14:23
wpwrakthat's not supposed to happen14:23
tuxbrainif I unplug/plug no complains14:23
wpwrakthat's cheating ;-)14:23
wpwrakcan you reach the pins to measure the voltage on MOSI ?14:24
tuxbrainit's challenging with my instrumental but I will try14:24
wpwrakatrf-gpio -p HHHHoHxH.HxHxHxHh.ohzzHoHH delay=1000014:24
wpwrakthen you should have 10 seconds to measure the voltage14:25
wpwrakwait, let me turn on the LED, for feedback14:25
wpwrakatrf-gpio -p HHHHoH1H.HxHxHxHh.ohzzHoHH delay=1000014:25
wpwrakwhile the LED is on, the voltage on MOSI should be stable 14:26
tuxbrainit marks 0 in several trys14:29
wpwrakah, and what does  ldd `which atrf-gpio` | grep usb   say ?14:29
tuxbrainlibusb-0.1.so.4 => /lib/libusb-0.1.so.4 (0x00007f6e69467000)14:31
kristianpaultuxbrain: do you ship by fedex?14:32
kristianpaulcan you low the declaration values for atben/atusb combo?14:34
wpwraklibusb looks good. i have the same version (valgrind complains a little about it, but nothing that looks too serious)14:34
kristianpaullow i mean not ilgellay just to raw costs.. :-)14:34
wpwraktuxbrain: put a sticker "MADE IN CHINA" on is and divide the cost by 10 ;-)14:35
kristianpaulanyway i'll try convince other nanonote owner to get a pair :-)14:35
tuxbrainkristianpaul: let's talk this on private later, I must leave now, see you later :)14:35
kristianpaulsure np14:35
kristianpaulbye !14:35
kristianpaulwpwrak: did tuxbrain already pass functional test?14:36
wolfspraulkristianpaul: there is no 'illegal'14:36
kristianpauli wasnt able to follow all the backlog from you14:36
wolfspraulI'm not going to redefine the world here, but well, in business there are perfectly 'legitimate' (and legal) cases every day where valuations for the same thing differ dramatically14:38
kristianpaulyeah,surelly i'm misundertood that word for other reasons.. i think i got that word wrong  use now14:38
wolfspraultax value, customs value, insurance value, price you paid - can all be totally different :-)14:39
wolfspraulnah it's ok, I know most people would feel like that14:39
wpwrakkristianpaul: we're in the process of testing. seems that the atusb boards work so far (still don't know if the RF side is okay, though)14:39
wpwrakkristianpaul: there is one mystery glitch in the tests i don't understand yet either14:39
wolfspraulnumbers are twisted in all directions14:40
wolfspraulsometimes when the government doesn't like how much tax they can charge, they just slap a (higher of course) 'estimate' on you14:40
wolfspraulfully knowing that the estimate is higher than the actual cost, but so what. if it's the law it's the law.14:40
wpwrakhmm, nokia's CTO threw the towel. they're going down faster than AF447 ...14:41
wolfspraulinsurance values are yet another completely separate number. you may under or over-insure for all sorts of reasons.14:41
wolfspraulso be a good happy international business man. pay some amoutn X, put a completely different amoutn Y on the invoice, use whatever number suits your purposes :-)14:42
wolfspraulthe other participants in the game will do the same thing, no worries :-)14:42
wpwrakwolfspraul: (estimate) heh, tell me. they quite generously upped the value of the collection of plastic parts you sent me a while ago. something like USD 50. (meaning ~USD 25 for customs)14:42
wolfspraulthat's what I'm saying14:43
wolfsprauleverybody uses the number that suits their needs14:43
wolfsprauland there are all sorts of reasons and justifications for that14:43
wpwrakof course, one could complain, but ...14:43
kristianpaulwolfspraul: having a fedex account may allow to reduce shipping costs some how?14:44
wolfspraula lot, fedex has big discounts for volume customers14:44
kristianpauljust this week fedex opened a office just under the office in wich i work, so i was thinking go to ask..14:44
wolfspraul30% or so, I would think it goes much higher if you send a lot14:45
wolfspraullike A LOT, like tens of thousands of packages per day14:45
wolfspraulyou can go there, but at your volume the discounts will be small14:45
wolfspraulI am sending a few hundred packges with them, I think I get 30%14:45
wolfspraulor maybe even 1000 now, somehting like that. over the course of a year.14:45
wolfspraul'small business customer' probably, to them14:45
wolfspraulif you are Amazon, you can probably negotiate over every penny14:46
wolfspraulmaybe if you register with them you can get 10% discount14:46
wolfspraulbecause you have an account and it will make communication, addressing etc. easier14:46
wolfspraulI think Adam gets 10% in Taipei after he did that14:47
kristianpauli'll try then14:47
wolfspraulyou should definitely try14:47
wolfspraulbut keep your expectations low :-) it will still be very expensive14:47
wolfspraulespecially packages originating from Buga/Colombia14:47
wolfspraultoo little volume there, not optimized14:47
wolfspraulthe computers cannot optimize for everything at once, so I can clearly tell you if you are swimming with a lot of other fish, it gets really cheap14:48
wpwrak"bugga ! where is that columbia ?"14:49
wolfspraulit's amazing what fleet of airplanes they are moving all the time, and how they can keep them full14:49
wolfspraulI've seen the DHL center in Hong Kong once - wow!14:49
wolfspraulI cannot being to imagine what the huge hubs in the US will look like14:50
wolfspraulfedex has 697 airplanes right now! (with 49 more on order)14:51
wpwraklargest airline in the world ?14:51
wolfspraullargest cargo fleet, according to Wikipedia14:56
wolfspraulthey must have weather-forecast like computers running this thing14:56
wolfspraulthat's why I say if you really send A LOT, by their standards, they can probably sit down with you and work out some amazing discounts. if they understand your traffic etc. well. But then you are talking about thousands of packages you send per day.14:57
wolfspraulI'm very happy with them.14:57
wpwrakyeah, me too. lowest rate of bad surprises when shipping to argentina.14:59
wpwraktuxbrain: ah, wait a minute .. that "stable voltage" test doesn't make sense. just MOSI is one of the lines that get changed before entering the delay15:23
wpwraktuxbrain: lemme rethink this ...15:23
Fusinhi qiots :)16:54
qi-bot[commit] Werner Almesberger: atusb/fw/: split board functions into app-only/shared part (boot overlflowed) http://qi-hw.com/p/ben-wpan/e67840117:07
qi-bot[commit] Werner Almesberger: atusb/fw/board_app.c: removed unnecessary includes http://qi-hw.com/p/ben-wpan/ed6d43b17:07
qi-bot[commit] Werner Almesberger: atusb/fw/boot.c (main): reset_rf implies spi_init - no need to call it here http://qi-hw.com/p/ben-wpan/814cf6f17:07
qi-bot[commit] Werner Almesberger: atusb/fw/: added "lazy enabling" of UART-SPI after ATUSB_GPIO http://qi-hw.com/p/ben-wpan/ec7c93717:07
whooohi, got my ben nanonote today, so I have two questions, first, the screen is flickering, is that just me or is it like that for everybody? and the second is question is regarding usb gadget support, can it to "anything" or is it hardcoded in the hardware?19:35
larscwhooo: the flickering should go away after an software update19:41
whooolarsc: ok, cool, thanks19:41
larscwhooo: you can use different gadgets19:42
larscthe nanonote could act as sound card/network card/game controler/usb drive...19:44
larsceven a printer ;)19:45
whooobesides making sure it's enabled in the kernel, how should I go around fiddling with that kind of thing?19:46
larscread the documentation Documentation/usb/gadget_*.txt in the kernel source dir19:47
larscnetwork and usb driver are straight forward19:48
larscbut if you want it to act as an HID device you need some extra platform code19:48
whooook, time for some fiddling then, thanks for the help :-)19:55
jluishttp://www.vanille-media.de/site/index.php/2011/06/09/the-eagle-has-landed/ mickey is online on om-cdevel21:00
rozzinlarsc: Using Ben as a USB modem is another nice idea... :)21:14
rozzinHm. Actually...,21:15
rozzinI bet that might actually be useful under Windows.21:16
rozzinHunh--it's even in the docs for USB_G_SERIAL: "This driver supports a CDC-ACM module option, which can be used to interoperate with MS-Windows hosts[...]"21:21
whitequarkwpwrak: hi21:32
whitequarkwpwrak: what do you think about this module: http://iteadstudio.com/store/index.php?main_page=product_info&cPath=7&products_id=53 ?21:33
wpwrakwhitequark: nice price :-) but doesn't this thing use a proprietary RF protocol ?21:39
whitequarkwpwrak: well, it does. but you know, the price...21:40
whitequarkalso, it looks like I can turn off that protocol, by the cost of much less effective power usage21:40
whitequarkas I don't need to use battery power anywhere, that's not a problem21:40
wpwrakso you're saying it implements a standard protocol in addition to the vendor-specific protocol ?21:42
whitequarkwpwrak: the DS says "if Enhanced ShockBurst(tm) is enabled [...]", which suggests that it can be disabled21:43
whitequarkI'm not quite familiar with chip yet, but looks like you can just turn off ESB (it's the link layer) and just use plain GFSK21:44
whitequarkmaybe it will even be compatible with other vendors (TI?)21:44
wpwrakGFSK at 2.4 GHz is in vendor-specific land ... compatibility is more coincidence than by design :)21:48
wpwrakbut yes, if you don't mind vendor lock-in, the price is certainly good21:48
whitequarkwpwrak: they're all 3.3V and are controlled by SPI. I can always dump one module, use another one and make just software changes otherwise21:50
whitequark... replacing all the modules in the process, of course21:51
whitequarkwpwrak: looking at the registers, there is no way to switch protocols apparently. anyway, I think this is perfectly good for learning, as I don't want to produce something in mass quantities anyway21:54
wpwrakif your volume is low, then the main factor would be engineering cost :)21:55
tuxbrainwpwrak: the firmware you have send me is the one I should put in production?21:55
tuxbrainor just for the testing phase?21:55
tuxbrainany make magic order to just upload the firmware instead of "make flash"21:56
wpwraktuxbrain: this firmware shuold work for all uses (in general, atusb-*.bin has code for regular use and for testing. so far, i haven't made any versions that would not be backward-compatible)21:58
wpwraktuxbrain: "make update" aka atrf-reset -a; sleep 1; dfu-util -d 20b7:1540 -D atusb.bin21:59
tuxbrainok just want to check I don't deliver anything wrong21:59
tuxbraingoing to flash and measure21:59
wpwraktuxbrain: if you want, i can try to put in proper DFU mode switching tonight (it's been on my to do list for a while anyway). that way, you wouldn't need the atrf-reset and the sleep.22:00
wpwraktuxbrain: the most important thing is the boot loader :) as long as that one is good, you can always update the rest22:01
wpwraktuxbrain: ah, the boot loader you flashed, did you build it from sources or did you use my binary ?22:01
whitequarkwpwrak: just curious, don't USB and RF parts interfere with each other on atusb?22:02
wpwrakwhitequark: i hope not to any major degree :)22:02
tuxbrainwpwrak: I have used your binary22:12
tuxbrainwpwrak: mosi stays at 0V22:18
wpwraktuxbrain: *hmm*22:20
wpwraktuxbrain: and  atrf-gpio -p HHHHoHxH.HxHxHxHh.ohzoHoHH   still fails (right after it)22:21
tuxbrainnop no fail22:23
qi-botThe build has FAILED, see log here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-06092011-0416/22:23
tuxbrainroot@Iluvatar:~/Proyectos/tuxbrain/productos/nanonote/8_10/wpan/ben-wpan/prod# atrf-gpio -p HHHHoH1H.HxHxHxHh.ohzzHoHH delay=1000022:23
tuxbrainroot@Iluvatar:~/Proyectos/tuxbrain/productos/nanonote/8_10/wpan/ben-wpan/prod# atrf-gpio -p HHHHoHxH.HxHxHxHh.ohzoHoHH22:23
wpwrakah. let's put it into failure more first then. "make usb", see if it passes22:23
tuxbrainyes gpio test fails22:24
wpwrakwhat does the _log say ?22:24
tuxbrainhey now is another one not mosi!!! 22:24
tuxbrainPD1/nSS         Z   1   0***22:24
qi-botThe build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-06092011-0416/22:26
wpwraktuxbrain: maybe the time i give those signals to settle is just too short. can you please try this one ? http://downloads.qi-hardware.com/people/werner/tmp/atusb-10ms.bin22:28
tuxbrainsame gpio fail , on consecutive make usb22:30
tuxbrainPD1/nSS         Z   1   0***22:30
wpwrakgetting weirder and weirder :-(22:30
wpwrakthat's the prototype or the production board ?22:31
tuxbraindo you want to try with the prototipe?22:31
wpwraknaw. let's keep the production board for now22:31
wpwrakwhat happens if you   atrf-gpio 02:8/0x1f reset 02:0/0x1f22:32
wpwraknot even one small little error ?22:33
wpwrak"register 0x02: got 0xe8 expected 0x00/0x1f" ? no ? please ...22:33
wpwrakwell, let's try something else then. http://downloads.qi-hardware.com/people/werner/tmp/atusb-10us.bin22:34
wpwrakthis one returns to the previous timing for letting the signals settle but increases the reset pulse width.22:35
wpwrakall the MOSI, nSS, etc. problems would be consistent with the transceiver not resetting. but then, the reset test above should have failed, too22:35
tuxbrainatrf-gpio 02:8/0x1f reset 02:0/0x1f still giving no error22:36
wpwrakthat's okay22:36
wpwrak"make usb"22:36
tuxbraingpio error 22:36
wpwrakwhat's in the _log ?22:37
tuxbrainnSS again22:37
wpwrakoh wait. i think i copied the wrong pattern22:38
wpwrakchecking ... no, looks right22:39
tuxbrainanother question , I know I have not set up yet a correct spectrum enviroment defaults due I have not yet placed the things correctly , but instead of a green dot on the corner an red arrow or a yellow arrow appears and I can pres P to pass... this is normal?22:40
tuxbrainwhen doing make usb or make ben22:40
wpwrakdoes this complain ? atrf-gpio -p 02:0/0x1f22:40
wpwrakyellow/red means that the signal is outside the profile. that's okay, as long as you don't have a profile valid for your configuration22:41
tuxbrainok, also is normal than atben give lower values than atusb?22:42
wpwrak(atusb) how about this one ? atrf-gpio -p HHHHoHxH.HxHxHxH0.ohzoHoHH22:43
tuxbrainnothing happens22:43
wpwrak(atusb vs. atben) how much lower ? :) also, did you try to vary the position/orientation ? the antenna has a few "blind spots", where the signal gets significantly weaker22:44
wpwrak(nothing happens) that's good ! :)22:44
wpwraknow if it just did all that a little more consistently :)22:44
tuxbrainwpwrak: I know now how feels to be the IA of an conversational game22:45
wpwrakso you'll be in good shape when they come and take you to be turing-tested22:45
wpwraki suspect that all these errors we're seeing may just be bugs in the test system. can't quite imagine which part of it, though22:47
wpwrakthere's also the issue that some of the behaviour i test is under-specified. i.e., the transceiver uses pull-up and pull-down resistors, but they don't specify their value. so at least in theory, they could simply be too weak, and thus yield unpredictable results.22:49
tuxbrain(atusb vs. atben) now they are in a total random position in side a wifi nightmare, with a lot of obstacles and twisted cables everywhere and me in the middle so definitively not a sane rf enviroment22:49
wpwrak(rf jungle) okay, that would explain pretty much any result ;-)22:49
wpwrakmaybe give the atbens a try now. see how far the test goes.22:51
wpwrakif you get stuck at the spectrum test, maybe just comment it out in prod/atben (at the bottom of the file)22:52
tuxbrainatben goes ok, just the spectrum part fails 22:52
tuxbrainthere is more test after spectrum?22:52
tuxbrainok the transmit22:53
tuxbrainatben all ok , at leat on the test :)22:54
tuxbrainwhet it sais send /recive is the atusb sending/reciving in the other side or is just some kind internal test about capability to send and recive?22:55
wpwrakthat's a real transmission between atben and atusb. so it's good if this works ;-)22:59
wpwrakbut, thinking of it, i could make this test a little more interesting ...22:59
wpwraki have to be afk for ~30 min23:00
tuxbrainok I will wait for you if I don't felt asleep first23:00
wpwrakyou can de-panelize boards while waiting ;-)23:01
wpwrakand clean up the rf jungle :)23:02
wpwrakah, and do the sugru trial !23:02
tuxbrainhehehe this is for the weekend dude, I have two meetings to prepare tomorrow and preparing my speach for the capitalist sharks23:03
tuxbrain(clean up the rf jungle) I will move during the weekend to a place where even GSM is scare, more easy than clean my sweet messy cave23:06
tuxbrain(sugru trial) first I have to have a sane rf setup to decide if sugru is what is afecting the values ...23:08
wpwraktuxbrain: (moving to the desert) so no internet and hard to repeat the tests ? not ideal :)23:28
tuxbrainwell is a little bit exagerated , there is enough movile signal to have a decent irc conversation and to read manuals online :P23:29
tuxbrainI go every weekend so the test is repetable 23:29
wpwrak(irc) okay, sounds tolerable :)23:29
wpwrak(every weekend) so let's hope you can answer all questions this weekend, or else ben-wpan will get delayed N weeks23:30
tuxbrain:P, well seems that at least they work, send/recieve :P enough for sale, now is matter of improvements for the next batch _P23:36
wpwraki'll improve the send/receive test a little23:36
wpwrakright now, it could pass things the kernel wouldn't like23:36
wpwrakthose things should get detected by other tests, but better safe than sorry23:37
tuxbrainthis test will be ready son? (I mean in a couple of minuts) or I better go to sleep?23:40
wpwrakshouldn't take long ... i'm on it23:41
wpwrakalmost there ...23:57
--- Fri Jun 10 201100:00

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