#qi-hardware IRC log for Monday, 2011-01-31

qi-bot[commit] Wolfgang Spraul: improved label http://qi-hw.com/p/m1/60127c601:00
qi-bot[commit] Werner Almesberger: atusb/cam/mkmk: updated for latest board parameters http://qi-hw.com/p/ben-wpan/e6b0ee503:04
qi-bot[commit] Werner Almesberger: atben.sch: corrected typo "sLP_TR" http://qi-hw.com/p/ben-wpan/14b008203:04
qi-bot[commit] Werner Almesberger: Added ATmega8U2/16U2/32U2 schematics symbol. http://qi-hw.com/p/ben-wpan/5b9263b03:04
qi-bot[commit] Werner Almesberger: atusb: replaced the C8051F326 with an ATmega32U2 (still needs cleanup) http://qi-hw.com/p/ben-wpan/6aa4e3003:04
qi-bot[commit] Werner Almesberger: atben/cam/mkmk: updated for latest board parameters http://qi-hw.com/p/ben-wpan/ded90ef03:04
qi-bot[commit] Werner Almesberger: atusb: cleaned up USB schematics http://qi-hw.com/p/ben-wpan/318f29c03:04
qi-bot[commit] Werner Almesberger: atusb: new layout for the ATmega32U2 http://qi-hw.com/p/ben-wpan/d8dfb5003:04
qi-bot[commit] Werner Almesberger: usb.sch: corrected symbol name http://qi-hw.com/p/ben-wpan/ebaffb503:10
B_Lizzardlarsc, the patch you sent me gives me an undefined reference to "usb_gadget_register_driver" during linking04:38
B_LizzardI noticed that the udc patch is different from this one https://dev.openwrt.org/browser/trunk/target/linux/xburst/patches-2.6.36/200-udc.patch04:39
rohhm.. hackabledevices down?04:43
B_LizzardI see the two files have a few differences04:43
B_LizzardI'll just rename the usb_gadget_probe_driver to usb_gadget_register_driver, see if it explodes04:46
B_LizzardWhoa, it doesn't build, conflicting types04:53
B_LizzardHow surprising04:53
wolfspraulroh: I saw your disgust when you had to enter a password into projects.qi-hardware.com without ssl in Berlin, and it made me think...05:02
wolfspraulso now I plan to move more stuff behind ssl by default, maybe I will experiment with projects.qi-hardware.com because there is not that much traffic there05:02
wolfspraulmy plan is to move the entire site behind https, and redirect all http urls to the corresponding https url. Is this a reasonable plan?05:03
wolfspraulyou have more ssl experience - any caveats?05:03
wolfsprauldo I need to be worried about server load?05:03
rohwolfspraul: atleast login should be behind ssl.05:16
wolfspraulsure but I'm thinking about moving all behind ssl05:17
rohyou know the firefox plugin which sniffs wifi and passwords?05:19
rohand enables you to login as people using no ssl around you ;)05:19
rohhttp://en.wikipedia.org/wiki/Firesheep05:20
wolfspraulno, but sorry I lost you. What do you mean?05:20
wolfspraulah (reading)05:20
rohthats stuff what happens when passwords are in the clear ;)05:20
rejonwolfspraul i think that is a good idea05:43
wolfspraulrejon: roh is from raumfahrtagentur.org and made the m1 case05:44
rejoncool roh are you in berlin?05:44
rejoncurious to meet up during transmediale week05:45
syeeborghey05:45
rohrejon: yes05:45
rohsure. when will that be?05:45
rejonits this week05:45
rohi see05:46
rejoni am working at betahaus.de all week05:46
rohheh05:46
rejonhi syeeborg05:47
rejonhi gbraad05:47
syeeborgdoes anyone know if tripple channel ram will run in an amd 3 board?05:49
rohi never visited betahaus... too much co-working (table, ip, coffee) and no enough hacking (tools, machines) for me05:49
syeeborgi know triple channel is optimized for intel based machines05:49
rohsyeeborg: no clue. check the compatibility matrix of your vendors.05:50
rohram is ucky. sometimes it doesnt even work when the specs say it should.05:51
syeeborgahh ok05:52
lekernelroh: I visited them a while ago. they do have some machines, but I haven't seen many materialized results yet.05:52
rejonlekernel roh right...05:53
rejonjust working upstairs my moz friends gave some space for the week05:53
rejonnot hardware hacking05:54
rejonbut its a productive space for computer time05:54
syeeborgis it worth putting win 7 64 bit on my new machine05:54
rejonand good food for us travelers05:54
syeeborgor shud i stick to win 7 32bit05:54
rejonroh anyway, it would be good to meetup05:54
rejonwhere do you work out of?05:54
lekernelsyeeborg: wrong channel05:54
syeeborgits got 8 gb off ram and a 6 core processor,05:54
rohsyeeborg: you are wrong here.05:54
rohthis is not computer 101 help channel.05:54
syeeborgoh what ist05:54
syeeborgis05:54
rohrejon: raumfahrtagentur05:54
rohrejon: prenzlauer berg, near alexanderplatz05:55
rejonok, cool06:00
rejonyou stay late?06:00
rejontrying to work full days06:01
rohdepends.. best is to call ahead or catch me here or in jabber06:01
rohwe dont have any regular times at the hackspace06:01
rejonok cool06:01
rejoncool i'm +491515531842706:01
rejonjon phillips -> http://rejon.org06:01
rohits funny to see all the opensource and hw guys tend to drift towards berlin lately :)06:03
rohbtw: something i realized only last winter06:10
rohi havent seen any hackspace i liked which didnt have a madly insane overspec-ed soundsystem.06:11
rohmusic is important for creativity and getting rid of disturbances by reality06:11
wolfspraultotally agree06:16
rohthats also something i would have lots of fun with. working on foss consumer electronics06:16
rohlike some ac3/dts decoder/amp/ thingie like there are hundreds.. just all closed and not intended for 'home integration'06:17
rohthere is some hw, but most of it is dongled up with bad sw like the roku stuff06:18
rohsame goes for all the 'hd media players' which are basically just dvd/blueray players with ethernet and usb/hdd instead of rotating disks06:19
roh .oO(anybody got 10-20million euro to play with?)06:20
rejonroh you need something like http://laoban-soundsystem.com/06:31
rejonor more like this with the faces off: http://www.flickr.com/photos/jonphillips/5345510732/in/photostream/06:32
rohheh. need more space first06:32
rohrejon: friends of mine built this: http://gregoa.de/wp/klondike-one/06:35
rejonthat's great06:36
rejonyeah, we need to send that big container soundsystem somewhere next06:36
rohif you want to see it used.. try festivals in the summer06:37
rohlike fusion or so06:37
rohthese guys http://www.kulturkosmos.de/06:37
rohthey have the space and the party06:37
rejoncool roh06:49
lekernelroh: (klondike) excellent :)06:53
rjeffriesIt is early morning on the Left Coast of USA.09:24
Action: wpwrak needs to write a bot that announces hourly the local time, wind speed, the temperature of my wine, and the amount of meat in my fridge :)09:46
rjeffriesvery funny. what time is it now? My guess: just before NOON on Monday??09:49
rjeffrieshttp://wwp.greenwichmeantime.com/time-zone/south-america/argentina/buenos-aires/time.htm09:49
wpwrakyour kworldclock seems to be working well :-)09:49
wpwrakah, or that :)09:49
rjeffriesis all of Argintina in same time zone09:50
wpwrakat the moment, yes. they the government gets it into their heads to introduce DST, things get a little complicated, with provinces changing their time zone at a daily basis09:56
wpwrakwell, this year, they didn't try DST. i think a lesson has been learned :) hopefully, this will put the issue to rest for a while09:57
qi-bot[commit] Werner Almesberger: atusb.brd: ground zone adjustments; corrected date code location http://qi-hw.com/p/ben-wpan/8e842ba10:08
qi-bot[commit] Werner Almesberger: usb.sch: cleanup and track layout http://qi-hw.com/p/ben-wpan/847b35410:08
larschm, only -5h explains why you are here most of the time when i'm here. especially given that my day starts 5 hours late ;)10:08
kristianpaullarsc: whats your GTM? +5?10:09
larscno, gmt+210:09
kristianpaulhere gmt-510:11
larscbut my day usually goes from 12pm-4am10:11
rohlarsc: *g*10:13
Action: roh has so many timezones in his gnome applet that the list doesnt show all10:14
rohscreen too small10:14
kristianpaulroh: he, same here ;-)10:15
kristianpaul8.9" laptop for work is too small, but damn light when i must travel around the city :-)10:16
kristianpaullarsc: do you have a list of apps that require MMU or way of notice that,may be grepping the source code?10:17
kristianpaulat least wich apps from the gnu familly10:17
larscnope10:18
kristianpaulk..10:18
wpwrakkristianpaul: it's hard to tell which applications really _require_ an mmu. i can't think of any library function that usually assumes an mmu that couldn't be rewritten to work without one.10:19
wpwrakkristianpaul: but this may result in excessive resource use10:19
larscthere might be some applications that take it for granted that they can mmap a file into their address space and have no open()/read()/write() fallback10:21
larscand there is no fork10:24
larsc()10:24
wpwraklarsc: yes, you would have to hack this in libc10:24
wpwraklarsc: (fork) just copy the process :)10:26
larsconce we have basic uclibc support for lm32, i'll add it to openwrt and make a full build.10:32
kristianpaullarsc: there are arch in owrt with support for no-mmu isnt?10:34
kristianpaulalso i guess busybox can handle that..10:34
larsckristianpaul: not sure10:36
larscabout whether there are other no-mmu archs in openwrt10:36
larscbut i guess busybox is written in a way that it can run on a mmu system10:36
larscalthough it's quite imperformant10:37
kristianpaul:/10:37
mthbusybox has a CONFIG_NOMMU option10:38
mthI never used it though10:38
dvdkhlngmay i throw in some opinion about nommu :10:56
dvdkhlngcauses lots of trouble with memory fragmentation10:56
dvdkhlngif an app wants to malloc() some 16MB of memory, and there is no contiguous chunk of that size, you're practically out  of memory.10:57
dvdkhlngi think that being able to alloc block-wise memory without fragmentation issues is much more important than the memory-protection that an MMU gives you.10:59
mth"any problem can be solved by adding a level of indirection"?11:06
dvdkhlng:)11:16
dvdkhlngBTW if we already replaced the FPU with gcc's softfp, why not hack gcc to generate 'softmmu' code for all memory references :)11:18
dvdkhlngi.e. MMU emulation11:18
mthis that really easier than adding an MMU implementation in the FPGA?11:22
dvdkhlngi was kidding (somewhat).  but note how transmeta crusoe/efficeon used software-based mmu on their x86-emulating cpus.11:24
rjeffriesLeft Coast USA during daylight savins (now) is GMT -7. I wish IRC nicknames carried an easy to finf associated GMT delta11:26
rjeffrieskristianpaul are you in same city as wpwrak a.k.a. Werner11:27
rjeffries08:23 <dvdkhlng> soft MMU is the perfect way to ensure MM pSOC runs as slow as posible, if not slower. <just kidding, but true statement>11:29
rjeffries08:23 <dvdkhlng> //you make a good point. maybe soft MMU is not a brain dead idea after all. Maybe I am breain dead. It happens.11:37
kristianpaulrjeffries: no11:38
rjeffriesmaybe I should go to sleep, wake up in 10 years to enjoy the MM SOC as asic I am just too damned impatient11:39
rjeffriesbassel what is your GMT delta? I am making a table I'll share later12:00
basselrjeffries GMT + @12:00
basselGMT + 212:00
wpwrakrjeffries: (irc and gmt) maybe they just realized that a lot of people aren't synchronized with the position of the sun :)12:07
larscB_Lizzard: you asked for the patch for 2.6.3712:08
kristianpaulrjeffries: gmt -5, buga, colombia :-)12:08
rjeffriesworks for me. time differnces make irc logs invaluable.12:09
rjeffrieshttp://en.qi-hardware.com/irclogs/latest.log.html12:09
B_Lizzardlarsc, 2.6.36 but it's fine, I fixed it12:09
B_Lizzardit works12:09
larscok12:10
dvdkhlngofftopic question wrt IRC (trying to catch up, didn't use IRC before): comments sent by rjeffries to me "08:23 <dvdkhlng>...)" look like being send via /msg; however they appear in the irc log.  answered via "/msg rjeffries ..." but result didn't show in irclog.  something wrong on my side?12:11
rjeffries<kristianpaul> in an earlier life I visited Bogotá on business to deliver technical prentation for a Swedish company12:12
larscdvdkhlng: they weren't private messages12:13
kristianpaulcold city :-)12:13
kristianpaulhere is warm and windy12:13
rjeffriesbut I worked form California for E12:13
wpwrakrjeffries: btw, thanks for the input on the "MADE IN". let's hope this helps to avoid unnecessary troubles.12:14
rjeffries <dvdkhlng> that time stamp happened because I double clicked your comment in the #qi-hardware irc channel. I use smuxi as my irc client. not great, but it works12:14
rjeffrieswpwrak sometimes facts are facts. It's funny, I have no dog in thsi fight, but would rather not see wolf create hassle for his distributors or sharism at work12:16
wpwrakdvdkhlng: it looks as if ron just copied and pasted your (/msg) replies. by the way, welcome to #qi-hardware ! i think now everyone who matters in ben-openwrt-land is here :)12:16
rjeffriess/tsi/this12:16
dvdkhlnglarsc, rjeffries: thanks for the info.  takes time to getting used to irc :)  looks like irc-client on this side (rcirc.el) changes color of comment depending on whether my nick is mentioned.12:16
wpwrakrjeffries: yeah. trying such things with customs is like wrestling with pigs. the pigs enjoy it and all you get is dirty.12:17
rjeffriesoh cool. So that is why I see some lines in RED here also. I thought wprak and others were treating me as a VIP12:17
wpwrakdvdkhlng: many irc clients do this. helps you notice quickly when people talk about you behind your back ;-)12:17
rjeffriescool beans12:18
dvdkhlngwpwrak: hi, werner; yeah, studdying the rcirc.el source concurrently :)12:18
rjeffries<kristianpaul> are you at the university or in industry12:18
wpwrakrjeffries: btw, when addressing someone, it is more common to use the  nick:  form (or some similar, e.g., command or semicolon would look right, too)12:21
rjeffriesI see so not use the <angle brackets>12:21
wpwrakrjeffries: the <nick>  form is usually associated with a message coming from that person. so it using it also when addressing someone can create confusion about whether you're quoting that person12:22
rjeffriesthanks for the useful advice it has been MANY years since I used irc. But it's great in this application.perfect in fact12:22
kristianpaulrjeffries: not university, industry neither :p, i'm just freelance for industry12:23
Action: dvdkhlng needs to disable colors in irc. looking at his screen it's too obvious that this is not source-code and he's not working :)12:23
wpwrak(irc) oh, i love it :) been quite a chat addict a long time ago and then decided just to stay away from that vile stuff, but i rediscovered it with openmoko, and it does really make cooperation more fluid.12:24
rjeffriesin many ways Twitter sorta kinda reinvented IRC, but for humans. but no "rrom" or channels so not good for this use case12:25
rjeffrieskristianpaul so you are what we call a hired gun. the only question would be, van you shoot straight? //smile12:26
rjeffriesso how are these COOL #qi-hardware logs created? I assume there is a bot that grabs stuff, then an offline program does some reformatting. I REALLY like the format12:27
wpwrakrjeffries: qi-bot is wolfgang's never sleeping eye that watches over us :)12:30
larscdvdkhlng: or add colors to your sourcecode ;)12:31
rjeffriesdid he also write the program that formats the logs I would assume?12:31
steve|mrjeffries: Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!12:35
wpwrakrjeffries: i think that's part of this eggdrop thing12:35
wpwraksteve|m: ah, other package then :)12:35
dvdkhlnglarsc: :)12:47
wpwrakroh: an AVR question. atmel traditionally gave the various timer pins names like "OC1A". e.g., here http://www.atmel.com/dyn/resources/prod_documents/2545S.pdf12:54
wpwrakroh: however, with the ATmega*U2, this has changed to "OC.1A" etc.: http://www.atmel.com/dyn/resources/prod_documents/7799S.pdf12:55
rohuh. didnt know that12:55
wpwrakroh: do you know the story behind that dot ? i find it rather troublesome to have two slightly different names for pins of virtually identical function blocks12:55
wpwrakhmm12:56
rohwell.. there are slim differences in different ones all the time for such stuff12:56
rohalso because some have parts of the interfaces twice instead one12:56
wpwrakroh: yeah, but a dot ? that seems like the mother of all gratuitous changes12:57
rohwpwrak: interresting anyhow12:58
rohin the end its just a define anyhow ;)12:58
wpwrakroh: well, i'm at the circuit symbol now. don't like to diverge from the vendor naming (except for inversion) but don't like to diverge from what the (sw) source may use either.12:59
rohwhat are you using? N_foo?13:00
wpwrakroh: for inversion ? nFOO13:00
rohah.13:00
rohi wonder if atmel doesnt have a usb-cpu-wireless in once chip already13:01
rjeffriessteve|m: Thanks for this info re logs irclog2html.py 2.9.2 by Marius Gedminas13:01
wpwrakroh: (usb-cpu-wireless) apparently not. at least not for ieee 802.15.4. haven't looked if they have any other rf.13:02
rjeffriesmaybe if wolfspraul will share source for his bot, I can (with help from someone else) set up my own irc channel for other purposes entirely13:03
rohATmega128RFA113:03
rohhttp://www.atmel.com/dyn/products/devices.asp?category_id=163&family_id=676&subfamily_id=2124&source=left_nav13:03
wpwrakroh: sans usb, no ?13:03
rohcorrect13:04
wpwrakroh: also quite a monster. if you don't need such a big core, you're cheaper off with two chips13:04
rjeffriesroh: do you know approx price for that chip13:04
wpwrakrjeffries: in digi-key we trust (also for pricing ;-)13:05
rjeffriesI have not mastered digikey i find it confusing. so it it $~10 usd or Less?13:05
rohnope13:05
rjeffriesmore?13:05
rjeffries$450USD13:06
rjeffriesSMILES13:06
rjeffriessorry for caps13:06
wpwrakrjeffries: price always depends on quantity. also, you have to consider if this is a chip major distributors stock. if they don't, it's a risk. (e.g., you may find your project delayed by several months while waiting for new parts from the manufacturer)13:07
rjeffriessomeone famous in semiconducter industry once said "all chips will be $5usd exc ept for those that cost less:13:07
wpwrakrjeffries: you really should learn to navigate digi-key. they're great.13:07
wpwrakrjeffries: yeah, buy a few millions and you can negotiate almost any price :)13:07
rjeffrieswpwrak I will I promise13:07
rjeffriesI have a buddy who uses a killer NXP Arm SOC taht costs maybe $5 in modest quantities13:09
wpwraktuxbrain_away: you may get your little uart board in the end. i realized that it's a convenient interim step on the way to making the avr-based atusb work.13:09
rjeffriesonce you design a complex chip, making a ton of them is pretty cheap13:09
rjeffrieswprak: a little uarts board would be goodness on so many levels. That is the fast track into Arduio land and Jee Labs13:10
kristianpaulwoah !13:10
kristianpaulWolfgang you did this http://en.qi-hardware.com/wiki/File%3ARfm12_microsd.jpg ?13:11
wpwrakah, fun chip. 0.45 mm pitch. my smallest so far :)13:20
rjeffrieswpwrak where did you but the micrsod extension cable? http://en.qi-hardware.com/wiki/File:Rfm12_microsd.jpg13:41
rjeffriess/Arduio/Arduino13:42
tuxbrain_awaywpwrak: yahooooooo!!!!13:43
kristianpaulO_o13:43
kristianpaulI think i heard that from here too :-)13:44
wpwrakrjeffries: i didn't make that one13:44
kristianpaultuxbrain_away: btw you should sell Jeenodes as well13:45
Action: rjeffries nods to kristianpaul and tuxbrain_away that would be Most Excellent13:45
kristianpaul3.3V arduino-like board is intersting, even mor if it can talk with your nanonote as well ;-)13:45
rjeffrieskristianpaul that is what gets me excited by Jee Labs, plus he is an above average human being13:46
tuxbrain_awaykristianpaul: is in the Todo list, I have contated them, margins are not very atractives but well , at the end I also sell NN isn't it? :P13:47
kristianpaultuxbrain_away: margins :'))13:47
tuxbrain_awayIf someone achive to comunicate with them previously and post a step by step tutorial task can rise some priory13:49
tuxbrain_awaypriority13:49
tuxbrain_awaywprank are you serious on that arduNN board?13:50
wpwraknow ... what connector on the serial side ? three option: 1) narrow pcb pads for soldering a 50 mil cable directly to the PCB. 2) larger pcb pads to solder a 100 mil header directly to the PCB. 3) a proper 100 mil header on the PCB.13:51
wpwraktuxbrain_away: there be no arduNN. but a uart, yes :)13:51
kristianpaulardu? i heard atmega13:52
tuxbrain_awayheheheeh ok ok :) enough +1 to a proper conector no soldering13:52
wpwraktuxbrain_away: i need something to try the whole avr-with-ben bringup chain, and that one is less overhead than the new atusb.13:52
kristianpaulactually is on stusd sch right now tuxbrain_away13:52
kristianpauls/stusd/atben13:52
wpwrakhmmm. i better terminate that thing ...13:54
tuxbrain_awayI also advocate for expose as many gpios on the chip , as many as  the board admit13:55
wpwraki was actually aiming for "small" :)13:58
wpwrakwell, let's see how the layout goes ...13:58
kristianpauljust let the pads, may be if some one want hack later can solder something on it..14:00
kristianpaulhey,i dint knew it about atmegau8/32/16u2 familly14:00
kristianpaulnice have a 8bit MCU plus USB support, (so i finally can drop the non gcc suported picF18 MCU...)14:02
tuxbrain_awaysorry I have just passed a sleepless night , was ending some task before going to sleep instead of having the space bar tatooed on the face , that news  have inject some adrenaline on the system but just enough to work with a smile on the face and not work with eyes closed. c you14:04
kristianpaulchao14:04
kristianpaulhmm, i think i soon can have enought stuff to order from digikey :-)14:05
rjeffriesThis is interesting. Looking at you, tuxbrain: http://www.robotfuzz.com/OSIF14:50
rjeffriestuxbrain_away: http://www.robotfuzz.com/OSIF14:50
rjeffriestuxbrain_away: ...expose as many gpios on the chip possible //YES! Like he said14:52
rjeffriesAT90USB82 Only $2.64 in qty 1 https://avnetexpress.avnet.com/store/em/EMController/Microcontroller/_/N-100185?CMP=KNC-ExpressGooglePPC&HBX_OU=50&HBX_PK=AT90USB82&action=products&cat=1&catalogId=500201&cutTape=&inStock=&langId=-1&myCatalog=&proto=&regionalStock=&rohs=&storeId=500201&term=AT90USB82&topSellers=&x=0&y=015:07
rjeffriesSweeT data sheet.15:10
rjeffrieshttp://www.atmel.com/dyn/resources/prod_documents/doc7707.pdf //data sheet I was talking about15:15
qi-bot[commit] Werner Almesberger: qfn.fpd: added draft for Atmel VQFN28 (ATmega48) http://qi-hw.com/p/ben-wpan/5a9ec3417:02
qi-bot[commit] Werner Almesberger: Added simple UART 8:10 card. http://qi-hw.com/p/ben-blinkenlights/e7ec5c817:11
qi-bot[commit] Werner Almesberger: cameo: allow for rounding errors KiCad produces with a metric grid http://qi-hw.com/p/cae-tools/6947b3a17:40
rjeffrieswprak: is there a PDF of schematic of UART 8:10? I clicked on that link but maybe I do not understand navigation, or no PDF yet. whatever17:42
steve|mrjeffries: just open the schematic with KiCad17:50
rjeffriesI will need to downlaod kiCad. I did a while back then deleted it when I was not using it17:51
rjeffriesso are you adding a hardware UART on Blikinlights, or is this just bit banging with 3.3 volt levels? sorry to be so ignorant17:52
rjeffriesby "just" I am NOT deprecating your effort, simply wondering17:53
rjeffriesbit banging could support a fairly decent serial rate. maybe 56kbps or so??17:54
steve|mwpwrak: by the way, is you PCB etching workflow documented somewhere? :)17:56
rjeffrieswhich chip is used yo do the 6 to 10 i/o muxing?17:57
rjeffriessteveM all I lnow is that over of Playboy or Maxim or similar girlie magazine us ESSENTIAL17:58
rjeffriesbetter tha the expensive photo paper17:58
rjeffriestell you wife of gf you buy it for the articles and the cover ONLY17:58
wpwraksteve|m: hmm, i described it here a few times but it's not "properly" documented. there's a also a fairly involved CNC workflow connected to this.18:00
wpwrakrjeffries: there's an AVR that handles the UART. talks SPI to the ben.18:00
steve|mwpwrak: ah, okay18:01
steve|mthis would be fun as well: http://www.youtube.com/watch?v=R0F0yLOnQZM ;)18:01
kristianpaulseems i lost the track about why blinkenlights have a UART port now...18:02
wpwrakrjeffries: doing it all with bit-banging would be cute, but that's not quite what i'm looking for this time. what i'm after is a simple AVR that gets flashed from the ben. once i have that working, i can apply it to the new atusb.18:02
kristianpaulah flash !18:02
steve|mhttp://www.mikrocontroller.net/attachment/99611/IMG_0340.JPG18:02
rjeffrieswprak that is the correct (smile) answer. yes yes and HELL yes18:02
wpwrakkristianpaul: well, i made a subproject in blinkenlights. seemed a bit empty :)18:02
kristianpaulwpwrak: uart?18:03
kristianpaulis not empty :-)18:03
kristianpaulwell missing code..18:03
wpwraksteve|m: nice :)18:03
kristianpaulbrb i must leave18:03
wpwrakkristianpaul: yes, now uart/ is there :)18:03
rjeffriesmy brain is empty. van we all agree on that?18:03
wpwraksteve|m: i use CNC only to make the board. the rest is chemnical.18:03
steve|mwpwrak: (programming the avr) ah, so you're using bitbanging ISP to program the avr.. nice idea18:04
rjeffriesBetter living through chemistry.TM Dupont18:04
steve|mthen you don't need ISP testpoints18:04
wpwraksteve|m: that's the idea, yes. and the SPI interface can then also serve as the host communication interface :)18:04
steve|myep, really nice that the MOSI/MISO/SCLK stuff is also uses for ISP on the avr18:05
LunaVoraxHello there :)18:06
steve|mwpwrak: by the way, here's the original thread.. http://www.mikrocontroller.net/topic/20654718:07
steve|mattach a 30$ laser from dealextreme to your CNC machine and you're set :)18:07
LunaVoraxIf I want to install the SDL lib on my NanoNote for C + SDL coding, does it works exactly like manually installing it on my x86 linux machines ?18:07
wpwraksteve|m: (laser plotter) hmm, seems rather complex. i thought this was milled.18:10
wpwrakhmm, does serial programming of avrs still work if nRESET is configured as I/O ? the data sheet isn't quite clear on that.18:28
wpwrak(reset disabled) i.e., RSTDISBL programmed to zero.18:28
wpwrakhmm, google says "no"18:29
steve|mwpwrak: no, doesn't work unfortunately (own bad experience)18:31
steve|mAVR fusbite mistake #118:32
steve|mif you don't have a high voltage programmer, you're screwed18:32
wpwraksteve|m: an ugly design "feature"18:35
steve|mwpwrak: indeed.. generally, the AVR has too much fusebits.. I really started to like other architectures which have no fusebits at all18:38
rjeffriessteve|m  Google Chrome translate extension is my friend.18:39
steve|mlike the msp430, which will run off the internal RC oscillator, and you switch to the external clock source in software, with fallback if there's no clock and so on18:39
steve|mguess there are many others out there which do it this way18:40
Action: steve|m has started doing some TI launchpad hacking, and reprogrammed the big MSP used for programming the target yesterday18:41
wpwraksteve|m: yes, that clock selection is rather antiquated18:42
rjeffriessteve|m did you garb some of those 430 Launchpad at the $4.30 price, incl. shipping in US?18:43
wpwrakhmm, not enough i/os for an interrupt line. sigh. wish atmel wouldn't insist on requiring nSS for slave mode ...18:43
rjeffriess/garb/grab18:43
steve|mwpwrak: interrupt Ben -> AVR? you could use a pin change interrupt on any other pin18:44
wpwraksteve|m: interrupt avt->ben. i.e., when data has been received. the problem is that i'm running out of signals.18:44
steve|mrjeffries: I have the launchpads from a german electronics seller for 4,50 each18:44
rjeffriessteve|m I assume 4,50 Euros still cheap18:46
wpwraksteve|m: an 8:10 card has 6 data lines. full spi already occupies 4 of them. i need one more for the clock and one for reset. leaving 0 for interrupts. well, unless playing games with configuring spi on and off. not sure if this works.18:46
Action: steve|m is looking at the schematics18:49
Action: wpwrak is already making radical changes :)18:50
steve|mif you have place for testpoints, SDA cand SCL ;) I2C is always good to have ;)18:51
wpwraksteve|m: naw, no room. i want this to be small :)18:53
steve|mor just connect them to rxd and txd, then you can configure them as high-Z and use either/or :)18:53
steve|mbut well.. might be over-engineered18:53
wpwraki would need a lot of vias for that.18:55
steve|manyway, bitbanged i2c would always be possible if someone should need that18:55
steve|mfor what is the CLK line used? for providing the a clock to the atmega? is the internal 8MHz rc-oscillator insufficient?18:58
wpwraksteve|m: the RC clock is only accurate +/- 10% (without additional calibration). alas, i don't see where in the data sheet they specify the clock accuracy the uart needs.19:04
wpwrakbut +/- 10% seems too much.19:04
wpwrakeven +/- 5% would be a lot19:05
steve|mwpwrak: and which clock can you provide via the CLK output of the ben?19:06
wpwrakwell, if we don't use it, we can recycle XTAL1 for other uses:)19:06
wpwrakthe SDIO clock. the ben's crystal clock is quite accurate. a bit too noisy for RF, but should be perfect for this use.19:06
steve|mokay19:08
steve|mwpwrak: if you should use the internal OSC, and can enable/disable the SDIO clock independently, this would also make an interrupt pin :)19:11
steve|meven quite resource saving, since you could use it as timer0 clock input and then generate a timer interrupt19:13
wpwraksteve|m: i can use the SDIO clock on the ben side also as regular GPIO, yes19:32
steve|mah, even better19:32
wpwrak(or interrupt, etc.)19:33
qi-bot[commit] Werner Almesberger: uart: the AVR requires nSS for slave mode http://qi-hw.com/p/ben-blinkenlights/d3ab4ab19:39
zrafaroh: do you know what did happen with openmoko ML?21:43
zrafait is not fair to see how the whole community dies because no more ML :(21:43
kristianpaulzrafa: http://humaneinfo.com/humane-20110131.tgz22:01
kristianpaulzrafa: i asked to the humane project guy for new dump and posibllityto parse recent wikipedia dump, and seems posible !!22:02
kristianpaulzrafa: just need a computer with 8Gb of ram.... :-|, he said..22:03
wolfspraul I'm writing the 02/01 news22:18
wolfsprauldoes anybody know why the AndroidStamp is called AndroidStamp?22:18
wolfspraulis it supposed to run Android one day?22:18
wolfspraulthere is no way anybody will ever get Android to boot or run with 64 MB memory, that made me wonder about the name...22:19
kristianpaulhe, good point22:19
wolfspraulalso Android without screen doesn't really make sense to me, but that's a separate point22:20
wolfspraulkristianpaul: well, to be critical of them, they do seem to operate in wishful thinking land once in a while :-)22:20
wolfspraulas in "xue is finished", we need to look at that one soon too, to not keep false promises around22:20
wolfspraulkristianpaul: you think it's called 'AndroidStamp' because it's supposed to run Android?22:21
wolfspraulor just because Android is the current fashion and they were so excited about it that they picked it for the name of the project?22:21
kristianpaulwolfspraul: i dont think, well, carlos just started the thread22:22
wolfspraulit's not supposed to run Android?22:22
kristianpauldunno... isnot pointed in the wiki actually22:23
kristianpaulto be fair with you, i'mn follwing linux en caja mail list22:24
zrafakristianpaul: looks like linux en caja ML has a nice amount of messages right?22:28
zrafakristianpaul: humaneinfo: great!... let me know if you get something in spanish :)22:29
kristianpaulzrafa: linuxencaja, well,, my coment could  be biased but i dint like the intro carlos did to adroid stamp..22:31
wolfspraulhe made an intro? will it run Android?22:31
kristianpaulno said nothing about android22:32
kristianpaul"las personas con conocimiento en HW pueden opinar sobre el"22:32
kristianpaulproceso, los que no porfavor no participen,22:32
kristianpaulzrafa: i dint like that22:32
zrafakristianpaul: yeah .. I just read that :D22:32
zrafa"las personas con conocimiento en HW pueden opinar sobre el22:32
zrafaproceso, los que no porfavor no participen, no vamos a dejar abiertas las22:32
zrafaespecificaciones, asi que no vamos a recibir sugerencias de HW ya que muchas22:32
zrafaveces las personas que no conocen el desarrollo HW no saben lo complicado y22:32
zrafacostoso que es agragar una nueva característica."22:33
kristianpaulseguro22:33
zrafakristianpaul: another thing more :22:33
kristianpaulpero en fin,,, no es mi estilo de decir las cosas22:33
Action: kristianpaul biased22:33
zrafakristianpaul: what does no vamos a dejar abiertas las especificaciones" mean?22:34
zrafais it something like open thing but closed ideas?.. or something like that?22:34
kristianpaulcopyleft like license22:34
zrafakristianpaul: but why he said "no vamos a dejar abierta las especificaciones"?22:35
kristianpaulzrafa: i dont want give personal comments about how i think emqbit works.. may be i'm wrong on that,22:35
kristianpaulzrafa: i guess it mean what you said, no more hw changes22:36
wolfspraulhe :-)22:36
wolfspraulI tried google translate22:36
wolfspraulit says the hw work is not open/collaborative, because outsiders don't understand the implications of the changes they propose?22:36
wolfspraulthat's a very good point, but the exact opposite of our beliefs over here :-)22:36
wolfspraulthose implications need to be documented, communicated, so that a collaborative effort becomes possible22:37
wolfspraulmaybe google translate got it wrong22:37
wolfspraulanyway I was wondering about the name 'AndroidStamp', guess we don't know right now22:37
kristianpaulzrafa: i dont want invest my free time time about learning about arm, i got my M1, good or not that my bet :-)22:37
zrafawolfspraul: you mean the spanish text we put here?22:38
wolfspraulyes22:38
wolfspraulI just pasted it to google translate22:38
wolfspraul"people with knowledge of HW can comment on the process, please do not not participate, we will not leave open the specifications, so we will not receive tips from HW because sometimes people do not know the HW development not know complicated and costly it appends a new feature. "22:39
zrafawolfspraul: "las personas con conocimiento en HW pueden opinar sobre el proceso, los que no porfavor no participen," : people with hardware skills can help with ideas, people without hw skills please dont22:39
zrafawolfspraul: "no vamos a dejar abiertas las especificaciones, asi que no vamos a recibir sugerencias de HW ya que muchas veces las personas que no conocen el desarrollo HW no saben lo complicado y costoso que es agragar una nueva característica." :22:40
kristianpauli wonder from where he learnt that last one :-)22:41
zrafawolfspraul: "we will not let open the specifications, so we will not accept suggestions about hw, because often people does not know how hard the hw development is, and how haw to add a feature is"22:41
zrafahow haw=how hard22:42
zrafawolfspraul: is that clearer than google translate? :)22:42
kristianpaula lot clearer22:42
kristianpauli dont get the point, if the list is not about hw then? just put linux on a box (linux en caja), well sounds coherent,.22:43
zrafakristianpaul: M1: he he.. great!.. yeah, I can imagine that you will not have time now.. just M1 for many days :)22:43
kristianpaulInicialmente, nuestro objetivo es proveer  de una alternativa a boards22:44
kristianpaulde 8 bit, también de bajo costo pero basada en un ARM de 454Mhz y el22:44
kristianpaulSO Android.22:44
kristianpaulCreemos que la board va a ser muy útil para los que  inician en22:44
kristianpaulembebidos basados en Linux y esa cosa rara salida de Linux, Android.22:44
kristianpaulthey will run Android OS !22:45
kristianpaulzrafa: m1 and gps stuff, yes :')22:53
wolfspraulkristianpaul: no they will not run Android, at least not on that board, if it's ever produced23:01
wolfsprauleven 1.6 already needed 128 mb just to boot (not really run any apps)23:01
kristianpaulhmm.23:02
wolfspraulI'm sure now with 2.2/2.3/3.0 etc. you better have 512 mb memory or so23:02
kristianpaulwell.. let see what android/linux hacked version they produce23:02
--- Tue Feb 1 201100:00

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