#qi-hardware IRC log for Thursday, 2010-08-05

wpwrakwolfspraul: hmm, you said that project creation wasn't automatic, yet i could add ben-counterweight without problems. seems that things work smoothlier than expected (either by me or by you :)02:59
wolfspraulwpwrak: I gave you super powers :-)03:01
wolfspraulIndefero has three levels, normal user, 'staff' and 'admin'03:01
wolfspraulI made you staff right away, actually I would also make you admin but it's only possible with a manual mysql command and I was too lazy and forgot what the actual extra power of admin is.03:02
wpwrakwolfspraul: aaah ! :) i kinda suspected that :) thanks !03:02
wpwrakwolfspraul: heh, being able to create my own projects already is plenty :)03:02
wolfspraulI think admins can delete stuff permanently03:03
wolfspraulit's actually good to make that a bit harder, accounts can be compromised etc.03:03
wolfspraulno need to be so public to have the 'kill' button right on the homepage :-)03:03
wpwrakyup. actually, a revision control system underneath would be nice. you never know why bad things may happen, so it's reassuring if you can fix them in any case.03:04
wpwrakwell, a feature request for the future perhaps03:04
wpwrakwolfspraul: btw, do you think you;ll have some use for the counterweight ? i expect that i'll have it in decent shape after the weekend.03:14
wolfspraulin mass-production not right away, but of course we can use it as a give-away, I'm sure David would love to have a few etc.03:15
wolfspraulwe can also use this chance to experiment how to make more03:15
wpwrakwolfspraul: david, for heavy metal :)03:16
wpwrakwolfspraul: for mp, you'd want to have something rohs-compatible anyway. lead doesn't look too pretty there ...03:16
wpwrak(make more) i'm using a wooden mold right now. already did ~12 pieces and the mold is still good. yield depends a lot on handling, though, so this is bad. there are a number of possibilities for making this more predictable. also, an aluminium mold should last quite a long time.03:18
wpwrakright now, it works like this: the mold is "flat", i.e., has the profile of the counterweight seen from above. i pour liquid solder inside, then i tap it with a pcb and let it cool03:20
wpwrakthe taping is critical: the faster, the more accurate the thickness. the slower, the more time for air to escape. there's also some effect on getting the solder to flow into all the corners.03:21
wpwrakso all this is nice for experimenting but too unreliable for anything else. the next step would be to try gravitation casting. maybe even the same mold will do. (mill is busy until friday with a scan, so reuse makes sense)03:22
wpwrakif that doesn't work, i may try heating the alloy while already in the mold. that may give more reliable definition. (i already got a nearly perfect cast with the primitive approach, so it's now all about making this reproducible)03:24
wpwrakfor large-scale production, you;d want to change the material and then you'd have to find out what your local companies can offer. cad files should be no problem.03:25
tuxbrainwpwrak: Yeah! (both for heavy metal and to have some counterweight once ready :)03:34
wolfspraulkristoffer: morning :-)03:37
wolfspraulgenius or weirdo?03:37
wolfspraulsounds tiring...03:37
kristofferwolfspraul, hey man :) Just answerd an email from you. Im off to school in a second (big exam in two weeks), but I can catch you later tonight.03:46
kristoffercya then03:46
mirkoanything special about the serial connection at sakc? i just get crypted symbols with baudrate 57600 (tried otehrs as well)04:21
xiangfumirko: do you use a TTL -- RS232 convert ? ( I don't have sakc. just some idea)04:36
mirkoxiangfu: sure i use - the setup is working fine with all my other embedded devices (such as the nanonote, the picture-frame, etc.)04:36
mirkoxiangfu: i also _see_ symbols scrolling over, however no human readable characters... usually that's the case of wrong settings... however i thought these are fixed within the SoC in this case, so i wonder...04:37
xiangfumirko: then I have no idea. :(. form the sakc uboot. it's definitely  57600.04:40
mirkoxiangfu: it's not just uboot/kernel - it's the same while flashing...04:40
mirkowhen flashning the jz4725 there's no software involved (except the stage-binaries which are the same as i used for the picture frame and the nanonote)04:41
mirkoweird - just dropped carlos an email04:41
mirkomaybe he can give me some hints04:41
uncloudedkyak: thanks for the fixes to the freedroid Makefile04:55
wolfspraulmirko: did you write your mail to Carlos in private? I don't see it over the list yet...05:07
wolfspraulalso, please remember to upload sakc binaries, no matter how preliminary they are, so Adam can get into the mode of flashing & trying them05:07
wolfspraulthese are different from our official releases, so don't hesitate to upload anything :-)05:08
wolfspraulotherwise Adam can only reflash Ben images and naturally they will not even boot05:08
mirkowolfspraul: i wrote to him in private, yes05:48
mirkowolfspraul: i need the serial connection to see what's going on05:49
mirkowolfspraul: it "looks" good from the outside-so if i get a shell via USB i'll upload bins05:49
wolfspraulgreat, thanks!05:49
mirkobut the serial connection is important05:50
wolfspraulyes sure, but if your image boots, at least that's a start05:55
bartbesthe ben automounts sd cards, but how do I make it mount it on /card (or at least symlink)?10:20
mththe automounting is probably done by a program/script that is run on hotplug events10:23
bartbescaps lock on 2nd keyboard10:24
mthcat /proc/sys/kernel/hotplug10:24
bartbesoh I think the conf is in10:24
bartbesand mounting is in 40-mount10:25
bartbes(just confirmed that)10:25
bartbesI wonder where it is configured..10:28
xiangfubartbes: under /etc/config/. name is "mount" I think. I am booing my nanonote now.10:31
bartbesoh there we go10:31
bartbesthe one named fstab10:32
bartbesright, so it basically says it will check fstab10:32
bartbesnow to add the device to fstab10:32
bartbesnot sure how I am supposed to do that10:33
bartbes(since fstab itself only states it is auto-generated)10:33
xiangfubartbes: http://wiki.openwrt.org/doc/uci/fstab10:37
bartbesof course openwrt docs10:37
Action: bartbes restarts nanonote10:41
bartbesxiangfu: thx, I got it to work10:59
xiangfubartbes: you are welcome :)11:00
kristianpaulnice pics11:34
bartbesnice cybering kristianpaul11:38
qi-bot[commit] Xiangfu Liu: add fw_env.config for change kernel commandline http://qi-hw.com/p/openwrt-xburst/e67fd6e11:39
kristianpaulbartbes: hola11:42
wpwrakrafa: you have to give me an ultimatum for the install session :) without a bit of pressure, i keep on dragging things out indefinitely.12:07
qi-bot[commit] Xiangfu Liu: enable [setfont2] and [uboot-envtools] package in config http://qi-hw.com/p/openwrt-xburst/33d268012:09
larscnanonote support is now upstream :) http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=e6b78c4f224925c71cce57033b1e6e30dd56add712:41
xdpiratenice one larsc!12:42
wpwraklarsc: congratulations ! so it's champagne and girls tonight ? :-)12:45
larscwpwrak: unfortunately not. to much work to do :/12:51
wpwraklarsc: sigh. what happened to the good old times, when returning heroes were welcomed with a parade and an orgy ?12:52
wpwraki really miss good old rome and the apollo program ...12:53
bartbesespecially the combination of the two12:54
wpwrakbartbes: hmm, apollo's antagonist in style, dionysus, might be the better companion for a bit of fun, though12:57
mthlarsc: congrats!13:05
bartbesdoes anyone have experience with compiling c++ programs which link against lua on openwrt?13:40
wpwrakhmm, no commit notifications for ben-counterweight yet :-(14:02
wpwraktuxbrain: so .. how many counterweights do you think you'll want ?14:23
rafawpwrak: haha.. yes, you are right :D15:26
rafawpwrak: just that I was with few free time :(..15:27
tuxbraincongrats lars!!!!!15:32
tuxbrainwpwrak: don't know how much is the price?15:32
viric_what is about 'congrats'? :)15:44
kristofferlarsc, congratz!15:48
viric_Ah, support upstream :)16:00
kristofferyes, nothing like that taste of vanilla16:01
viric_I also enjoy that! In what release it is going to appear first?16:01
larscthe next16:02
viric_great. Congratulations16:02
qi-bot[commit] Bas Wijnen: start wrapper; proper serial port baudrate http://qi-hw.com/p/iris/765967316:20
wpwraktuxbrain: depends on how many you need and in what form :) varies between "i send you files" over "you pay the shipping" to "now it's actually turning into real work".16:32
wpwrakrafa: (you not having time) better than you angrily watching the clock while i'm busy doing other stuff :)16:32
tuxbrainwpwrak: then I will reformulate the question... what is the scale price? :P16:34
tuxbrainprice scale sorry, too sleepy16:35
shevekDoes anyone know how the kernel powers the Ben off?  According to the programmer's manual, this can be done using the rtc, but setting the PD bit doesn't cause a power down.16:36
larscshevek: use the source ;) http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=arch/mips/jz4740/reset.c;h=5f1fb95c0d0daf980bc20194be8bb4c6f35f002d;hb=HEAD16:46
wpwraktuxbrain: haven't figured out what i would have to charge for it if i have to charge for it. so it's better to try to keep it simple enough that it's not worth the effort to put a price on it :) but just tell me what quantity you intend to use. then we'll see how to best get this done.16:46
tuxbrainlets say 20 , do an estimation of what will be the cost including shipping and we can decide if it worth the effort (yours and mine)16:48
wpwraktuxbrain: 20 should still be within the range of "pay me a beer or two when we meet anywhere" :) weight would be about 400 g, so shipping shouldn't be too evil either. provided that there are no restrictions on lead.16:55
wpwraktuxbrain: do you have a fedex/dhl/ups/tnt/etc. account the shipping could be charged to ?16:56
tuxbrainwhere are your from?16:57
tuxbrainI have no direct acount but I can ask to my courier to pickup there.16:58
wpwraktuxbrain: i live in argentina16:58
sheveklarsc: There may be a big power saving opportunity there, then. :-)  The kernel does what I did, followed by while (1) asm("wait"); which means it just sits there waiting if power down doesn't work.  So if it doesn't, the system doesn't really power anything down.16:58
tuxbrainwell then you can send it to rafa, I will meet him at the end of this month or begins of september16:59
larscshevek: line 7016:59
wpwraktuxbrain: oh, the two of you will meet. perfect then.16:59
sheveklarsc: Yes, that's what I'm doing in Iris as well, and it doesn't do anything.16:59
larscwell, here it shutsdown the system17:00
wpwraktuxbrain: he'll have to pick them up himself, though :) (he lives within walking distance from my place)17:00
sheveklarsc: Hmm, you should notice indeed; otherwise it can't be powered up using the power button.17:00
tuxbrainhehehe :) ok I will ask him also to transfer you some beer , I wil transfer to him here in spain slow-fi beer17:02
wpwrakbeer store-and-forward :)17:04
wpwrakrafa just agreed :)17:04
wpwrakrafa: of course, you could also have answered right here ;-)17:05
rafatuxbrain: can I drink the half of beers for wpwrak?17:07
rafatuxbrain: btw, I need to buy a nanowar version as well17:09
rafafor a friend17:09
tuxbrainrafa:You have a full slow-fi beer to you, don't need to split the packet :), you will act as proxy and also as destination address17:13
sheveklarsc: It works now, I forgot to open the gate again.  Thanks!17:50
bartbesthis is weird18:07
bartbesthe official tcl package18:07
bartbesis semi-broken18:07
Action: bartbes removes it18:07
bartbesno use for a broken interp18:08
mthbartbes: they might have upped the version without testing whether it still works18:16
mthI had a lot of cross compilation problems with Tcl in openMSX18:16
rafawpwrak: so, which is the deal?.. tuxbrain gets the lead and we get the same weight on beers?19:39
wpwrakrafa: half a liter of beer ? hmm, i've seen better deals :)19:47
wpwrakrafa: maybe the weight in nanonotes. that would be about two of them ;-)19:47
rafajaja.. well, you define that with tuxbrain, I am happy just with two beers as usual :D19:51
rafabah.. we, south americas, are so cheap :)19:51
Textmodewpwrak: with or without batteries?19:55
wpwrakTextmode: good idea :) the battery is something like 20g, too :)20:19
kristianpaullarsc: :)21:11
wolfspraulwpwrak_: ben-counterweight commitlogs should work now23:37
wolfspraulthere are 2 steps that need to be done manually right now after the first commit23:37
wpwrak_wolfspraul: thanks !23:38
wolfspraulwpwrak_: since you are the kicad expert...23:40
wolfspraulwhat do you think about adding some server-side scripts where a commit of kicad files will automatically generate a PNG or PDF version of the schematics23:41
wolfspraulor other magic like some of your bom scripts etc.23:41
wolfspraulI would be very open minded to working on some kicad automation on the projects server23:41
wolfspraulof course it needs to make sense23:41
wpwrak_it's a bit tricky: if you use my print from command line patch, then you need to find the top-level .sch file, then print that one23:44
wolfspraulwell first I would like to understand what could create value23:45
wpwrak_but yes, easily/browser-viewable schematics are definitely a good thing23:45
wolfspraula lot of people like to look over schematics, but installing kicad is a big burden23:45
wolfspraulso if a commit could trigger the generation of PNG/PDF schematics, I think that would be valuable23:46
wpwrak_yup, agreed23:46
wolfspraulsimilar for the BOM maybe, although as you know I am somewhat cautious on the value of some of your scripts in that area :-)23:46
wolfspraulkicad is still an electrical design tool for me, at some point the business/price/sourcing side starts, with separate tools23:47
wolfspraulwarehouse management, etc.23:47
wolfspraulso maybe we start with the png/pdf schematics then?23:47
wpwrak_my bom scripts are definitely just a first step23:48
wpwrak_the problem with the schematics conversion is that you need to plot from the root. if the .sch file is a sub-sheet, and you try to plot just that file, it will come out wrong.23:50
wpwrak_"plot just that file" = open foo.sch and try to plot it, or use eeschema --plot foo.sch, there foo.sch is the sub-sheet23:51
wolfspraulone radical solution would be to commit the live script into the repository :-)23:51
wolfspraulthe perfect storm in terms of taking over the server :-)23:51
wpwrak_yeah :)23:52
wpwrak_also running kicad on arbitrary schematics may be a bit dangerous23:52
wolfspraulso we don't know which .sch is the root23:53
wpwrak_less than heekscad, though ;-) fewer bugs23:53
wpwrak_you could scan them for ... $Sheet ... F1 "foo.sch"23:54
wpwrak_that is, F1 "foo.sch" in a $Sheet ... $EndSheet block23:54
wpwrak_the one(s) that have such a block are parents of foo.sch23:55
wpwrak_recurse until you find no further parents. of course, they could probably be in a different directory23:56
wolfspraulsounds complicated23:56
wpwrak_a good hint are also .pro files23:56
wpwrak_e.g., if there's a bar.sch and a bar.pro, then you may want to --plot the .sch whenever a .sch in the same directory changes23:57
wolfspraulI think much easier maybe to have a file named 'git-hooks' or so in the root dir, and then 'root-sch path/to/root.sch' in it23:57
wolfspraulthat would also allow us to add more logic maybe around bom later23:57
wpwrak_yes, that would work, too23:57
--- Fri Aug 6 201000:00

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