#qi-hardware IRC log for Thursday, 2010-12-16

qi-bot[commit] Werner Almesberger: cameo/ops.c (tool_comp_paths): inside/outside logic was reversed for outer path http://qi-hw.com/p/cae-tools/ed6bbff00:39
qi-bot[commit] Werner Almesberger: atusb/cam2/mkmk: new-style CAM script http://qi-hw.com/p/ben-wpan/000d27000:40
qi-bot[commit] Werner Almesberger: cngt: added support for setting default positions from toolpath file http://qi-hw.com/p/cae-tools/ffaded702:50
qi-bot[commit] Werner Almesberger: mkmk: further improvements and updates http://qi-hw.com/p/ben-wpan/1bef41f03:20
qi-bot[commit] Werner Almesberger: atusb.brd: removed stray pad http://qi-hw.com/p/ben-wpan/d7719af03:20
adamwangwolfspraul, everytime I need to generate shopping list I must 'make clean'?05:59
adamwanghmm...it must. I got my answer.06:22
qnrqhi, I just received my nanonote today and I just hooked it up to the net. when I run opkg update it tries to connect to a file that returns 404, more specifically http://downloads.qi-hardware.com/software/packages/openwrt/xburst/latest/Packages.gz is it just temporarily down or is there a new location that was exluded from my release? did ofc google :-)13:30
wpwrakqnrq: i think the whole package database is very out of sync. it's at ttp://downloads.qi-hardware.com/software/packages/NanoNote/Ben/latest/13:53
wpwrakbut from april ..13:53
qnrqwpwrak: yeah, heh.. is there any way to change the download dest the same way as one would edit /etc/apt/sources.lst?13:53
wpwrakprobably ;) but i never tried to find where this lives. you may have to poke around /etc for a bit13:54
qnrqthat's what I thought, I set up a temporary mirror and edited /etc/hosts for now :)13:55
wpwrakand i'm not sure if the packages will actually be useful. i think there were one or two libc changes in between.13:55
qnrqI'm just trying to play around with the sound for now anyway13:55
qi-bot[commit] Werner Almesberger: cameo: added optimization of movements between tool paths http://qi-hw.com/p/cae-tools/57b4e2714:38
qi-bot[commit] Werner Almesberger: atusb: updated and improved CAM script http://qi-hw.com/p/ben-wpan/ba310a814:38
kristianpaulwpwrak: qi-bot may be is fault of apache indexer16:00
kristianpaulpage indexer soemthing, as i know all stuff from download.qi-.. is not listed as it is on the directory16:01
wpwrakkristianpaul: hmm ? which problem are you referring to ? "not listed as it is on the directory" = how does it differ ?16:09
kristianpaul"not listed as it is on the directory"16:10
kristianpaulor well i think that may be a cause16:11
wpwrakcause for what ?16:11
wpwrakyou speak in riddles ;-)16:11
kristianpaulhmm let me check i'm following tha backlog right16:12
wolfspraulqnrq_: maybe xiangfu can help you20:08
wolfspraulxiangfu: if you read the irclogs, you see that qnrq had some problem with the OpenWrt packages URL20:09
Action: FrankBlues waves20:10
xiangfuqnrq_: /etc/opkg.conf :)20:11
wolfspraulwaves back but doesn't know how to do so in a proper IRC way... :-)20:11
FrankBlues"/me does action"20:12
FrankBluesThanks wolfspraul!20:12
Action: wolfspraul thanks FrankBlues20:13
wolfspraulyeah, it works!20:13
wolfspraulnow it only needs to sink into my brain...20:14
Action: wolfspraul is happy, just got up getting my coffee!20:14
qi-bot[commit] Xiangfu Liu: bash-files: update the package url http://qi-hw.com/p/openwrt-xburst/6f74fce20:19
wolfspraulthe first task my coffee has to help me over: setting up a new server, using the shiny GNU parted instead of old fashioned fdisk. but then... mklabel MSDOS!20:20
wolfspraulyes, it's true. it reminds me. the partition sector format most of us are using is still horrible old MSDOS style, with endless broken special cases. argh.20:21
kristianpaulyup :(20:21
wolfspraultruth hurts...20:21
kristianpaulAt least there is openbsd users around? :)20:21
wolfspraulwhat is the general replacement these days? or is there none because partitions are 'bad' anyway?20:21
wolfspraulI guess I still go with msdos20:22
kristianpauldont know really20:22
wolfspraulmac has long switched to gpt20:22
wpwrakgood old dos partition table seems to work fine :-) never had any troubles20:22
kristianpaulBtw is that fedora, what are you setting up?20:23
Action: kristianpaul already got FEL and really curios of tring more in depth20:23
wolfspraulwpwrak: good that you don't (seem to) know how it counts cylinders, heads, etc. and how much magic is needed to translate it to modern disks without rounding problems etc.20:24
wolfspraulkristianpaul: yes, and in such opportunities I always try to improve several things at once, to not waste too much time.20:24
wolfspraulso I will also improve the backups substantially, use LVM snapshots, prepare for an actual rescue/recovery scenario on another server, etc.20:25
kristianpaulgood :)20:25
wolfspraulso far I'm just running rsync, but without actually ever walking through a rescue scenario20:25
wolfspraulso that's a mess20:25
kristianpaulsnaphots are very usefull20:25
wolfspraulrsync from underneath running mysql databases and such...20:25
wolfsprauldon't even think about it20:25
kristianpaulsure dont (*sigh*)20:26
kristianpauli wonder if that worked (considering the running thing)20:26
kristianpauli guess it did :)20:26
wpwrakwolfspraul: oh, i know very well what ugly things it does ;-)) remember lilo ? in "linear mode", it translates things to that CHS format20:27
wpwrakwolfspraul: but recently, i never ran into any troubles with that stuff20:27
wolfspraulyes true, seems it is successfully 'masked'20:28
wpwrakwolfspraul: (of course, having handed off lilo maintenance many years ago may help :) so, end-user's perspective here20:28
wolfspraulbut still there, and the name is a reminder...20:28
kristianpaulah wikireader content is package in some binary way20:31
qnrq_wolfspraul, xiangfu: thanks a lot for the the help with /etc/opkg.conf :-)21:04
xiangfuqnrq_: we try to build the new packages whis weekend. hope we don't get much errors :) , if all goes fine. we will have a update packages.21:06
qnrq_xiangfu: awesome! I just got my chance to start playing around with my nanonote today :)21:07
wolfspraulqnrq_: thanks for buying one!21:08
qnrq_wolfspraul: not sure if I'm the one to thank here, so far it's a beautiful device to hack :>21:09
wolfspraulok keep us posted...21:11
qnrq_for sure :)21:13
kristianpaulwpwrak: (DMA) is just for mmc as far as read on qi wiki21:21
kristianpaulwell at least in nanonote21:29
wpwrakkristianpaul: yeah, it seems that there's no external DMA. there's only an empty chapter in the 4720 data sheet about it, that's all.21:36
kristianpaulargg (empty)21:38
kristianpaulwpwrak: ddi you look at page 107 from  jz4740_pm ?21:46
wpwrakwell yes, that page and the ones following it ;-)21:54
kristianpauli dont undestand :/21:54
kristianpaulthat pages, well i still readin what DMA is ;-)21:55
wpwrakthe description is a bit cryptic. they also jump between register-based and descriptor-based operation21:55
kristianpaulhmm, Is like a bulk transfer in USB?21:56
wpwrakhmm, kinda. basically, instead of setting up one transfer, you define a list of transfers21:56
kristianpauli need that ;)21:57
kristianpaulcan i set up 2048? (theorically)21:57
wpwraknot sure if it helps you much. you still need a trigger. and there are no simple external DMA triggers.21:57
wpwrak2048 what ?21:57
kristianpaultransfers of 8 bits each21:58
wpwrakprobably, yes21:59
wpwrakbut again, how do you trigger it ? it's messy21:59
adamwangwolfspraul, i think next time we need re-layout for Jtag/serial cable22:01
wolfspraulwhat did you find?22:01
adamwangfor JTAG 14pins receptacle i found http://search.digikey.com/scripts/DkSearch/dksus.dll?lang=en&site=US&WT.z_homepage_link=hp_go_button&KeyWords=79109-1006&x=0&y=022:01
wolfsprauladamwang: ahh. have a story for you there. wpwrak thinks it's super easy to source a 14-pin jtag + 4-pin serial connector set that will have the same height22:02
adamwangclick the datasheet link,22:02
wpwrakit is :) just don't mix product lines ;-)22:02
adamwanglet me finish first22:03
wpwrakhey, how hard can it be if even the arduino crowd does it ? ;-))22:03
wolfspraulwow that is crazy expensive22:03
wolfspraul3.18 USD22:03
wolfspraulshould be more like 10 cents :-)22:03
adamwangno matter you selected a 3M part or I found Molex, the receptacle parts always has 'positioner' to hook up pcb,22:04
adamwangcurrent pcb gerber there's no two 'through holes' can be inserted!22:05
wpwrakadamwang: would going through-hole be an option ?22:05
wolfspraultoo late, gerber already out22:05
wolfspraulbut in the future, sure22:05
wolfsprauladamwang: at least for 100 the price is 1.87 USD. still crazy expensive22:05
wpwrak yeah, i meant in the future22:06
wpwrakwolfspraul: even this connector needs new holes for the positioning pins22:06
wpwrakah wait, no22:07
wolfsprauladamwang: maybe you can ask Yanjun Luo where he sourced the connectors and we buy from there?22:07
wolfspraulthat's another option...22:07
wolfspraulbut if this Molex one works just buy that, is faster probably22:07
wpwraknaw, yes, it need holes22:07
wpwrakgot confused because the data sheet first shows one without22:08
adamwangthey are matched each other, so i want directly buy this.22:08
adamwangif you go to see the bom to find p/n: 87832-1420 , then you can find 79109-1006 matched in Digi-Key.22:09
adamwangwell...so this means when the parts arrives, I need to manually cut those two positioners. man!22:10
adamwangand probably loose accurate placement!22:10
adamwangwell, i think i still can control this.22:11
adamwangumm..i'll ask him..but if no datasheet from him available...then i order Molex this one.22:12
wolfspraulthe digikey p/n of the new one is WM18702-ND22:13
wolfspraulshould I update the bom?22:13
adamwangnot now22:13
wpwrakadamwang: are you sure about the positioners ?22:14
wpwrakadamwang: the ~-1006 doesn't seem to have them22:14
wolfsprauladamwang: the old one was http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=3M5414CT-ND22:14
wolfspraulwhat was wrong with that?22:14
adamwangsure! why not? the datasheet needs positioners, even without this, trying to imagine how this SMD part will be mouted accurately on board.22:15
adamwanggo to page 422:15
adamwangyou will see22:15
wpwrakadamwang: yes, page 4 had those pegs. but this is for the parts listed on page 5. the ~1006 is on pages 1/222:17
adamwang1.45mm height22:17
adamwangsee page 4, there's recommended PCB layout22:18
wpwrakadamwang: yes, that's for the parts listed on page 5 :)22:18
adamwangthere's two circles in there(means through holes needed)22:19
wpwrakadamwang: do you want the 79109-1006 ? or another one ?22:19
wpwraklook at page 2. it explicitly says "WITHOUT PEG OPTION"22:19
wolfspraulthe 3M one has no 'locator posts'22:21
wolfspraulit's an option but with p/n -2000- it doesn't have them22:21
adamwangwait..let me see page 222:21
wolfspraul(and for 7 positions it's not available, only 8 or more if I understand correctly)22:22
adamwangwpwrak, you are right! the p/n: 79109-1006 without peg positioner22:23
adamwangbut this is still very bad...22:24
wolfsprauladamwang: what is wrong with 3M 153214-2000-RB-WB ?22:24
adamwangmeans smt machine needs to pick & place in 'correct' location.22:24
wpwrakadamwang: yup22:25
wpwrakadamwang: if the machine doesn't have a positioning error > 0.1 mm, you're probably fine. if the connector needs a little extra push, nobody will notice :)22:26
adamwangthe terminal won't be contact well to 87831-1420 pin header even they are the same 2.00mm pitch22:26
adamwangwpwrak, hahaha..22:26
adamwangi don't like this way..22:27
adamwanganyway...just fyi and records this to be fixed next time.22:27
adamwangbut bad thing is now the thickness is 1.6mm, also there's routed trace on both side..22:28
wpwrakwith positioning pegs/pins/posts/nipples it's nicer, agreed22:28
adamwangsurely , that's the rules for matching mechanical parts!22:29
wpwrakwhy is 1.6 mm (PCB thickness ?) a problem ?22:29
wolfspraulstill don't know what's wrong with the 3M part22:30
wpwrakadamwang: i've been fighting all day long today with such positioning posts ;-) in my mill, the board moves a little (sideways) when the tool pushes against it. so holes that i mill (instead of drilling) end up a little too small. great fun ...22:31
adamwanghmm...well...always careful the peg/positioner/pins while design stage...then safe later22:33
wpwrakfor my DIY board i usually just cut them off. too much work to drill those holes :)22:34
adamwangha...well but you know if those go to smt/dip process during MP production, it saves time without your diy! :)22:35
adamwangof course now you must diy.22:36
adamwangand hates those holes to hell.22:36
adamwangwolfspraul, the 3M first needs peg/positioner22:36
wolfspraulwhat do you mean with needs?22:37
adamwangthen when you select matched parts, also check inlet of Mate point .22:37
wolfspraulit does not have what they call 'locator posts'22:37
wolfspraulthe 3M part number we have does not have positioner, I think22:38
wpwrakadamwang: in the last days, i improved my collection of programs for milling boards and such. now i can actually precision-drill things. so the pegs no longer scare me :) (well, just a little - still need to add a compensation for the board deflection)22:38
adamwangno , it has. see the recommended solder pad layout, there're two.22:39
adamwangwpwrak, wow..man! incredible you are!22:39
adamwangwell...i have to work others...cu22:39
wpwrakwolfspraul: what's the problem with the 3M part ? it sounds as if you're looking for an exact replacement ?22:41
wolfspraulno I don't care which one is best22:42
wolfspraulbut the 3M part we have has no positioners22:42
wolfsprauladamwang: I don't see it. note 4 says it's optional, and in the ordering information if the p/n is -2000- it has no positioner.22:43
wolfspraulif it's -2020- then it has the positioner, but according to the datasheet that option is only available for 8 positions or more (see Note 4)22:44
wpwrakadamwang: look carefully - the connector exists with and without pegs. they just didn't show the footprint without them22:44
wpwrakadamwang: (see "Ordering Information"). the equivalent part with pegs would be 153214-2020-RB-WB22:45
wolfspraulwpwrak: yes but positioner is only available for 8 or more positions, if I understand correctly22:45
wolfspraulbut not sure about that it says "optional locator posts available for positions 8 and greater"22:46
adamwanghm..you both are right22:46
wpwrak.. which mdigi-key have as 153214-2020-RB22:46
wpwrakwolfspraul: 14 > 8 :-)22:46
wolfspraulok makes sense so they count each pin22:47
wolfspraulwell then, we can have 14 with or without pegs, and the current p/n we had was without22:48
wolfspraulsame as the Molex part only that the Molex one is more expensive22:48
adamwangwait..guys..i'm drawing ...i am not only to see this part only22:48
adamwanglater you will see..i haven't decide to buy which one exactly22:49
wolfspraulok sure22:49
adamwangjust a second22:49
wolfspraulall fine22:49
adamwangwolfspraul, http://downloads.qi-hardware.com/people/adam/m1/jtagserial/finding_a_correct_height_for_P1.png23:12
adamwangmy goal is try to find p1 with correct height and 14pins receptacle with good 'contact point'.23:14
wolfspraulok looks good23:14
adamwangthe '?' is the P1 on Jtag/serial board....so let us see carefully with datasheet!23:15
adamwanghope we won't make mistakes again. :)23:16
adamwangi'll find a suitable height for P1, be careful keep the J6 and receptacle with good 'contact point' enough!23:18
adamwangsee you later23:18
kristianpaulnice see adamwang more often at irc :)23:37
kristianpaulsee and read :)23:37
wpwrakyup :)23:39
kristianpaulhe finally got the SYNC/423:41
kristianpaulnow lets if if sinthetize ;-)23:41
kristianpauli need read this doc from xilinx about xst, i wonder what lekernel_ had to deal when making MM SoC ;)23:42
kristianpaulgood ir pass :)23:45
kristianpaulDDR3 in Xue is to much challeging?23:50
kristianpaulis not to*23:50
--- Fri Dec 17 201000:00

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