#qi-hardware IRC log for Thursday, 2010-09-09

xiangfuHi all. I found the 0.9.32 uClibc enable the NPTL. I am not sure if this NPTL make "dgclock" "mkfs.ext2" not work.01:34
xiangfuI found those info about NPTL in uClibc's menuconfig:01:34
wolfspraulxiangfu: I think I remember Mirko said that nptl might be necessary for some new qt version. not sure though.01:35
kyakthere is a problem: 0.9.32 uClibc stop some programs from working correctly. At the same time, those programs work correctly with uClibc, but then qt4 acts weird. For example, it can't correctly exit an application that uses directfb01:49
kyaki'm talking about NanoMap, but i also tested it with a qt4 "hello world"01:49
kyakand i also mentioned several times it has something to do with nptl, because uClibc lacks nptl and some threads of qt4 are not terminated correctly (just an assumption). at the same time, some other application (by their old design, i assume) can't work correctly with nptl version of uClibc01:52
wolfspraulkyak: ok! :-) and what now?02:01
wolfspraulI suggest we try to fix the 0.9.32 with nptl version.02:02
kyaknow nothing02:02
kyaki pinged mirko some time ago with questions about qt4-* pacakges, but he's not around02:02
kyaki think he's the only one who can figure out this issues with qt402:03
wolfspraulhe's in Bali :-)02:03
wolfspraulhe will only return in October02:03
kyaki remember02:03
kyakhe went to Bali by mistake :)02:03
kyakpretty funny02:03
elricsfatehello all03:00
elricsfateanyone not idling?03:01
elricsfatewell anyway I am looking for some fun ideas on what to do with my soon to be nano note03:02
elricsfateanyone have any suggestions?03:03
kyakdefine "fun"03:05
elricsfatefun= something thrilling, entertaining, or educating.03:06
elricsfateI am already looking into using it to flash arudino which will be rather useful03:07
elricsfateI have plenty of ideas for the Arduino I have coming in but not to many for the nanonote03:09
kyaki've found it interesting for myself to play with openwrt toolchain03:11
kyakand porting some software, too03:12
kyakvery educating03:12
wolfspraulelricsfate: thanks for buying one, first of all! what are you good at, and what do you want to learn?03:12
elricsfatekyak: Roger. I have thought about installing an internal wireless card by stripping a small linux compatible wireless usb card.03:13
kyaknot possible, cause Ben lacks for usb host -\03:15
kyakhowever, you can buy a microSD Wi-Fi card03:15
kyak(i had one, but never got it working, and at was broken at the end)03:16
elricsfatewolfspraul: I have general linux knowledge. Most of my knowledge pertains to troubleshooting and system/network administration. As far as what I would like to learn? As much as I can about everything haha. I am dabbling in electronics and the such (have an Arduino coming in the mail. Lots of ideas for that) and I am currently a computer science major (I am a freshman and haven't started in on the programming part yet) . I wo03:17
elricsfatewolfspraul: So basically I am telling you I intend this to be a large learning experience. Haha.03:17
elricsfatewolfspraul: Also your welcome.03:17
kyak...and IRC's 512 bytes limitation comes in action :)03:18
wolfspraulkyak: your wifi card broke? didn't know that. what happened?03:18
wolfspraulelricsfate: wow, yeah. you seem open minded.03:18
elricsfatekyak: Damn :/ there goes that idea. I am sure I can find some way to make it internal.03:19
kyakwolfspraul: i'm not sure.. at some point i was gentle enough inserting it into microSD slot -\ first it started pilling off (the external plastic), then it became completely useless03:19
kyak*i wasn't gentle03:19
elricsfatewolfspraul: I try to be. I enjoy learning. I started working on computers when I was about 12.03:20
wolfspraulelricsfate: the best challenge, and way to learn for you, around the NanoNote is probably to make it as useful as possible on the software side.03:21
wolfspraulthe arduino will be better suited to hookup some simple peripherals etc.03:21
wolfspraulso maybe like kyak dive into the OpenWrt system03:22
wolfspraulwe are trying to bring more and more valuable software and content to the NanoNote03:22
wolfspraullike NanoMap, offline wiki readers, dictionaries, etc.03:22
wolfspraulit's not a smartphone, it lacks tons of 'sexy features', like 3G, Wi-Fi, touchscreen, USB Host/on-the-go, GPS, whatnot03:23
elricsfatewolfspraul: Roger. Luckily thats where most of my experience is at. Making things mesh together into something coherent. Now if I would just learn to code haha.03:23
wolfspraulyes the NanoNote, if you are patient, will allow you to dive deeper that's for sure03:23
wolfspraulbut it's an exercise in frustration tolerance, gotta know :-)03:23
elricsfatewolfspraul: thats the great thing about the system. it lacks "useless feautures" lol.03:24
wolfspraulso I use mine mostly for note taking on trips (vi)03:24
wolfspraulthat's not enough!03:24
wolfspraulbut getting more onto it is a big challenge, luckily many people help, we have a projects server, people are contributing etc.03:24
wolfspraulif I look at the hardware specs, and I imagine what it _could_ do (theoretically) after I press the power-on button, oh well, we are maybe able to actually do 5% of that today03:25
elricsfatewolfspraul: Trust me frustration I am used to. I actually work for a GSP company. linuxgamingservers.com. I once stayed up until 5 in the morning helping my boss try and figure out how to get a non-linux native garrysmod server running correctly. That was fun. SSH FTW.03:25
wolfspraultheoretically, you press the power-on, it's up in 3 seconds, and it gives you lots of useful apps, pim, music/video viewer, games, dictionary/wiki/maps. it can sync with your notebook when you plug it in. software can easily be updated. and so on.03:26
wolfspraulthat's the theory03:26
elricsfatewolfspraul: That sounds pretty easy to make happen if thats what you envision for this device. :D I am optimist I guess.03:26
wolfspraulthe reality today is 5% of that, especially for non-technical end users who cannot install this or that or work their way through configuration and terminals03:26
wolfspraulbut ... all is not lost yet :-) we are around, improving the little thing, and slowly people are joining too, so I'm glad you show up here!03:27
wolfspraulthere are many things going on on the hardware side too, but they are a bit harder for you to jump into03:27
wolfspraulsome GPS hacking, some 802.15.4 RF hacking, 3D work, a board that adds an FPGA (SIE), a whole GPL licensed CPU actually (Milkymist)03:28
wolfsprauland a camera too (Xue)03:28
elricsfatewolfspraul: I have seen that debian (which I will most likely install) has been ported to the nano note.03:28
wolfspraulon this channel, the most active Debian guy right now seems to be nebajoth03:29
elricsfatewolfspraul: I believe (I am going to see how it works and all) it would be much easier for an end user to work with something such as debian03:29
wolfspraulwell you have to try, there are many pros and cons03:30
wolfspraul32 MB SDRAM is a big limitation for Debian03:30
elricsfatewolfspraul: I know you have your vision of the device and I think debian would allow you to reach that vision rather quickly.03:30
elricsfateAgreed the RAM is a problem03:30
wolfspraulthe Debian situation should get better after more kernel features go upstream, which will happen in a big way in 2.6.3603:30
wolfsprauland then boot time03:30
elricsfatebut I am curious. Shouldn't it be possible to use the nand flash as a sort of cache? I know it won't be as fast as RAM but it should help.03:31
wolfspraulI'm all for diversity, nothing against Debian. Personally I enjoy battling with OpenWrt, another effort you might want to check out is the OpenEmbedded based Jlime (www.jlime.com)03:31
elricsfateor if not the nand possibly the SD car03:31
wolfspraulyes people put a swap on the sd card03:31
wolfspraulyou will find out all about that soon :-)03:31
elricsfatewolfspraul: I am not sure but is the nand faster or the sd?03:32
elricsfatewolfspraul: Also I am looking at the jlime as we speak.03:32
wolfsprauldepends on the SD card I think some SD cards may be faster03:32
wolfspraulif you want to really learn kernels botom-up, check out the Iris kernel03:33
wolfspraula little LED play http://en.qi-hardware.com/wiki/File:Leds.ogv03:34
elricsfatewolfspraul: Also checking out iris. I am sorry for all the questions but as I said I enjoy learning. If someone is handing out free knowledge then why not take it :D.03:34
wolfsprauloh there is plenty of that here, too much maybe :-)03:35
wolfspraulso you need to follow a path that makes sense for you03:35
elricsfatewolfspraul: I guess thats my biggest problem. I try to learn and dabble in a little bit of everything. I am trying to learn about the hardware and software portion of things as they are both related. Ya know?03:36
elricsfatewoflspraul: Also are you part of the dev team? Community member?03:38
wolfsprauldev team, server admin, NanoNote salesman03:40
elricsfatewolfspraul: Awesome. Software or hardware side of things as far as dev goes?03:40
wolfspraulwell 'dev' may be the wrong word. the server admin takes up an awful amount of time.03:41
wolfspraulI just setup a buildhost so hopefully we will get better and quicker with OpenWrt images.03:42
wolfspraul(building right now :-))03:42
elricsfateI have an idea03:42
elricsfatenot a unique idea03:42
elricsfatebut something practical03:43
elricsfategetting some sort of install manager done for the end user if that hasn't been done already03:43
elricsfatepossibly something graphical that takes the small resolution into mind03:43
wolfspraulgood idea although I think we are just not there yet03:46
wolfspraulapt-get in Debian and opkg in OpenWrt/Jlime...03:47
elricsfatethose are both non graphical are they not? I have something in mind similar to (and I hate to say this) the Apple Store03:49
wolfspraulthat's OK, but we can only go step by step by what is possible for us today03:51
wolfspraulI'm not aware of a GUI installer for qvga resolution03:52
elricsfatewolfspraul: Of course. One should have a vision of the vision though shouldn't they?03:52
elricsfatewow typing fail XD03:52
elricsfate*vision of the future03:52
elricsfateperhaps its time for me to go to bed03:52
elricsfatewoflspraul: One last question. How many days does it normally take to arrive?03:56
wolfspraulelricsfate: where did you order and where do you live?04:16
wolfspraulmostly it's 'a few days'04:16
wpwrakelricsfate: i like the concept of "a vision of a vision" ;-)04:33
kristianpaulyes about raid1 when dont need mirror something wich no backup is planned ;)07:21
kristianpaulfree it we will get twice space i guess :D07:22
kristianpaulwolfspraul: you changed something in fidelio? i got @    WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!     @07:29
kristianpaulwolfspraul: yes raid1 just will low performance and also we dont need backup07:30
wolfspraulkristianpaul: yes, I did a complete new setup :-)07:30
kristianpauland more space free i may guess07:30
kristianpaulwolfspraul: oh riht07:30
wolfsprauldidn't bother to copy the server keys07:30
kristianpauli noticed ;)07:31
wolfspraulit's raid0 now, and ext4 (journal disabled) instead of ext2, and some other performance tweaks07:31
kristianpaulgood :)07:31
kristianpaulYou really make worth to the "is sanbox server, explore !"07:32
wolfspraulwell I won't just nuke it again, only now because I realized raid0 is so much better than raid1 in this case07:32
wolfspraulI could have preserved the data on it, but since it was only one day old I just started over07:33
wolfspraulnow it's good to go!07:34
wolfspraulI am building already, and will start some more builds07:34
wolfspraulwill try to make all packages too07:34
wolfspraulsee how long that takes07:34
Action: kristianpaul 07:43
kristianpaulgot a go, byw07:43
qi-bot[commit] Werner Almesberger: atusd/atusd.brd: bumped the version to 100908 http://qi-hw.com/p/ben-wpan/284b55708:25
qi-bot[commit] Werner Almesberger: Updated atusd driver for new hardware. Make use of the interrupt line. http://qi-hw.com/p/ben-wpan/18eec5508:25
qi-bot[commit] Werner Almesberger: New atusd errata and a few small CAM updates. http://qi-hw.com/p/ben-wpan/93f0f4008:25
wolfspraulwpwrak: I'm thinking about the practical aspects of the breakout sdio cable more. What's so bad about the ribbon cable?08:28
wolfspraulit looks like the ribbon cable is stronger/more versatile than an FPC cable08:29
wolfspraulis the signal integrity in a ribbon cable worse?08:29
wpwrakwolfspraul: it's probably a bit more difficult to manufacture. FPC may also be more stable where soldered.08:29
wolfspraulit's probably more expensive, yes. but my feeling is the mechanical stability is much better.08:30
wpwraksignal integrity should be similar (if you use the same spacing)08:30
wolfspraulmaybe even better if the ribbon cables are individually shielded?08:31
wpwrakfor better mechanical stability, you'd also have to attach the isolated part of the cable somehow. glue or such. that may get tricky.08:31
wpwrakhmm, let's see what shielded ribbon cable costs you ...08:32
wolfspraulyou mean the isolation around the individual wires needs to be attached to the pcb?08:32
wolfspraulin terms of soldering, the wires from inside the ribbon cables would be soldered to the PCB as well, don't know why that should be less stable than an FPC?08:33
wpwrakyes. otherwise you pull on the solder joints. since they'll have to be soldered manually, they're likely to experience an uneven load. also, ribbon cable doesn't take too frequent bending lightly.08:33
wolfspraulI need to see what standard techniques/processes exist to attach ribbon cables to a PCB08:33
wpwrakthe normal approach is to use a connector08:34
wolfspraulwait a sec loosing you08:34
wolfspraul"you pull on the solder joints"08:34
wpwrakyou can of course do this. just adds some overhead08:34
wolfspraulwho pulls on what?///08:34
wolfsprauluneven load?08:34
wolfspraulthe ribbon cable itself is a lot more stable than the FPC08:34
wolfspraulwe agree on that I think08:34
wpwrakwhen there mechanical stress on the cable, then all that goes to the solder joints08:35
wolfspraulin terms of shielding, the ribbon cable will be as good or better than the fpc?08:35
wolfspraulyes but same for an FPC, no?08:35
wolfspraulof course we could go pro and have a flex-rigid PCB solution :-)08:35
wpwrakin the case of fpc, the cable lies flat, so there's less of a lever. also, you probably have a larger contact surface and a more uniform load distribution08:36
wolfspraulwhy? I an expose the wires and solder it as strongly to the PCB as I can with an FPC08:36
wpwrakribbon cable and fpc will typically be unshielded :)08:36
wolfspraulhmm. many fpcs nowadays are shielded.08:36
wolfspraulsome silver stuff I think08:36
wolfspraulbut it's only on top, so I think there is nothing _between_ the wires (sideways)08:37
wolfspraul(in a shielded fpc)08:37
wolfspraulwould you prefer a shielded ribbon cable? (where each wire is shielded) does that help?08:37
wpwrakjust look at how thin the wires are :) that 300 V+ isolation does have its cost in terms of wasted space.08:38
wpwraki'm not sure you can even get an individually shielded ribbon cable08:39
wpwrakthere are some with a shield around everything, though08:39
wpwrakwhat you can get is twisted pair, or put ground wires between signals. the latter is what IDE (PATA) does.08:40
wolfspraulok one by one08:40
wolfspraulfirst I try to understand the pros and cons of ribbon vs. fpc08:40
wolfspraulthen - if it's fpc, whether fpc shielding (coating on top and bottom) is worth it08:40
wolfsprauland, if it's ribbon, whether ribbon shielding (either around the entire thing or individually or with GND wires) is worth it08:41
wolfspraulthe thing is when I look at Ornotermes pictures, the ribbon cable looks perfect to me :-)08:41
wolfspraulfpc? too breakable!08:41
wpwrakyou probably don't need a lot of shielding. this is for fairly coarse experiments anyway.08:42
wolfspraulshielding is my second question, what about the first one - fpc vs ribbon08:43
wpwrakfpc have their failure modes too, agreed.08:43
wolfspraulI don't understand the fpc advantages08:43
wpwraki think soldering a ribbon cable directly on a PCB is a non-standard process08:44
wolfspraulyes that I have to investigate :-)08:44
wpwrakif you can find a way to get this done properly, including some pull relief, a ribbon cable is okay08:45
wolfspraulwith an fpc, you can also run the fpc all the way into the microSD connector, and only underlay it with a piece of plastic in the last part (inside the connector), to reach the necessary thickness08:46
wolfspraulso basically no PCB inside the connector at all, just run the fpc in there08:46
wpwrakyes, good idea08:47
wolfspraulactually I've made some of these last year and gave to some people08:47
wpwrakjust need to get some good glue ;-)08:47
wolfsprauland the idea was stolen from???08:47
wolfspraulmaybe you remember that you had the idea of a 'breakout fpc' at our fantastic previous employer08:47
wpwrakyes, that i remember08:48
wolfspraulI will take a picture of the one I have here tomorrow (daylight)08:48
wolfsprauldidn't I include one in the last package to you?08:48
wpwrakah, there was one odd little cable ...08:48
wpwrakfull-size SD or such, if i remember right (?)08:49
wolfspraulmicro to full, yes08:49
wpwrakhaving a look ...08:49
wolfspraulI have several types, some are PCB based, some are FPC based08:49
wolfspraulDavid is using the PCB based one in this video: http://en.qi-hardware.com/wiki/File:Leds.ogv08:49
wolfspraulif you have the FPC-based one, then you can see how it's done inside the connector08:50
Ornotermesthe biggest chance i failure i suspect so far is bending the cable near she solder points where the tin stops supporting the strands in the wire, but that is solved by hot glue08:50
wolfspraulmaybe I'm too worried and actually FPC is the best08:51
wpwraki have the fpc version. uSD to SD. yes, looks pretty tidy.08:51
wolfspraulmaybe even with some shielding08:51
wolfspraulwpwrak: yes I think it's neat. only needs to end in those metal heads, 2.54mm apart08:52
wpwraksome beads may do more for signal integrity than shielding. but with adam, you already have an expert on such things :)08:53
wpwrakhehe, a cable for metal heads. tuxbrain will like that :)08:53
wolfspraulwhat's the right term?08:53
wpwrakthey're called headers08:54
wolfspraulyeah, so basically the cable you already have, just longer and ending in 8 headers 2.54mm apart08:55
wolfspraulI still like the flexibility of the ribbon cable, oh well... :-)08:57
wpwrakyup. btw, the whole thing with N pins is called "header" (singular). they also exist in an SMT variant08:58
wolfspraulif I make the FPC 20 cm long, it won't fit in regular letters, so it has to be bent around. not good.09:00
wolfspraulif I try to keep the total size under 10cm, it looses functionality09:00
wpwrakuse longer enveloppes ;) i think a total length of ~20 cm should be okay, no ?09:01
wpwrakso maybe you get 15 cm for the fpc, 5 for uSD and connectors09:02
wolfspraulI would definitely want the header to be flat, pointing in the same direction as the cable is running. again for shippability, just put into a letter...09:02
wpwrakare 4 mm flat enough ?09:03
wolfsprauldon't know, but are you trying to have the header point up (90o angle), or point outward (flat)09:05
wpwrakah, would you solder the header directly on the FPC or go via a small PCB ?09:05
wolfsprauldirectly onto the fpc if possible09:05
wolfsprauljust look at that fpc cable/adapter you already have09:05
wpwraki was thinking to make the coplanar with the FPC. although this introduces some bending09:05
wolfspraulkristianpaul: config.full_system build took 7.5h now, down from 9.x yesterday09:05
wolfspraulcoplanar = flat?09:06
wolfspraulwhat bending?09:06
wolfsprauloh you mean if the breadboard sits next to the ben, for example, the fpc has to run in a curve09:06
wolfspraulbut I think there are many setups we cannot predict the exact angles outside the device anyway09:07
wpwrak(curve) yes, that's what i was thinking about. connecting from the top is the most typical use for these things.09:09
wpwrakoh, you could of course leave the contacts open. no header. that way, people can solder whatever they like to have there09:09
wpwrakalso solves the shipping problem ;-)))09:09
wolfspraulyes but it's harder, and desoldering is always easy anyway09:10
wolfspraulso I'd rather get some work done...09:10
wpwrak(whatever they like) e.g., they may want to have a female connector there so that they can use the (cheaper) male header on their board09:10
wpwrakor they may want to have a polarized connector, again same spacing and everything. there's a number of systems that are all compatible with the plain 100 mil header.09:11
wolfspraulsure but that is true anyway, and you can always desolder the connector that's already there09:11
wpwrakif it's not glued, too.09:11
wpwrakyou probably need glue, for stability09:11
wolfspraulI think a simple male 2.54mm header to get started is fine09:11
wolfsprauldon't know. I can put a plastic underneath the end of the fpc, just like on the other (microsd) side09:12
wpwrakyeah. the plastic is pretty strong.09:13
wpwraksound good then. you just have to pick the sort of header you like. an smt variant will be a bit thicker but more standard to solder.09:13
wpwrakif you just solder a throuh-hole variant on the edge, it will be thinner but i don't know how they'll like it from the manufacturing side. well, since it's small numbers, manual soldering should be okay.09:14
wolfspraulregular letter maximum width: German 23.5 cm, Hong Kong 26 cm, USA 29.2 cm09:15
wolfspraulyeah so if the total is 20 cm or even 22 cm, no need to fold it09:16
wpwrakperfect :)09:16
wolfsprauldo you think shielding the FPC is a good idea?09:17
wpwraki don't think you need a shield that parallels the traces. not sure about EMI and cross-talk, though. the latter may benefit from a signal-GND pairing. uSD already does a little of this by sandwiching the bus clock between VDD and GND. but CMD and the data lines run openly.09:19
wolfspraulso silver coating the fpc on top and bottom is not worth it?09:20
wpwrakit would seem excessive to me09:20
wolfspraulshould we make a much shorter variant too? or one that basically just exposes the 8 wires where they come out of the connector?09:23
wpwrakthe latter could be useful too. 20 cm may be a bit long but the minimum useful size is probably a bit above 10 cm, so i'm not sure there's much room for a smaller variant.09:25
wolfspraulwhat do you mean with 'much room'? you mean no applications?09:26
wolfspraulwhat if someone wants to solder something more high speed, and the wires just need to be short?09:26
wolfspraulof course you may be able to cut open the fpc and get to the wires that way09:27
wpwraka "just expose the 8 wires" variant could help here.09:27
wpwrakanyone who's planning something a bit more sophisticated can just make their own pcb anyway09:28
wolfspraulI think there are lots of small EVBs nowadays09:28
wolfspraulsome in full-size SD form, that's one reason why I made this adapter you have09:29
wolfspraulbut some just a small PCB, say 2x2cm, and you need to find out for the first prototyping how to hook that up (with several options offered by the evb)09:30
wolfspraulI think in many cases, making another PCB is quite a bit of work, so if we can make some and 'bridge' a few cases here and there later that's cool I think09:30
wpwraka "just bring out the wires" adapter may help with such things. it eliminates the finicky mechanical work and you'll need some evb-specific cable anyway.09:31
wolfspraultrying to find the 'one size fits all' solution :-)09:33
wpwrakah, maybe adding an inductor and/or a polyfuse may also be a nice touch :)09:34
wolfspraulwhere and how?09:36
wpwraksolder them on the fpc, inrush current and short circuit protection09:37
wolfspraulmaybe before the male header?09:40
wpwrakwe also need to specify how much power is actually available09:40
wolfspraulwhere would they sit? (on which wires)09:41
wpwrakjust on VDD09:41
wpwrakthere's of course also the issue of ESD protection. there's none in the Ben, except whatever the CPU may be able to take.09:42
kyakecho 665 > build_dir/linux-xburst_qi_lb60/linux- // :D09:43
wolfspraulwpwrak: what type of inductor and fuse do you have in mind?09:59
wpwrakinductor maybe around 1 uH. fuse, a polyfuse.10:00
wpwraklike this http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=F2901CT-ND10:00
wpwrakunfortunately, they're not cheap10:00
wpwrakcurrent rating would depend on how much we're actually allowed to draw10:00
wolfspraulok I'll go from there, thanks!10:07
wpwrakgood luck ! :)10:09
bartbesurandom__: ah you're finally here again11:15
bartbes(while I am)11:15
bartbesso, did you ever get around to creating a more advanced port of the snake game?11:15
qi-bot[commit] Werner Almesberger: Improved clock stability by using a capacitative divider and found more minor http://qi-hw.com/p/ben-wpan/6e726d111:44
kristianpaulwpwrak: what the 'a' mean when you said 10 MSa/s ?12:55
kristianpauli undertand is likle said 10 millions of samples per second..12:56
wpwrakSa = Sample12:56
wpwrakto avoid confusion with S = Siemens :-)12:56
kristianpaulbut even TI uses MSPS12:56
kristianpaulSDL hackers around?12:57
wpwrakyeah, Siemens isn't the most commonly used unit. i think i've seen it once or twice outside textbooks ;-)12:57
kristianpaulsure and in germany i guess ;)12:57
wpwrakit's an official SI unit :) but it gets better: http://en.wikipedia.org/wiki/Siemens_(unit)#Mho13:00
wpwrakand there are whose people who already find it difficult to place regular greek letters into text or schematics. little do they know ;-)13:01
kristianpaulwpwrak: btw how is your wpan going?13:01
kristianpaulthere is a list called linux en caja, i pointed a guy to yout project13:01
wpwrakpretty good. i made a new set up boards that don't have the multiplexing bug. one has trouble with clock stability, so i changed the resistive voltage divider (the clock input can only be driven with a signal of 400-500 mV peak-to-peak)13:02
kristianpaulwhat,  inverted Omega? grazy people13:02
mthkristianpaul: as in libsdl? I've done some programming with it13:03
bartbesurandom__: you here yet?13:03
kristianpaulmth: hello13:03
kristianpaulmth: do you have a hello world example of basic plotting with sdl, lile sin or cos functions?13:03
wpwrakto a capacitative divider in the second board. that seems to work well. clock is stable and i can send and receive. not quite sure what the capacitative divider does to EMC, though. I guess i should find some radio that listens in the low harmonics of 16 MHz :-)13:04
kristianpaulgreat wpwrak :)13:04
mthnot really, but I guess there are tutorial around13:04
mthI did write this yesterday: http://dingoowiki.com/index.php/Dingux:OpenDingux#Development13:04
mthsee the fragment at the end for a quick overview of how to push pixels13:05
wpwrakkristianpaul: (uinverted omega) and then there's the beautiful programming language called APL. not satisfied with the regular characters set, extended with greek and some other weird symbols, they invented the overprinting of characters. one can't say that APL programs aren't *dense*, though. Perl looks rather chatty and trivial in comparison.13:06
kristianpaulmth: what programinhg you did before?13:06
mthfor NN it would have to be 32bpp instead of 16bpp13:06
mthmainly on openMSX (MSX emulator)13:06
wpwrak(wpan) now i need to see how far these critters can go. i hope they're doing better than their USB predecessors. (not because of USB vs. uSD but because i also improved a number of things in the RF design)13:07
kristianpaulwpwrak: i think will look nice a wap with uSD and a little case for it :)13:07
wpwrakthe case would be the next challenge indeed :)13:08
wpwrakwell, it's not so small. the antenna is big.13:08
wpwraklemma take a picture ...13:08
kristianpaullovelly !13:18
wpwraki wish i could make it a bit narrower13:19
kristianpaulsell me one !13:20
wpwrakheh, i'll let wolfgang and tuxbrain figure out how to produce those beasts ;-)13:21
kristianpauland the case13:21
kristianpaulwell is small13:21
kristianpauli can print it at home :)13:21
kristianpaulhmm i have ABS in colors13:22
wpwrakah yes, you can make a case :)13:22
kristianpaulwill be cute yellow13:22
kristianpaulor strong red13:22
kristianpaulgreat work wpwrak :)13:24
wpwrakyou can always paint it black to follow the style of the rest of the ben :)13:25
kristianpaulhehe yesi just dont have black ABS...13:25
kristianpaulmay be blueABS13:25
wpwrakthanks :) but it's not done yet. still need to see if i get any kind of range. right now, my test consist of transmission between antennas about 20 cm apart :)13:25
kristianpaulwpwrak: i guy asked me if posible to send encrypted data with the current bandwich13:27
kristianpaulwhat do you think?13:27
kristianpaul(talking about wpan)13:27
wpwrakof course you can encrypt. doesn't have much to do with bandwidth. the cpu should be more than fast enough13:28
kristianpaulgreat :)13:28
kristianpauli now just wait you finish to start selling devices with build-in ecrypted comunication :D13:28
wpwrakthe transceiver doesn't have hardware-accelerated encryption. so the cpu has to do the work. this saves us from us export restrictions for the chip :)13:29
kristianpaulbtw is this first aprouch to wifi freedom but i guess next move is something more SDR that SCR ?13:29
kristianpaulSCR (software controlled radio)13:30
wpwrakSDR will be a bit of a challenge :)13:30
kristianpaultell me to me13:30
kristianpaulGPS-SDR is really challging all this correlations are heavy for 336Mhz ben...13:31
kristianpaulwell you have USRP, soway to go :)13:32
kristianpauljust an ADC DCA and the nanonote :p13:32
wpwrakheh, connect the ben to the usrp ;-)13:34
kristianpaulyou tried?13:35
kristianpaulno kidding13:35
kristianpaulit will fry with that bandwich13:35
wpwraknaw. i already know that the usrp2 needs a little more to be happy than what the ben can deliver :)13:35
kristianpaulwpwrak: whats the current cost for awpan board?13:36
wpwrakyou mean the BOM cost ? depends a lot on the quantity at which you buy components.13:43
wpwrakright now, you need an Atmel AT86RF230, a Wuerth 748421245 balun, 2 good 22 pF capacitors, 4 x 1 uF cap, 1 x 220 pF, 1 x 33 pF, and 2 0R resistors. all the small stuff is 0402.13:45
wpwrakone of the 0R should probably become an inductor13:45
urandom__bartbes i am here14:11
bartbesso, have you tried the new build of nlove (0.2, from.. 2 days ago, I think)14:12
urandom__i have not, i think14:13
bartbesoh noes14:13
bartbesyou promised me a new version of snake when fonts arrived14:13
urandom__so where can i download the latest build?14:15
kyakare you guys not satisfied with bsd-games/worm? ;)14:23
urandom__nah the point is to turn the nanonote into a love-maschine kyak14:24
kyakso one could make love with it?14:25
urandom__well you can play love games on it, making love might be .. slow14:27
kyaknot productive, too14:28
urandom__bartbes next version of snake comes friday or on weekend or when it is done, really have other stuff to do14:29
bartbeskyak: well, it all depends on how fast your fingers are14:29
urandom__i am wodering if i should make the snake smaller so you can grow it longer14:30
urandom__but the current size feels very retro, i think14:31
urandom__nah i will offer different modes14:34
bartbesI should write a game too14:37
bartbesmaybe tomorrow night14:37
urandom__you could port the sea otters but that needs sound first14:38
bartbesthe sea otters have been done14:40
bartbesoh wait, I can re-add the text now14:40
bartbesoh god no14:43
bartbesI modified it, on /tmp14:43
bartbesand now it crashed14:43
bartbesor.. maybe it didn't14:44
bartbeswhat the hell just happened?14:45
bartbesI think love segfaulted14:45
bartbesit did14:46
qi-bot[commit] Bas Wijnen: mass storage nand almost working http://qi-hw.com/p/iris/87acc4715:50
qi-bot[commit] Werner Almesberger: Cleaned up interrupt handling in atspi-txrx. Report ED. Shut down at end. http://qi-hw.com/p/ben-wpan/916268916:57
qi-bot[commit] Werner Almesberger: tools/Makefile.common: added quiet compilation http://qi-hw.com/p/ben-wpan/be6241616:57
qi-bot[commit] Werner Almesberger: Make wait_for_interrupt available to all atspi tools. http://qi-hw.com/p/ben-wpan/2a245e116:57
qi-bot[commit] Werner Almesberger: atspi-rssi cleanup: wait for interrupt, exit cleanly. http://qi-hw.com/p/ben-wpan/4df6d2c16:57
jegccan someone give me information about http://projects.qi-hardware.com/index.php/p/ben-wpan/17:01
kristianpaulwpwrak: ^17:25
wpwrakoh ah ...17:28
wpwrakjegc: i probably can :)17:28
jegcwpwrak, hi :) thanks17:28
wpwrakjegc: so what would you like to know ?17:30
jegcwpwrak, i like nkow how much is the cost of elaboration of a ben-wpan17:30
wpwrakheh, we had this a moment ago :) lemme copy the answer ...17:31
wpwrakyou need an Atmel AT86RF230, a Wuerth 748421245 balun, 2 good 22 pF capacitors, 4 x 1 uF cap, 1 x 220 pF, 1 x 33 pF, and 2 0R resistors. all the small stuff is 0402.17:31
wpwrakthat's the BOM. prices depend a lot on how many units you make.17:32
wpwrakyou also need the PCB (0.8 mm thickness)17:32
kristianpaulwpwrak: yes i was setting up you for this conversation :)17:33
wpwrakfor small quantities (< 25 units), the material cost should be around USD 8.17:34
jegcwpwrak, ok, thanks again17:35
wpwrakthat is, unless i find a bug that needs expensive fixing :) still haven't done much analyzing of the RF characteristics17:35
wpwrake.g., there's a 0R resistor that could become an inductor in case we need some noise suppression17:36
jegcfor my need the cost is in the budget17:37
kristianpauljegc: me curious, you got Ben?17:37
kristianpauljust sell that frerunner and buy one !17:37
kristianpaulat least you are _raelly_ using it17:38
wpwrakit's pretty cheap. you could make it a bit cheaper by replacing the balun, but i won't touch that before i'm a bit more confident about playing with RF.17:38
jegckristianpaul, no, but if the project is approved, would buy a few17:39
kristianpaulgood :)17:39
wpwrakwolfgang will be happy :)17:39
jegckristianpaul, for now is a preproject...17:39
kristianpauljegc: good to know you care about Ben :)17:41
jegckristianpaul, shure, i like one, but i wait while get money :( or the project run ;)17:44
jegcwpwrak, kristianpaul, thanks for the help, now i will go out17:56
qi-bot[commit] Juan64Bits: Kicad labels problem. http://qi-hw.com/p/xue/1cd37e718:58
qi-bot[commit] Werner Almesberger: Added interrupt polling support for atusb. http://qi-hw.com/p/ben-wpan/b6a80ba19:42
qi-bot[commit] Werner Almesberger: Performance comparison of atusb vs. atusd. http://qi-hw.com/p/ben-wpan/0a0a50d19:42
qi-bot[commit] Werner Almesberger: Oops. The atusb boards are of course 20100813. 20100902 was the first http://qi-hw.com/p/ben-wpan/b32827d19:46
qi-bot[commit] Werner Almesberger: Moved the ECNs from atrf/ecn to the top-level. http://qi-hw.com/p/ben-wpan/d3bc27423:11
qi-bot[commit] Werner Almesberger: New ECN about clock circuit of atusd boards. Measured (absence of) effect of http://qi-hw.com/p/ben-wpan/df071c323:11
--- Fri Sep 10 201000:00

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