#qi-hardware IRC log for Wednesday, 2012-11-14

wpwrakwolfspra1l: some people may be interested in buying atusbs. does sharism.cc sell them ?01:46
xiangfuwpwrak, if he in China mainland: atUSB: http://item.taobao.com/item.htm?spm=a1z10.1.w4.5.NgYcpn&id=1585650689602:08
xiangfuatBen: http://item.taobao.com/item.htm?spm=2013.1.31-2129074574.4.0CLsSV&id=1585647466102:09
wpwrakxiangfu: heh, thanks :) but i think they're not in china10:13
wpwrakxiangfu: one question, though: where do you keep the kernel tree for the nanonote openwrt, with atben/atusb ?10:14
xiangfuwpwrak, http://projects.qi-hardware.com/index.php/p/openwrt-xburst/source/tree/master/target/linux/xburst/patches-3.3/800-WPAN-new-files.patch  and  http://projects.qi-hardware.com/index.php/p/openwrt-xburst/source/tree/master/target/linux/xburst/patches-3.3/810-WPAN-patches.patch10:24
wpwrakthanks !10:28
xiangfuwpwrak, checkout http://en.qi-hardware.com/wiki/Mini-slx9, any comments will be great. 10:33
xiangfuwpwrak, I already etched 6 boards. I can not take good picture at night. more document on etching tomorrow. 10:33
wpwrak(picture at night) they should build more hacker-friendly cameras, wouldn't you agree ? ;-)10:34
wpwrakah, i see that you're making good progress with the paper. the results look good indeed10:36
wpwrakfor the list of papers, if you have product codes for the kodak papers, it would be good to include them. in case someone wants to buy exactly the paper that worked for you.10:36
wpwrakfor "Plot KiCAD design out", maybe add the pcbnew command line ? i.e., pcbnew --plot=ps --plot-fill-all-zones --layers=Front --plot-mirror PROJECTNAME.brd10:38
wpwrak(requires our patches, of course)10:38
wpwrakwolfspraul: sharism.cc has now all the atben/atusb boards, correct ?10:39
rohhmm.. any ideas where one could get mylar or kapton film (without glue) .. like 0.1mm thick?10:39
rohor whatever works well for smt stencils.. ive read that mylar and kapton can be lasered... it would be cool to make pcb stencils outselves10:40
wpwrakxiangfu: "oxidized dust" ... you don't have much dust at that point. what the steel wool removes is the oxide layer that builds on the entire surface.10:44
wpwrakxiangfu: (removing the paper) soaking it in water can be very counter-productive. many papers come off cleanly just by pulling. what water does it that it breaks the bond between the layers of the paper and you have to remove each layer separately, which is messy.10:46
wpwrakxiangfu: after paper removal, you still want to clean the board, though, because there can be some residues from whatever the "glossy" stuff is. just holding it under the water tap and rubbing it with a finger usually does the trick. a toothbrush is probably fine, too.10:48
wpwrakxiangfu: (etching) the last process is hydrogen peroxide (H2O2) with muriatic acid (HCl). advantages: works at room temperature, the etchant is a clear solution, the ingredients are very easy to find and cheap, and it's fast. disadvantage: releases small amounts of clorine gas, so you don't want anything metallic near it and good ventilation is important.10:52
wpwrakxiangfu: (removing the toner) you shouldn't need any sand paper. the acetone will get rid of everything. if you have problems getting acetone (its sale is regulated at some places since it's also used for some processes in the production of illegal drugs), you can try paint thinner/remover. pure acetone is better for your health, though.10:55
wpwrakxiangfu: (soldering) 315 C seems pretty high. is that with lead-free solder or leaded solder ? i rarely go above 270 C for components. (i go higher for large objects or for tinning the pcb surface)10:56
wpwrakxiangfu: and flux isn't optional ;-) you can also get flux in pens, which are probably more convenient to handle than a bottle10:57
wpwrakah, i saw your dispenser. yes, that should do, too10:57
wpwrakwhat kind of flux do you use ? there are RA (good but sticky), water soluble (not as good as RA but nicer to handle), and no-clean (not very effective in my experience)10:59
wpwrakwhatever the flux, you should clean the board afterwards. the chemicals in the flux can be quite conductive (< 100 kOhm/mm) and your circuit may act in very strange ways if there's flux on the board, particularly when still wet.11:00
xiangfuwpwrak, thanks11:13
xiangfuwhat is 'RA'?11:13
xiangfuthere are 3 type flux here.11:14
rohmaybe a weird question... .. but did making pcbs get expensive in china recently?11:14
xiangfusoldering paster, no-clean flux. another no-clean flux but the liquid is a little bit red11:15
xiangfuwoakas, updating the page.11:15
xiangfuroh, no. for learning and know what kind of board can made at home. what are not. :)11:15
wpwraki think "RA" is "Resin, Activated"11:16
wpwrakthere's also RMA "mildly activated". but so far, i treat them as the same. they're all sticky, which helps a bit with placing components, but your hards get quite dirty. "water soluble" is much friendlier.11:17
wpwrak(flux types) solder paste is not a flux. solder paste is solder ("tin") plus flux. you can also get solder paste with no-clean or water soluble flux.11:18
wpwraknot sure if they also make solder paste with RA/RMA. checking ...11:18
wpwrakno, apparently only water soluble and no-clean. which is fine. RA/RMA would be a mess :)11:20
wpwrakroh: don't forget turn-around times. 2 h vs. 2-3 days. and you can work on weekends/holidays :)11:21
kristianpaulxiangfu: the shops are in english?11:29
kristianpauli have no problem to chinesse (actually withs to know more) just a comment..11:29
kristianpauls/withs/wish/11:29
rohheh.. i see11:30
xiangfukristianpaul, no english. 11:34
kristianpaulno prob, brand actually seems to be more generic than i tought11:35
xiangfuwpwrak, updated page base on your command. also added th H2O2.  yes the H2O2 is faster. since the APS/etching box easy to buy. so I try the APS/etching box first. 11:36
xiangfuI have 2 type copper board. 0.8mm and 1.6mm. there are different color.11:37
xiangfuI mean the PCB color is different11:37
wpwrakit's nice that you have easy access to etching boxes. i wouldn't even know where to begin to look for them here ... and it would be expensive. luckily, an etching box isn't really needed for all this. i use glass kitchenware, with a plastic cover to prevent evaporation.11:38
wpwrakcolors can vary. that doens't necessarily mean anything.11:39
wpwrakhowever, there's also FR4 and pertinax. FR4 is epoxy-glass. pertinax is paper. the latter is cheap but lousy quality.11:39
wpwrak(paper) with some chemical treatment. so it has a hard surface and it quite stiff. but it's still paper.11:40
xiangfuand burn paper is much more easier. :)11:41
wpwrakoh yes. "FR" is for "Flame Retardant". there's probably a reason why they thought they had to emphasize this :)11:44
xiangfuthere are 2 other issues with 600DPI laser printer.12:06
xiangfuthe bga ball footprint is not a circle.12:08
xiangfusame width track on X or Y coordinate give different width after printing.12:09
wpwrakthat's a little odd. your printer shouldn't do that :)12:09
xiangfuespecially inside bga footprint.12:10
xiangfuwpwrak, checkout this : http://en.qi-hardware.com/wiki/File:Photo_paper.JPG12:10
wpwrakwell, if it's a general X/Y distortion. you could get quantization effects, which would be normal, given that 600 dpi = 1.6 mil12:12
wpwrakhmm yes, look a little distoted12:13
wpwrakare the large dimensions okay ? board heights to board width ?12:13
xiangfularge dimensions is okay.12:14
wpwrakwell, then there's probably not much to worry about. some small imperfections are to be expected at 600 dpi. besides, the rest of the process isn't all that precise either12:18
xiangfu(600 dpi = 1.6 mil). yes. the minimal track here is 6mil. 5V/GND track is 10mil.12:22
wpwrakso you have to expect best-case +/-13% variations, worst-case (if they use bad rounding) +/-26%12:33
rohhm...12:36
wpwrakthe hp 1320 i use for toner transfer has 1200 x 1200 dpi. i guess that's why i never even noticed :)12:37
rohwe got a big plotter recently... maybe that can be used for pcb purposes12:37
roha hp one.. for A012:37
roheats hpgl12:37
xiangfuhp1320 can buy 4 HL2140 :) 12:39
wpwrakxiangfu: it also has four times the resolution ;-)12:40
xiangfuyes. that is what I mean. :-)12:40
rohwpwrak: does your mill not also eat hpgl? how do you generate it? we played with inkcut (plugin to inkscape) .. the plotter is on a (usb) serial.. maybe one can plot directly to pcb and just etch then12:45
rohplot with etch-resistant ink. cut the transfer step12:46
wpwrakmy mill eats some bastardized hpgl12:47
rohahm.12:47
wpwraki generally work in gnuplot, when convert to the mill at the very end. machine parameters get inserted at that point.12:47
rohever put a pen in instead of a millhead?12:48
wpwrak(and my mill isn't the big mean kind, so the set of parameters is relatively benign. no need to worry about bringing seven tons of angry steel to a sudden halt :)12:48
wpwrakno, no pen yet. but pen + spring ought to work just fine12:49
rohjust thinking.. ours is a roll-plotter (120cm wide paper,, or with the rotary blade also vinyl)... but these pen plotters are dead cheap on ebay12:49
xiangfuwpwrak, hmm... SamsungML-1676 says "Up to 600 x 600 dpi (1200 dpi effective output)"?12:49
wpwraki'm not so sure about the ink, though. in my experience, etch-resistent whatever often doesn't12:49
xiangfuwpwrak, what is effective output?12:49
rohso i was thinking of getting another one, flatbed type12:49
wpwrakxiangfu: you'll have to ask samsung marketing about that ;-)12:49
roh .oO(we need more wallboards to store cnc machines on)12:50
wpwrak;-)12:50
xiangfuor I can print something, then compare. :-) ask samsung will be messy.12:52
rohanyway.. cool project. nice to see it documented so scientific. thats important12:53
wpwrakxiangfu: those "effective" parameters are usually marketing lies. they're basically saying that, in some cases, their 600 dpi printer produces results that look just like those of a real 1200 dpi printer would. but chances are that, if you actually count on having 1200 dpi, you'll just get 600 dpi.12:55
lindi-wpwrak: I have ML-2010, if you have some test page I can print it and take photo :)12:56
xiangfuyes. that is true.12:56
wpwrakit's similar for, say, oscilloscopes. some have an "effective bandwidth" that's sky-high. but only for perfectly periodic signals. in real life, what's causing the trouble are often the little anomalies, which you'll never find in such a mode.12:56
wpwrakxiangfu: there's your lab rat ;-)12:57
rohn812:57
xiangfugood night.12:58
xiangfulindi-, thanks. here is the file: http://downloads.openmobilefree.net/tmp/mini-slx9-ftg256-Front.ps12:58
xiangfulindi-, just print it to normal paper then take a photo is enough. 12:58
xiangfuML2010: 1,200×600dbi13:01
lindi-xiangfu: http://lindi.iki.fi/lindi/tmp/IMG_2403.JPG13:06
wolfspraulyes the atben/atusb are 'somewhere' :-) but not in the sharism.cc online shop right now, it's too much trouble (and too little demand) to add them there...13:13
xiangfulindi-, thanks. much better then my printer.13:13
xiangfulindi-, I will upload your picture to qi wiki.13:16
wpwrakwolfspraul: heh :) i indicated that sales@sharism.cc would be able to help in case anyone wants to buy some of them. does that sound reasonable ?13:17
wolfsprauloh sure!13:18
wpwrak(not sure if anyone will actually want any. but since it's been brought up, i'd at least like to send them to the right place.)13:18
wpwrakkewl. thanks :)13:18
wpwrak(printer) interesting difference. there were several traces where i thought you made contact with pads intentionally but not it's clear that you didn't :)13:20
wpwraks/not/now/13:21
xiangfuwpwrak, lindi- : http://en.qi-hardware.com/wiki/Mini-slx9#Printer_issues13:26
xiangfuthe bga ball is square under 600x600 DPI printer. :-(13:27
wpwrak"DO NOT USE photo paper that have glue-stuff at back" ;-)))13:31
wpwrakseems that you'll need a new printer quite soon anyway ;-)13:31
wpwrakthe track width difference is quite severe indeed. i think it's safe to say that the HL-2410 is rather unsuitable for any such work.13:32
xiangfu"seems that you'll need a new printer quite soon anyway ;-)" yes. I already do that. then I have to print ~10 paper for remove those glue form printer-inside. good news it's still working. but have some bad smell today.13:34
wpwrakin fact, the width difference that looks like -50%/+100% even slightly exceeds my "worst case". well, there's always rooms for a "worse than worst case" :)13:34
wpwrakat some point in time, the drum will get dirty. there isn't much you can do about that (except for replacing it or the drum/toner assembly)13:35
wpwrakfortunately, even a fairly dirty drum will still not cause a lot of trouble13:36
wpwrakbut that issue is why i do my toner transfer on the hp 1320 and keep my much nicer p2055 clean.13:37
xiangfuwpwrak, I guess the photo paper also leave some stuff somewhere.13:38
xiangfu(glue on back)  I can see dirty on even on those outside-small-wheels.13:39
xiangfuthen I know the laser printer works at ~160C degree when printing13:39
xiangfu1200x1200DPI printer cost more money. >2000RMB(~350USD) here.  600x600DPI printer ~700RMB(~130USD)13:45
lindi-this ml-2010 was very cheap13:46
xiangfuthat is why they stop production ml-2010. :-)13:47
lindi-definitely in the 130 USD range13:47
e2580we have launched our open source hardware encrypted storage device on indiegogo who wants to see it22:11
viricthis is a bot22:11
virice2580: help22:11
roh*yawn*22:12
e2580help ?22:12
virice2580: part22:12
e2580http://www.indiegogo.com/CryptX222:13
rohstop spamming or we will.22:13
e2580not sure if you are asking for this link or what22:13
wpwrake2580: it is a little odd to pitch a device you can't export to an audience that's largely located in what your country considers enemy territory :)22:51
wpwrak(germany, spain, russia, china, columbia, argentina, to name a few)22:53
e2580i didint write the rules, but since i live here i have to follow them22:53
e2580i wish i was able to sell worldwide22:54
e2580we are giving the info for free worldwide :)22:54
wpwrakpity that you couldn't find a solution that has no such restrictions, e.g., with a cpld or such23:17
--- Thu Nov 15 201200:00

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