#qi-hardware IRC log for Monday, 2011-07-18

errordeveloper__I just found the WPAN project00:02
errordeveloper__excellent! I was considering this idea at some point ..though haven't got all the PCB skills and time for working on this00:03
errordeveloper__that's a bit on my wiki which covers that :)~00:04
mthwpwrak_ is the one who designed the WPAN00:04
mththe channel is rather quiet today though00:04
errordeveloper__any how ... I'm glad to find this channel so I'll stick around :)00:04
mthanyway, I'm going to bed now00:04
errordeveloper__mth: me to ..00:05
errordeveloper__it's 1am here00:05
mth2 am here00:05
kristianpaulhi and wellcome, errordeveloper__ :)00:17
GitHub36[milkymist] kristianpaul pushed 9 new commits to gps-sdr-testing: http://bit.ly/qiBUFz00:27
GitHub36[milkymist/gps-sdr-testing] top code clean up plus pin definition renamed to real-mode - Cristian Paul Peñaranda Rojas00:27
GitHub36[milkymist/gps-sdr-testing] added module for flag crossing - Cristian Paul Peñaranda Rojas00:27
GitHub36[milkymist/gps-sdr-testing] added test bench, WIP - Cristian Paul Peñaranda Rojas00:27
kristianpaulall good so far,  i can read teh hw id from the namuru, afaik... seems endianess aligment pop up.. but i was already fixed in hw i tought.00:55
kristianpaullet see00:55
kristianpaulafaik i cant run rtems... i guess is something in my soc, but so far i'll stick on mm1 bios hacking :-)00:56
kristianpaulor milkymist demo, hum i guess i should have some code about interrupt handling wich is my first concern now00:58
kristianpaulwell, once i verify NCO is working..00:58
kristianpaulwhere is the nanonote when you need to solve some equations...01:28
kristianpaulah, got it, carrier_nco should be computed based on IF, that fore SiGe 4162 i real mode is 2.556 Mhz01:39
kristianpauland also namuru run at 16.348 Mhz, instead of the 40Mhz of the original namuru01:40
kristianpaulfControl (aka carrier_nco) = 2^30 * 2.556 / 16.38401:42
kristianpauleasy :)01:42
GitHub88[milkymist] kristianpaul pushed 1 new commit to gps-sdr-testing: http://bit.ly/pyrkRW02:14
GitHub88[milkymist/gps-sdr-testing] Update carrier nco comments to fit SiGE 4162 in realmode - Cristian Paul Peñaranda Rojas02:14
kristianpaulokay, i'm cofuse there is a accumulator interrupt period, that is 500us03:44
kristianpauli cant be upper than 1 ms,03:45
kristianpaulthis is rasoable and is related to gps specs i remenber03:45
kristianpaulbut there isalso a register called TIC, wich is recomended to set to 100ms period03:46
kristianpaulwell, this should be the interupt flag,03:46
kristianpaulman i forgot what is that 1 ms in GPS spects, i thnk is a chip but i need to cofirm with the book03:47
GitHub4[milkymist] kristianpaul pushed 4 new commits to gps-sdr-testing: http://bit.ly/r9CtqB03:57
GitHub4[milkymist/gps-sdr-testing] Minor name fix - Cristian Paul Peñaranda Rojas03:57
GitHub4[milkymist/gps-sdr-testing] Minor value fix - Cristian Paul Peñaranda Rojas03:57
GitHub4[milkymist/gps-sdr-testing] comment on code nco update to SiGE 4162 real mode - Cristian Paul Peñaranda Rojas03:57
kristianpauldamn bot dint show last commit..03:58
kristianpaulbrb, cold wind at midnight.. time to close windows :)04:31
kristianpaulokay, i think i have all right values now04:32
kristianpaulahh, the problem with endianess..i forgot04:33
kristianpauldamn,inverted "soft" reset signal that explain all those zeros..05:28
kristianpaulit tought i wasn lacking resgister initialization, but i see now there was something else..05:30
kristianpaulha, this namuru code is full of surprises, like why the accumulator register was 32 bits then it was commented and now is 16.05:59
kristianpaulmost if are you supposed to get navigation data from it... i need find tresholds, but that will be tomorrow i hope05:59
errordeveloper__hm .. so is the "8:10" card slot effectivelly an sd card or not at all ?07:25
wolfspraulerrordeveloper__: yes, like microsd07:26
errordeveloper__it's just the matter of branding .. isn't it ?07:31
errordeveloper__so the wpan module will work in any sdio slot ...07:32
wolfspraulhmm, that I'm not sure sure, we have to ask wpwrak_ 07:33
wolfspraulyou need to run Linux on that machine for sure, and you may need certain access to the sdio pins07:33
wpwrak_errordeveloper__: no, atben doesn't speak SDIO. you need to talk SPI to it. it de facto means that you need to be able to bitbang07:35
larsci.e. your sd-card host pins have to be accessible as gpio pins08:16
errordeveloperwpwrak_: I'm not worried about the SDIO so much08:36
errordeveloperjust the eclectrical and geometrical compatibility08:37
errordeveloperi.e. - would it fit in a card slot at work with linux on ... well say a laptop or a beagle board ..or sheeva plug ..08:38
wolfspraulyou need to be able to drive the individual pins, I think08:38
wpwrak_errordeveloper: if there's enough space, then it would fit, yes. the contacts are like on a uSD card08:38
wpwrak_errordeveloper: but you must be able to bit-bang your pins. your average laptop can't do that08:39
wolfspraulfor that you need to check the manuals/datasheets of the controllers that are driving the sd/sdio peripheral08:39
errordeveloperok, so it will be down to the host controller and the driver for it08:39
wpwrak_errordeveloper: and you should be able to disable/enable the 3.3 V voltage supply (to reset the chip)08:39
wpwrak_errordeveloper: (reset) you may be able to get away with just doing power-on reset, though. but that's a bit risky.08:40
errordeveloperwpwrak_: I had been using contiki and the Econtag hardware08:44
errordeveloperI really want to have direct communication between a linux host and wpan08:45
errordeveloperI had already discovered that sdio can be used for a drop-in solution08:45
errordeveloperthough at the time I could only find Speactec products08:46
errordeveloperwhich is a very obscure chinese brand08:47
wpwrak_errordeveloper: if your linux host has USB, you could use atusb. there, you can put whatever you want into the AVR. you have up to 28 kB to play with :)08:48
wolfspraulerrordeveloper__: Spectec is not obscure, you can consider it a real 'brand' almost08:49
wolfspraulit's a Taiwanese company, not Chinese08:49
wolfspraulthe real obscure stuff is very different ;-)08:49
wolfspraulSpectec is a brand company, and care for quality quite a bit. I can recommend the company, in general. You will also not find obscure Spectec clones, because the brand is too small for that.08:50
wolfspraulthe only thing to watch out is that their website is more a wishlist, what they really have in stock is maybe 10% of those products08:50
wolfspraulthe others can be built if you order 1000 or more08:50
wolfspraulit's a small company in Taipei, 20-30 people08:51
errordeveloperwolfspraul: yeah, just the spectec website is really badly designed08:52
errordeveloperhm .. I can see that there is also #milkymist channel10:30
errordeveloperwell, I quite like the idea ..tough not that much of a visuals person10:30
wpwrak_you can also ignore all the pretty graphics and do just FPGA hacking ;) although the graphics are nice to behold10:32
wolfspraulerrordeveloper: yes correct, the #qi-hardware channel is mostly meant about copyleft hardware, manufacturing, ben-wpan, ben nanonote, etc.10:32
wolfspraulwhereas #milkymist all starts with the milkymist ic design, then the milkymist one video synthesizer... at some point there's the overlap to #qi-hardware :-)10:32
wolfspraulyou can be relaxed in both channels though, nobody will complain about off-topic since it's all connected10:33
wpwrak_particularly not here10:33
errordevelopera) I kind of really need to change the job and 11:05
jivsneed some help with reflashing nanonote13:10
jivsI am trying to reflash nanonote with my own toolchain's images13:11
jivsI get all the three files I need for reflashing from toolchain i.e. u-boot.bin, uImage.bin and root.ubi13:11
jivsi can flash it successfully with the script, but nothing loads after the "starting kernel" message13:12
jivswhat might be causing this? boot loader, kernel or rootfs ?13:13
jivsplz help13:13
kyakdid you flash all bootloader, kernel and rootfs?13:27
jivsyes all of them13:28
kyakshould be fine then.. strange13:29
jivsyes it is, taking alot of time13:31
jivswhat do you think might be troublesome by looking at that situation13:31
kyakthere are some tasks to complete on first boot, but you should be brought to gmenu2x interface anyway13:31
jivsdoesn't go further than "starting kernel"13:32
jivseven if i get the terminal with al+ ctrl f2 , i will be happy13:32
kyakdo you build based on the release image or the one from trunk?13:32
jivsthen i can ssh to it13:32
jivsi tried both of them, but current one is from trunk13:33
kyaki can only suggest that you start a new clean build13:35
jivskyak, done at least in 3 new toolchain13:36
kyakdid you follow the isntruction here http://en.qi-hardware.com/wiki/Building_Software_Image exaclty?13:37
jivsi am trying to figure out which out of three is faulty, rootfs, kernel and bootloader13:38
jivsyes 13:38
kyakcan you upload your .config and build_dir/linux-xburst_qi_lb60/linux- somewhere?13:40
jivskyak,  plz wait13:41
jivshttp://pastebin.com/yK5qaCZD .config13:43
jivshttp://pastebin.com/RCiQdDbR build_dir config13:46
kyakyour configs are different from default ones13:53
kyaknot sure if it causes any problem, but you should try building from scratch based on default configs13:53
kyaki.e. cp feeds/qipackages/nanonote-files/data/qi_lb60/conf/config.full_system .config13:54
kyakthe kernel config is different, too13:54
jivsi have tried building with the default config as well13:55
kyakhas sound build as modules13:55
kyaknot sure where you got it13:55
kyakit has been built in kernel for some time now13:55
kyaki have to go now, sorry13:55
jivskyak, thanks for ur time13:56
jivsi will try with the .config from wiki again13:56
jivshopefully will work.13:56
kyakjivs: and do it from scratch. Previous (older) builds can mess up your compilation15:06
wolfsprauldoes anybody know whether the fisl videos are up for download now?15:50
kristianpauli think rejon should push for that, or wpwrak_ ....15:53
kristianpauli may be wrong but tv.softwarelivre.org looks not very updated, as the major tag is fro m 2009 :-(15:54
kristianpaul404 <- http://tv.softwarelivre.org/category/tags/fisl/fisl12 bad sign for me15:55
kristianpaulori'm looking bad place sinec fisl over :(15:55
wpwrak_kristianpaul: the site is a mess :-(15:57
wolfspraulsebastien was quick enough to take his own snapshot for most of Jon's talk15:57
wolfspraulbut the begining is a bit messy, and werner's talk is missing entirely15:58
wolfspraulso I was hoping to find a proper download at some point before the next news15:58
kristianpaulwpwrak_: claim for you maaterial to FISL ! :)15:58
kristianpaulor thing will neverbe there for sure15:59
kristianpaulthe page dint look like ccc video recondirg page or other floss related events indeed16:00
kristianpaulbad :(16:00
wpwrak_i already asked them. they weren't quite sure what was happening. the problem is that the conference is run by a company that does conferences. so they have all their weird tools and procedures. the free software folks aren't quite welcome to mess with that. they take care of program and content.16:01
wpwrak_in the old days, everything was handled by free software people. but it just became too much work for them.16:01
wolfspraulwhat does this mean? where is the download link?16:02
kristianpaulmean on download link anytime soon, (i hope i'm wrong )16:02
wpwrak_it's supposedly *somewhere* in the pipeline :)16:02
wolfsprauljust kidding, I know you cannot do anything either16:02
wolfspraulso Sebastien was 100% right16:03
wolfspraulkeep your tech in order, capture the stream, move16:03
wolfspraulvery Chinese16:03
wpwrak_yeah ;-)16:03
wolfspraulworst case have a cam on tripod ready, film off your notebook16:03
wolfspraulno questions, no delays, no licenses, just pure fun16:03
wolfspraulso I guess no werner talk for now, too bad16:04
wolfspraulwe had similar cases in the past, with Bas supposedly giving a superb talk at FOSDEM 201016:05
wolfspraullost in cyberspace forever...16:05
tuxbrainhi dudes :)19:54
tuxbrainhi wolfspraul!19:54
tuxbrainhi rejon :)19:57
wolfspraultuxbrain: hi19:57
tuxbrainvideos and presentations will be collected next week but the f4hc on Bilbao was a great success :) a lot of interesting people knows now about Qi!19:58
wolfspraulah btw, I like your laser stuff19:58
wolfsprauljust a little overloaded19:58
wolfspraulI added marcan's blog to the qi planet19:58
wolfspraulwe can think about a way to hookup a laser pointer to m119:58
wolfspraulone thing I dont' understand is how they go from a laser pointer to a laser that can be positioned on x/y coordinates?19:59
wolfspraulit seems they use audio to encode the picture? I need to read some more...19:59
wolfspraulmaybe we can include a future laser pointer in m1, who knows. if I can source it for a few USD and we can connect it and get it to work etc. then of course, why not :-)20:00
wolfspraulbut first I need to understand this stuff more, and I don't even have time to look into it right now20:00
wolfspraulit's 4 AM here :-)20:00
tuxbrain :) breif summary ,the laser is "moved" by mirrors conected to servo motors, trhough the "audio" using the ILDA protocol you pass X,Y, bright data to the microcontrolers that moves the servos. and tada laser projector :P20:05
kristianpaulhola :)20:06
tuxbrainI will reconect in short20:08
wolfspraultuxbrain: wow. and the motors and mechanical reaction of the mirror are fast enough?20:09
wolfspraultuxbrain is a busy man!20:09
tuxbrain_awaywolfspraul: yes I see them in action with a DIY one20:09
rohnice. which kind of servos? normal rc-model stuff?20:26
tuxbrainlet me search if I found some info on the projector20:32
tuxbrainAh! here it is http://marcansoft.com/blog/openlase/hardware-mark-1/20:35
tuxbrainroh: the servos with mirrors for laser are called galvos20:39
tuxbrainis matter to find a good cheap source from china :)20:41
tuxbrainroh: http://spacelas.com/html_products/35k-scanning-system-31.html seems the new version of the one's  that marcan is using in his DIY projector, seems they cost arround 88$ (~30k points per s) the full kit, higher res (50k points per s) is about 250$20:48
rohtuxbrain: yeah. was already wondering which servos would be fast enough20:58
tuxbrainwhat amaze me is that he wants to do a RGB laser... how he will align them to go out a a same beam? 21:01
tuxbrainwell watever, I have see it in action and I just see a MM conected to the Laser projector and with a vga projector and techno music louder as hell :)21:03
tuxbrainkristianpaul: me preguntaron en el f4hc como iba el proyecto GPS :)21:08
tuxbrainkristianpaul: fui lo mas evasivamente educado posible vi unos cuantos ojos brillar con el tema21:09
kristianpaultuxbrain: espero tener buenas noticias la proxima semana, almenos de la recepcion de detoas de navegacion21:24
kristianpaulinteresante, preguntaron algo especifico? o esperan algo especial?21:25
tuxbrainkristianpaul: que funcione y que sea barato :P21:28
kristianpaulfuncionara, eso si,pero el precio se sale de mi tema..21:29
tuxbrainkristianpaul: genial lo de los datos de navegacion :), espero ansioso las novedades21:29
tuxbrainya eso ya es cosa del wolfi y mia21:29
tuxbraino si se apunta alguien mas of course21:29
kristianpaulestos chips no son costosos, almenos  no de Maxin, el detall es la orden minima21:29
tuxbrainde momento eso que me has dicho me gusta :), necesita mucho procesador extra? se podria hacer un modulo para arduino?21:30
kristianpaulprocesador sip, bueno si haces todo en software es lo mas seguro, pero por ejemplo el nanonote podria hacer el procesamiento de la se~nal, pero no esperes mas luego21:32
tuxbraintalking about arduino modules: who was here interested/skilled to develop an Arduino version of the atben/atusb? also DocScrutinizer you were almost skilled in rf , how about bost mA of emision on the atben/atusb?21:33
kristianpaularduino, bueno, poder si, pero ya seria a nuestro nivel una board con fpga que emulara un receptor de gps hasta entregar datos NMEA..21:33
kristianpaulwhy arduino version?21:33
kristianpaulatben is mor that ready i think, just missing the lib/stack for it21:33
tuxbrainkristianpaul: I have some inertia in the arduino world and the target is quite wider than ben 21:34
kristianpaulsure sure21:34
kristianpaulmy point was, i think the hardware as it is is re-usable21:34
kristianpaul(i dont have nothing against arduino* btw)21:34
tuxbrainalso is easier to enter in the oem world though arduino than from ben21:34
kristianpaulOEM, oh really?21:35
kristianpauli tought was merelly hobbist and artits..21:35
tuxbrainyep, I have some costumers planing to do machines directly with arduino mini integrated in the board :)21:35
kristianpaulso is also ready to industrial grade?21:37
tuxbrainthey prototipe with UNO and then integrate with mini, is quite straigh fordward and adds not too much cost , an arduino mini clone I source is about 6¬21:37
tuxbrainis atmega chips industrial grade :P21:38
kristianpaulwill be good to know :)21:39
DocScrutinizertuxbrain: I'm available for development. What was it you were interested in?21:40
kristianpaulmay be an arduino atben shield :)21:40
tuxbrainwell whatever I'm talking about coffee machines , milds and some not critical dosificators.21:40
tuxbrainyes that's the idea21:40
tuxbrainan atduino shield :)21:40
kristianpaulnice, name !21:41
tuxbrainbut actual power (10m) is a little scarce to real applications, we must to empower the signal to achieve at least 100m optimal 1'5Km21:42
tuxbrainhi DocScrutinizer!21:43
tuxbrainok bbl is bye bye later :P , bye DocScrutinizer :)21:44
kristianpaulyeah, i still thinking about at* boards because power...21:46
tuxbrainkristianpaul: of course fisrt is to stabilize the soft with the ones we have but as is, they are enough for this right now, but I want to start thinking and make think others what is the best approach for that two goals atduino and +mA, and/or if this goals make sense to make them at time, an 10meters RF on arduino is almost useless, or at least hard to sell21:48
tuxbrainbut also maybe there is away in the arduino software side with what we have as is, conenectinging in some fancy way an atben or atusb21:49
tuxbrainthere is away in the -> there is a way to advance in the21:50
kristianpauli see, so you see that 10m is better sellable among arduinos, wich are most commmon for short distances apps right?21:52
tuxbrainI can provide with material(Arduino/atben/atusb) to anyone that want to try :)21:52
tuxbrainof course with the promise he will try hard :)21:52
kristianpauli dont mind having a low speed than wifi, but at least same/near range transmit power :)21:53
kristianpaulto be honest :)21:53
tuxbrainkristianpaul: I agree, or even more if possible :)21:54
tuxbrainkristianpaul: Gps arduino , which fpga chip will be needed to make to that NMEA conversor? very expensive?21:55
kristianpaulexpensive yes.. :(21:56
kristianpaulwell, we'll see, first make it work well, later optimize :)21:57
tuxbrainsure kristianpaul, don't want to unfocus the actual developer work :) nor on GPS or AT*, I just want to draft the lines of what I would like to be Tuxbrain nexts steps on hardware production.22:07
kristianpaulsure,is wise from you to tell us what you can buy/invest next :)22:09
tuxbrainAlso I want to know if I'm wrong in advance :)22:19
--- Tue Jul 19 201100:00

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