#qi-hardware IRC log for Thursday, 2011-03-24

wpwrakrjeffries: i don't think this is how it works. there are other marks with crosshairs, but they have a different role. the idea behind the fiduciary is that it's on the copper layer, which is the reference for everything else, and the resolution for lines/traces there is limited01:22
wpwrakrjeffries: at least now i understand what these strange circles do. before, i thought they were supposed to be ground points for use during testing. yet, strangely enough, they often didn't connect to any ground ...01:36
wpwrakgrr. s/fiduciary/fiducial/01:46
kristianpauloh dear, big storm today :S01:50
wpwrakkristianpaul: ah that's where it went. for today, heavy rain was announce for buenos aires. i went shopping tonight, hoping to get a free car wash. but not a single drop fell.01:51
kristianpauldrops are , but Lightning and wind are getting anoying..01:54
kristianpaulare oka*01:54
wpwrakkristianpaul: as long as there are no power cuts ... :)02:01
kristianpaulwpwrak: two atm02:01
kristianpaulvery short02:01
kristianpaulUPS is strong (up 6 hrs) also all my computers are low power :_)02:02
wpwrakkristianpaul: (6 hours) wow02:05
kristianpaulas i said my computers are low power :-)02:05
kristianpaulis a cheap USP02:05
mthcan anyone test this on the NanoNote? "hwclock -ru ; sleep 10 ; hwclock -ru"02:13
mthunder OpenDingux way more than 10 seconds pass on the hardware clock02:14
mthin fact, the RTC seems to be running 1024x faster than it should02:14
mthlars_: who should program JZ_REG_RTC_REGULATOR? the boot loader? the board init code? the RTC driver?02:29
kristianpaulhe, ethernet cord is a bit loose :_)02:46
kristianpaulwolfspraul: I have some pics !02:46
kristianpauli'll upload02:47
wolfspraulkristianpaul: cool!02:51
kristianpaulwolfspraul: multi upload dont support multi file description ;-)03:00
wolfspraulyes, I think you are right03:01
kristianpaulwolfspraul:  OBK-3130C default lense is M12 right?03:10
wolfspraulthe lense holder, yes03:13
wolfspraulI think all your lense holders are M12, no?03:13
wolfspraul12 = 12mm03:13
wolfspraulM9 = 9mm03:13
wolfspraulM12 = 12mm03:13
wolfspraulthat's just the size of the thread03:13
wolfspraulthere are other sizes like back focal length etc. that have to match for it to focus03:13
wpwrakreviving xue ?03:14
kristianpaulokay, i'll upload pics with default lense for all tree cameras and same "landscape"03:14
wolfspraulwpwrak: no, thinking about including a camera with Milkymist One03:15
kristianpaulwpwrak: how you in english  un vaso grande para tomar agua y demas?03:16
wolfspraulmaybe some of those insights can be used to revive a mobile milkymist or mobile camera project one day, but right now it's very specifically about whether/which camera to include with Milkymist One03:16
kristianpauli wonder all the names in english for that... :o03:16
kristianpaulah, holder..03:17
wpwrakkristianpaul: hmm, not quite sure if there's such an expression in english03:17
kristianpaulok i'll just call it gree vaso03:17
kristianpaulmore generic03:17
mthlars_: the Dingoo native OS is to blame: apparently the power-on setting is 1Hz, but the native OS reprograms it to 100 Hz (not 1 kHz as I thought) and that reprogramming survives a push of the reset button03:17
wpwrakwolfspraul: wouldn't the camera be very specific to the use(r) ? well, i guess it can't hurt to have an optional "first steps" camera for someone who just wants to explore03:19
wpwrak(waiting for asia to reconnect ...)03:19
kristianpaulwpwrak: so far from what i can see it is meant to be generic camera03:20
kristianpaulor what you mean with veryspecific?03:21
wpwrakkristianpaul: a good one or just something cheap that works ?03:21
kristianpaulwpwrak: actually the camera i liked was not the cheap one03:21
wpwrakkristianpaul: well, if you go to a club, you may have specific needs depending on the scene. and if you want a camera that can handle "anything", it would probably quite expensive03:22
kristianpaulwait some minutes i'm uploading some pics i hope explain better03:22
kristianpaulwpwrak: sure, for club we need to be very picky03:22
kristianpaullux*** picky :-)03:22
kristianpaulequals expensive !03:22
wpwrakyup. low light, wide focus/zoom range, etc. probably right in the semi-professional segment. USD 1k+03:23
kristianpaulfocus is other history indeed03:23
kristianpaulso all what we found now lets call it cheap after your nice point about semi-professional stuff :-)03:25
wpwrakhehe :)03:27
wpwrakwhat's actually the maximum camera resolution mm1 can handle ?03:28
kristianpaulthere is a chip for composite video, NTSC/PAL..03:28
kristianpauli dont remenber the specs03:29
kristianpaulbut PAL resolutios is given in TV lines i think03:30
wpwrakhmm, if it's only VGA or such, then that would mean a fairly unsophisticated camera03:30
kristianpauloh, sure03:30
kristianpaulflicernoise dont work on HD (yet)03:30
wpwrakhalf of asia is back ...03:31
kristianpaulwpwrak: http://en.qi-hardware.com/wiki/Xue/sensors#Milkymist_One_camera_comparison03:32
wpwrakkristianpaul: hmm, a comparison shot with a reasonably good still image camera (i.e., a real camera, not a mobile phone) might be useful03:36
kristianpaulwpwrak: i used my canon a470 for pics03:37
kristianpaulscreenshots are not posible yet03:37
kristianpaulI just a pic from LCD03:37
kristianpaulyeah :(03:37
kristianpaulwhat is real camera for you?03:38
wpwrakcanon a-series is okay-ish03:38
kristianpaulah good :-)03:38
kristianpauli was scared ;)03:38
wpwrakof course, a Canon EOS 5D Mark II or similar would be better (-:C03:40
kristianpauli'll stick for nanon as is i can run chdk, may be i can buy that after several years of saving ;)03:41
wpwrakyeah, i'm a canon fan, too. and i also like chdk ;-)03:42
wpwraksd880is here. i like 'em small03:43
kristianpaulnot bad, i need better mechanical zoom and aperture, chdk cant add that to hw ;)03:45
kristianpaulokay, wolfspraul, i still need take pics with less lighe enviroments..04:12
kristianpauli think i'm off bed now04:12
wolfspraulkristianpaul: wow, great pictures!05:11
wolfspraulthe artefacts are scary, but it's a starting point05:11
wolfspraulwe certainly will not revolutionize hollywood anytime soon :-)05:11
qi-bot[commit] nbd: mac80211: initialize the last rx time when creating a station http://qi-hw.com/p/openwrt-xburst/3ae6fa106:18
qi-bot[commit] cshore: [package] hotplug2: Added zaptel subsystem to /etc/hotplugs2.rules so that the zaptel kernel module package only needs to had a script to create the correct device nodes (default names differ from what all apps that use zaptel actually use, so a script is necessary). http://qi-hw.com/p/openwrt-xburst/06a1c1c06:18
qi-bot[commit] jow: [package] ncurses: install ncurses5-config and ncursesw5-config (#9044) http://qi-hw.com/p/openwrt-xburst/f73979c06:18
qi-bot[commit] thepeople: sysupgrade works on the wp54 http://qi-hw.com/p/openwrt-xburst/d3db86b06:18
qi-bot[commit] jow: [tools] ipkg-utils: utilize the tar -p flag to preserve permissions (#7667) http://qi-hw.com/p/openwrt-xburst/ebdf11806:18
qi-bot[commit] jow: [include] image.mk: only upgrade permissions instead of overwriting, utilize tar -p flag for targz image targets (#7667) http://qi-hw.com/p/openwrt-xburst/d80e0f106:18
qi-bot[commit] nbd: add support for md5sum checks for mirrored tarballs of packages with version control source urls http://qi-hw.com/p/openwrt-xburst/45e427506:18
qi-bot[commit] nbd: hostapd: update to 2011-02-21, use PKG_MIRROR_MD5SUM, includes fixes for WPS http://qi-hw.com/p/openwrt-xburst/f67f19506:18
qi-bot[commit] nbd: ath9k: add a patch from linux-wireless@ for reducing driver size http://qi-hw.com/p/openwrt-xburst/d69293906:18
qi-bot[commit] kaloz: enable the 'big kernel lock' by default http://qi-hw.com/p/openwrt-xburst/5bac58406:18
qi-bot[commit] kaloz: [ixp4xx]: add 2.6.38 support http://qi-hw.com/p/openwrt-xburst/70a23a106:18
qi-bot[commit] jow: [include] target.mk: don't enable coreutils by default, it does not even exist anymore... fixes the Image Builder http://qi-hw.com/p/openwrt-xburst/909ceae06:18
qi-bot[commit] jow: [ar71xx] fix image generation in the Image Builder environment http://qi-hw.com/p/openwrt-xburst/957986706:18
qi-bot[commit] nbd: hostapd: make entropy collection contribute to the kernel pool http://qi-hw.com/p/openwrt-xburst/2dbaa7806:18
qi-bot[commit] nbd: mac80211: update to wireless-testing 2011-03-22 http://qi-hw.com/p/openwrt-xburst/5034ff806:18
qi-bot[commit] hcg: [omap35xx] Add patch to correct sub-page alignment http://qi-hw.com/p/openwrt-xburst/09cfea706:18
qi-bot[commit] hcg: [omap35xx] kernel: add CTI usb-id's to FTDI driver http://qi-hw.com/p/openwrt-xburst/d88135f06:18
qi-bot[commit] hcg: [omap35xx]: Correct CTI patch http://qi-hw.com/p/openwrt-xburst/4a907de06:18
qi-bot[commit] nbd: ath9k: merge a pending patch for fixing a stopped queue issue (mostly for client mode) http://qi-hw.com/p/openwrt-xburst/f8662e206:18
qi-bot[commit] nbd: ath9k: add a few de-bloating and optimization patches http://qi-hw.com/p/openwrt-xburst/b139cef06:18
kyakxiangfu: i just updated the trunk :)06:20
kyakxiangfu: hope it went well... since we didn't commit anything to qi trunk yet06:20
wpwrakwolfspraul: the images are thrice "analogized" ... i don't think you can really tell what quality you have there before getting digital screenshots. also, things like gamma correction may make a huge difference.08:03
wpwrakwolfspraul: the photos basically scream "PLEASE correct my gamma !!!" ;-)08:07
qi-bot[commit] Werner Almesberger: atben.brd, atusb.brd: set solder mask opening of fiducials to 2 mm http://qi-hw.com/p/ben-wpan/7e20be908:11
qi-bot[commit] Werner Almesberger: README-PCB: relaxed antenna via size from 8 mil to <= 15 mil http://qi-hw.com/p/ben-wpan/af6d56308:11
qi-bot[commit] Werner Almesberger: atben. atusb: bumped version numbers to 110324 http://qi-hw.com/p/ben-wpan/f24236808:11
wpwraktuxbrain: http://downloads.qi-hardware.com/people/werner/wpan/tmp/atben-110324-overview.png and http://downloads.qi-hardware.com/people/werner/wpan/tmp/atusb-110324-overview.png08:11
wpwraktuxbrain: also, http://downloads.qi-hardware.com/people/werner/wpan/fab/*110324*08:12
wpwraktuxbrain: i also checked the prices at digi-key of the major components - no change since i did the last BOM (with prices from february ~14-15)08:17
tuxbrainwpwrak: thank dude, today I will pretty offline, but I will keep an eye on mail to see if smt vendor finally decide to tell us how much08:19
wpwrakyeah, they seem to be dancing around that topic. i guess if you have to ask, you can't afford it ;-)08:20
kyakxMff: ping10:18
kyakhey :)10:21
xMffhm sorry, I have to run. Will check in later10:21
kyakxMff: could you explain how CONFIG_ALL works? It's pretty strange, because when i choose it via menuconfig, it doesn't select all packages as "m"10:22
kyakok, no problem10:22
kyakthen to make CONFIG_ALL work, i have to rm the .config10:22
kyakand run menuconfig again and choose the CONFIG_ALL10:23
kyakit feels like CONFIG_ALL respects those packages, which are already "is not set"10:27
kyaktherefor CONFIG_ALL=y only works the first time you generate .config10:27
kristianpaulwolfspraul: he, well screenshots i think will make things more clear10:46
wolfspraulyes, we'll get screenshots to work :-)10:49
kristianpauli'll try what lekernel said about black background10:50
kristianpaultake some pics, record a video :-)10:50
wolfspraulkristianpaul: I will also look for a (manual) zoom lense10:53
viricare the x264 decoders floating-point only?10:53
wolfspraulthinking about your pictures, I realized a 'good' setting will very much depend on the location/environment10:53
wolfspraulso that will only work with a zoom lense10:53
kristianpaulwolfspraul:  (manual) zoom lense10:55
kristianpaulnice !10:55
kristianpaulviric: i dont know, but you ask lekernel at #milkymist10:56
kristianpauli think they are but not sure at all, i dint take a look at the pfpu10:57
virichm ok11:02
viricthank you!11:02
wolfspraulkristianpaul: so you agree about the zoom lense idea?11:12
wolfspraulI think our system is not about end-to-end quality right now11:12
wolfspraulwe don't even know where the bottleneck is11:12
wolfspraulbut from a use-case, I think you want to adjust the zoom setting to the location you use the m1 at11:13
wolfsprauland how is that possible with a fixed zoom lense, whether it's wide angle or not11:13
wolfspraulso we rather have a manual zoom lense 4-9mm or so11:13
wolfspraulI need to look at cost, but no matter how cheap a lense we take, I doubt the lense performance will be our optical bottleneck.11:14
kristianpaulwolfspra1l: (zoom) i agree14:09
kristianpaulWell, sebastien already point us the main use case for the camera, so that zoom at least you fit that case14:10
wolfspra1lmy network is crazy unstable lately14:28
wolfspra1lI think the Chinese government is going all out in attack mode14:28
wolfspra1lhopefully I can get a few more things done today :-)14:29
tuxbrainwpwrak: smt vendor confirm fiducials are ok,14:41
tuxbrainwpwrak: also smt vendor is in charge to aply the solder paste, so they will be the destinators of the stencils14:42
wpwraktuxbrain: but will they need stencils to be made, will they make the stencils themselves, or will they apply the solder paste without stencil ?14:43
tuxbrainthey ask for the stencils, if not they will charge us for them I supose14:44
wpwrakokay. good. then you need a quote from the pcb fab for the stencil as well.14:44
tuxbrainthis info are not in the gerber files_14:45
wpwrakit's just the solder paste layer14:45
wpwrakbut you have to tell the pcb fab that you need a stencil as well, and not just the pcb14:45
tuxbrainmmm this is not what the tell I guess,14:46
tuxbrainEl proceso normal de un montaje es : recibes los circuitos tal cual , y aquí14:46
tuxbrainen nuestras instalaciones , primero pasan por Serigrafia  ( donde se aplica14:46
tuxbrainpasta de soldar , por lo que se necesita una pantalla de Serigrafia )y luego14:46
tuxbrainpasa a  pick&place y horno14:46
tuxbrainc&p of the mail14:46
wpwrakserigrafias = printing, using the stencil14:46
tuxbrainso I might ask to the pcb vendor to make the stencils isn't it?14:47
tuxbrainmight-> should14:49
wpwrakyup. there are also different kinds of stencils. but i guess they'll know what the smt fab uses.14:49
wpwrakhere's an example: http://www.pcbunlimited.com/smt-stencils.php14:49
wpwrakshouldn't be overly expensive14:49
tuxbrainI hope so you scared me with the prices of that page14:50
wpwrakUSD 150 ? why ? that's peanuts :)14:51
wpwrakah, one thing to cut costs could be to combine atben and atusb on the same panel. you may want to ask the pcb fab if they can do that (and if it really does lower the cost)14:53
tuxbrainpeanuts and water will be my diet in short so for me are quite important :)14:53
wpwrakdrawback: you'd have a fixed atben/atusb ratio14:53
wpwrakadd a beer and a little coffee, and you have something from all the important food groups, i.e., fat, alcohol, caffeine ;-)14:54
tuxbrainmmm I don't think it will low costs two much... also it has the big draw back than if for example atusb are good selling and atben not, you will have to pay again to do a atusb setup only...14:54
wpwrakyup. okay, not mixing them is cleaner. it's just one option to consider if you must cut costs.14:57
wpwraknow we just need to find out what the smt would cost :)14:59
tuxbraingive me a good software stack than make them plug an play and brings up a /dev/wpan0 able to accept dhcp and cost will be reduced by volume :)15:00
wpwraknetwork interfaces don't need /dev nodes :)15:01
tuxbrainmore in my fabour15:02
wpwraktuxbrain: oh ... i see what you mean. indeed, no solder paste layer in the fab packages. fixing ...15:18
tuxbrainthat means I have not to ask to pcb maker?15:19
wpwrakno no ... it means that i forgot to include the file ;-)15:19
wpwrakmaking these fab packages is a heck of a lot of work. good that the next set will be much easier, with all the tools in place15:22
wpwrakof course, i'll probably hear adam's scream all the way to buenos aires when he sees this little file: http://projects.qi-hardware.com/index.php/p/ben-wpan/source/tree/master/makefiles/Makefile.kicad15:23
wpwraktuxbrain: ah, have you gotten a response regarding the mounting of the usb connector ? i.e., should the stencil provide solder paste for it ? and if yes, on both sides ? (both sides means that you need either two stencils or a pcb that has half of the parts flipped)15:28
tuxbrainI they will just will be in charge of the usb due is out of the smt process15:34
wpwrakparse error :)15:35
tuxbrainwpwrak: (adam screaming) a real fun image  to have in mind15:35
wpwraktuxbrain: so .. what will happen with the usb connector ?15:37
tuxbrainthey will just put in place soldered15:38
wpwrakyou mean they'll solder it by hand ?15:38
wpwrakor will they insert it by hand, but reflow it with the rest ?15:39
tuxbraindon't know, ok I will satisfy your curiosity, I will ask :)15:39
wpwraki need this information to determine whether to apply solder paste or not. the options are: 0) no solder paste for the usb connector, 1) only on front, 2) on both sides.15:40
wpwrakhmm, i'll add a via to make sure ground on both sides is connected. just in case ..15:43
wpwraktwo vias. now nothing can go wrong :)15:45
tuxbrainchanging the topic, how on hell spetec could achive to put the wifi stuff in such ridiculous narro space... what would requiere to us to do at atben of such dimmensions (apart of surelly indecents amounts of money)15:50
wpwrakeasy: use microvias, put components on both sides, use a chip antenna15:55
wpwrakand use a four-layer pcb while you're at it15:55
wpwraknothing insane. just a little outside of the reach of my DIY process -> expensive trial and error cycles. but okay for a 2nd generation product.15:56
wpwrak(expensive) both in terms of money and time. maybe USD 1000-2000 per revision. and 1-2 weeks turn-around time for each. but again, that's okay if you know what you're doing, because then you'll need maybe 2 trial versions and you're good.16:00
tuxbrainby the procces you are descriving then is even posible to make a Nanonote watch :)16:01
wpwrakhmm, if you like your watches rather big, like this one: http://gizmodo.com/#!5445595/thomas-cruyls-wristwatch-is-cuckoo16:02
tuxbrainlol , well ok I must leave16:06
wpwraktuxbrain: i'll have to wait until midnight for the next fab package. already used the 110324 codepoint :)16:11
tuxbrainhahahaha then the codification is inneficient for you productivity :)16:12
tuxbrainman just smash the old one :P16:12
wpwrakor rather my error rate :)16:12
wpwraki could start adding letters ... 110324a, b, ...16:13
wpwraktuxbrain: well, the final version will depend on the response we get regarding the USB connector. so i'll have to wait for this anyway16:16
wpwraktuxbrain: btw, you can already ask the pcb fab what the stencil will cost. the cost shuldn't change it we make small modifications16:18
qi-bot[commit] Werner Almesberger: README-SMT: various small clarifications http://qi-hw.com/p/ben-wpan/38bea8e16:21
qi-bot[commit] Werner Almesberger: atusb.brd: added two vias to connect front and back ground at USB shield http://qi-hw.com/p/ben-wpan/1c4c78e16:21
qi-bot[commit] Werner Almesberger: added the front solder paste to the PCB fab package (for stencil production) http://qi-hw.com/p/ben-wpan/269f77016:21
qi-bot[commit] Werner Almesberger: atusb.brd: moved FIDUCAL label away from mounting hole; bumped version number http://qi-hw.com/p/ben-wpan/6ff274016:21
wpwrakhmm, i should have used this USB connector: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=WM3983CT-ND16:30
wpwraka little cheaper and also a bit "less through-hole". ah well, something for atusb v2.016:31
tuxbrainwpwrak: the usb connector can be smt? smt vendor suggest that to avoid two stencils and two printings18:47
tuxbrainwpwrak: one of your naps_19:23
wpwraktuxbrain: i haven't been able to find any USB connector that completely avoids being through-hole. there is one that is slightly SMT-friendlier, but still requires a hole19:23
tuxbrainok I think they have one of those19:24
wpwraktuxbrain: an more radical alternative would be to eliminate the USB connector entirely and to make one of the PCB19:24
wpwraklemme show you a picture ...19:24
tuxbrainI  know how they look like19:24
tuxbrainbut I don-t like it19:25
tuxbrainsave this option when we use microvias and 4 layer pcbs19:25
tuxbrainI will thell the smt to go ahead with his smt USB connector and only one face to deal with19:26
wpwrakwait ... you say they propose a different connector ?19:28
tuxbrainan smt one19:28
wpwrakdata sheet ?19:29
tuxbrainok I will ask, will make any difference?19:29
wpwrakthat's what i'll know when i see the data sheet ;-)19:30
wpwraktuxbrain: here's the pcb-only variant: http://downloads.qi-hardware.com/people/werner/tmp/usbpcbconn.jpg19:36
wpwrakthe white stuff on the bottom is to add thickness19:37
tuxbrainI have asked for all datasheets of their alternatives they propose in the quoting , to be validated by our "Enginering department" so I prefer to tell them to go ahead with the quoting an then we will enter in matizations if something don't convince us19:37
wpwrakmatization ?19:39
tuxbraindoesn't exist this word in english?19:40
tuxbrainok matizacion(ES)->Clarification(EN)19:41
wpwrakah, learned a new word ;-) i was already wondering if you meant something related to "matanza" :)19:43
tuxbrainhaahaha this is reserved if things go wrong :P19:44
tuxbrainand after production started, before it starts we are all friends :P19:45
kristianpauleliminate the USB connector !!20:14
kristianpaulhmm, creo que matizaciones era solo para luces y colores20:15
kristianpaulinteresting :-)20:15
wpwrakkristianpaul: (eliminate the connector) unfortunately, this isn't entirely trivial. my board is 0.8 mm, and the total thickness has to be around 2.0 mm. so you either use 1.6 mm and add some material (e.g., the white stuff in my picture), or you use a 2.0 mm board20:16
wpwrakkristianpaul: the problem with making the board thicker is that this changes its RF characteristics. so the feed line gets wider and the antenna needs to be redesigned20:17
kristianpaulyeah, i missed the thickness :/20:21
kristianpaulbut hey, this is atben, is so dramatic therorically tought about a thick change?..20:23
wpwrakit's atusb. atben is exactly right at 0.8 mm :)20:23
kristianpaulyes yes, sorry20:24
wpwrakwell, if i change the board thickness, i need to redesign the antenna, make various samples, measure them, etc. that takes a while. this sort of thing: http://downloads.qi-hardware.com/people/werner/tmp/antfarm.jpg20:26
methril_workwpwrak, nice picture20:34
tuxbrainyeah! nanocoder draft compiled and running on NN :)22:37
tuxbrainnanokoder sorry :P22:37
tuxbrainand as I fear qt file dialog is too big for use on NN22:38
tuxbrainat least as I call it, maybe it has some conf to adjust to a 320x24022:39
rjeffrieswpwrak cn you ship me a 40 ft container full of chairs similar to this? http://downloads.qi-hardware.com/people/werner/tmp/antfarm.jpg22:43
wpwrakrjeffries: i''ll pass the request to my carpenter :)22:46
wpwrakrjeffries: he'll be happy to hear that you like this avantgarde design. like stone-washed jeans, just for furniture22:48
tuxbrainwpwrak: sell it pricy , is an RF chair22:49
mthtuxbrain: are you using Qt Embedded or Qt X11?22:49
tuxbrainI'm using qt4, using the openwrt tool chain than use frambuffer22:50
rjeffrieswpwrak I knew you were an authority on ATM and communications matters, had no idea you were also an artisan with wood22:50
tuxbrainhi Ayla22:50
Aylaanyone uses gmenu2x?22:50
rjeffriesElectric Chair, used once, from Argentina.22:50
wpwraktuxbrain: anechoic wood22:51
wpwrakrjeffries: (-:C22:51
tuxbrainwpwrak: you enjoy using words than I don't understand , isn't it?22:52
wpwrakrjeffries: i actually did a few things with wood. when my carpenter was too slow, i made my lab benches and a cupboard myself22:52
Aylawould you like to have gmenu2x work in a more unix way, with user conf/skins in $HM22:52
Ayla$HOME, and data in /usr/share/gmenu2x?22:53
wpwraktuxbrain: anechoic = no echo (or, in the case of RF, no reflections). you use something called an anechoic chamber for proper RF or also audio tests22:53
Aylamy question is more, can I get rid of the current behaviour, ie. reading only the files located on the binary path22:53
tuxbrainwpwrak: another piece of brain occupied by werner's knowledge22:54
tuxbrainAyla: if the questions is than if I like the idea, yes I like it , but I'm just a user not developer :)22:55
Aylawell I don't know how it is used right now by the nanonote,22:56
Aylabut I believe it is an improvement22:56
tuxbrainAyla: If you have the knowledge to do so go ahead :), who was the gmenu2x maintainer?22:57
Aylawell, I have the necessary rights to commit on the repo22:58
Aylabut I don't want to break things by introducing policy22:58
mthmaybe ask xiangfu?22:59
Aylamaybe Qi guys prefer how it's currently set up22:59
Aylaif I can suppress the current behaviour, that saves me lots of time and code23:00
mthI think a more unix-style dir would be better, but it could break existing packaging scripts23:00
mthyou could create a branch and if the change is accepted, merge it to master23:00
tuxbrainusing you approach will be more easy to get rid of root as normal user but will also not hurt meanwhile it don't occurs , isn't it?23:00
Aylabasically, I just want to change the directories gmenu2x checks for its files23:01
Aylamth, what do you mean by packaging scripts?23:01
tuxbrainAyla: Openwrt makefiles23:02
mththe integration into openwrt23:02
wpwrakAyla: if you could search the old and the new locations, that probbaly wouldn't break anything23:02
wpwrakAyla: then, after a while, turn off the old23:02
mthnot sure how openwrt does it, but buildroot copies files from the build dir to the target dir23:02
mthit doesn't run "make install"23:02
Aylayes, that's what I've done right now23:02
Aylabut the code is quite ugly as there're exceptions everywhere23:02
mthan advantage of branching is that you can show the new situation to other people23:05
mthand they can test it to see what kind of impact it has23:05
Aylayes, maybe it's a better idea23:07
Aylathan keeping the current way while trying to improve it23:07
qi-bot[commit] Werner Almesberger: fab/drl2gerber: new tool to convert an Excellon drill file to Gerber http://qi-hw.com/p/eda-tools/813ac4723:11
qi-bot[commit] Werner Almesberger: prettygerbv: new option  -d file  to show holes from a coverted drill file http://qi-hw.com/p/eda-tools/e52f0db23:11
qi-bot[commit] Werner Almesberger: makefiles/Makefile.kicad: also show drill holes in PNG illustrations http://qi-hw.com/p/ben-wpan/0ee8b4223:11
rohwpwrak: at some point you should make a flowchart of fileformats und and tools to convert them23:12
wpwraktuxbrain: now with holes: http://downloads.qi-hardware.com/people/werner/wpan/tmp/atben-110324-overview.png http://downloads.qi-hardware.com/people/werner/wpan/tmp/atusb-110324-overview.png23:12
rohflowchart for 'what is done in which format and how is the workflow'23:13
wpwrakroh: yeah, i thought that earlier today as well. it's getting a little ... nontrivial :)23:13
rohone thing i havent understood yet: why is there a gnd plane on the microsd-end of atben?23:13
tuxbrainwhat are that holes for?23:13
rohon the bottom.. sure that doesnt make trouble?23:13
wpwraktuxbrain: vias, through-hole (usb connector), positioning (usb connector)23:14
wpwrakroh: it can only improve things :)23:14
wpwrakroh: (by making a better ground contact)23:14
rohwpwrak: hehe.. just mentioning that because its getting difficult for ME to keep track of such things.. and i am more into such processes as the regular guy.. so getting an overview is harder than a year ago now23:14
wpwrakroh: signal and 8:10 shield ground are connected in the ben23:15
rohi see.23:15
wpwrakroh: yeah, when the dust settles, i'll have to make a little drawing with a lot of arrows23:16
rohwpwrak: is there really a gap in the antenna on atben?23:16
wpwrakroh: rounding error23:16
rohat the 90° point of the gnd and the signal feed onto the 'antenna'23:16
tuxbrainthat ones I understad , was that little ones speraded arround the board23:16
rohah. so its only in the png, yes?23:16
wpwraktuxbrain: that's vias23:17
wpwraktuxbrain: see also http://downloads.qi-hardware.com/people/werner/wpan/tmp/2boards-20110305.jpg23:17
tuxbrainok to connect to the other side, the big ground , isn't it?23:17
wpwraktuxbrain: yup. or also for signals23:18
tuxbrainok that ugly bumps :)23:20
tuxbrainroh, any idea on atben/atusb envelop?23:20
rohnot yet. its _really_ small23:21
rohits much smaller than i expected.23:22
roheven if i knew... its just the.. having it infront of you... hm.23:22
wpwrakwelcome to the marvelous world of microelectronics :)23:23
tuxbrainanother motto , price are high but damn it, they are _really_ small! :P23:24
tuxbrainroh, then maybe the best solution is a bath of plastic or silicon, with a beatifull tuxbrain logo on it :P23:28
rohi think for the atusb its a question of finding the right usb-stick-plastic case23:30
rohsomething made in china consisting of 2 parts being pressed together23:30
wpwrakroh: yeah, that would be an elegant solution. they may require some openings in the pcb, though.23:31
wpwraktuxbrain: i think silicone is what we'll end up with, at least this time23:31
rohwpwrak: i just tried one which i had around from a fake-class-2-bluetooth-donge i took apart23:32
wpwraktuxbrain: if you use transparent silicone, you can simply put a tuxbrain sticker at the bottom, where you have plenty of space and where there is no other text23:32
wpwrakroh: pictures ! :)23:32
rohand it doesnt fit.. the usb plug needs to stand out further than its possible with the notches.23:32
wpwrakroh: (stand out) hmm, you mean the board+case assembly can't be fully inserted into the usb receptacle ?23:34
rohmeans the pcb needs to be tighter to the socket23:34
rohwpwrak: exactly.23:34
rohi think if i dont find another case i will trim away a bit of pcb there and fit it in23:34
wpwrakroh: hmm, it's already at the limit23:34
wpwrakroh: ah, so the case has some press-fit "columns" on the PC side, after the PCB ?23:35
rohit can be wider further away from the socket. just not at the 'corners of the case'23:35
rohexactly. something like small noses and tubes23:35
wpwrakah, the width. okay, that can be trimmed a little23:36
wpwrakand another fped bug fix ...23:46
--- Fri Mar 25 201100:00

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