#qi-hardware IRC log for Wednesday, 2012-05-16

wolfspraulwpwrak: thanks again for taking the time to go through the pcb process yesterday!00:54
wolfspraulmade me happy, I am thinking about buying things now and experimenting :-)00:54
wpwrakwas a pleasure :) and you'll have fun, too ;-)00:55
wolfspraulI start easy with some of the smaller utentils, and then I go to an electronics market with xiangfu for some of the more elaborate things, especially also to talk with guys and see what they think and/or use00:55
wolfspraulI will drag xiangfu into this right away as well00:55
wpwrakelaborate things = scope, ultrasonic cleaner, ... ?00:55
wpwrakand let's not forget the CNC :)00:55
wolfspraulyah, basically. also chatting with the guys there00:55
wolfspraullaser printers cost as little as 80 USD now, amazing00:57
wolfspraulI was wondering whether one could directly print onto the pcb?00:57
wpwrakthere have been attempts. with ink, though.00:58
wolfspraulhow about other printers?00:58
wolfspraulah ok, but you think the toner transfer is a good standard still, well known to work...00:58
wpwrakthere are basically three approaches:00:58
wolfspraulbut maybe a laser printer is so flat that it can pull a 0.8mm (or thinner) pcb right through it? :-)00:58
wolfspraulwell I can imagine that there are potentially infinite ways how to get this printed00:59
wpwrak1) lithography, with UV light and photo-coated PCBs. popular but a bit messy. never tried that. that's probably still what industry uses.00:59
wpwrak2) toner transfer. less capable but reasonably straightforward.01:00
wpwrak3) milling of the non-copper areas01:00
wolfspraulso you apply the etch resist (positive or negative), then uv light to remove parts, then etch?01:00
wolfspraulyeah I'll just go with the toner transfer first01:00
wolfspraulfeels right and uses only cheap tools and simple chemicals and steps01:01
wolfspraulexcept for your stepper-motor laminator, which I hope to bypass01:01
wpwrak3) should be cool, because you avoid the chemistry and it's just one step. drawbacks: the finer your traces, the easier your mills break. and you need very precise depth control. i experimented a bit with this but eventually gave up. never succeeded in making a board that actually looked right.01:02
wolfspraulnah I don't feel good about 301:02
wolfspraullithography or lasering, chemicals - that sounds better01:03
wpwrak(photo process) you make a transparent mask, then expose with UV, then develop (chemically), then remove the developed/undeveloped part (not sure if you need special chemicals for that), then etch, then remove all the photoresist (chemically)01:04
wpwraka heck of a pipeline01:04
wolfspraulmaybe you can print the mask with a regular printer just like for those overhead projectors01:04
wolfspraulwould need someone who still has a simple lab for analog photo development01:05
wpwrakand it also complicates the handling of the PCB material, since there's the photosensitive coating to worry about01:05
wpwrakyou don't need the entire lab. just a few bits of it.01:05
wpwrakbut still seems messy to me :)01:06
kristianpaul80usd !!03:13
kristianpaultoner transfer, indeed is nice, but is not one step solution, so more probably to make mistakes03:15
kristianpaulat least thats my experience03:15
kristianpaultought laser printing tech had improve a lot, now i should try again :)03:15
kristianpaul3) is messy, even more mechanical posibillities to get a failure...03:17
wolfspraulyou tried toner transfer before?03:27
wolfspraulbut it seems you have no idea for a better process?03:28
kristianpaulyes i did when in the University03:29
kristianpaulto be honest i tried no more than 5 times, then decide outsource :)03:30
kristianpaulno idea for better indeed03:30
kristianpaulwhen i tried laser printed was new.. all was most ink so...03:30
kristianpaulresolution wasnt the best03:31
kristianpaulalso i dint knew about the mix of chemicals werner pointed, thats is a cleaner solution03:31
wolfspraulwpwrak: so maybe on my initial purchase, 100ml h2o2 and 100ml hcl is enough?05:34
wolfspraulor more?05:35
wpwrak100 ml of each should be sufficient for several boards. but you may not be able to buy in such small quantities. especially not the HCl.05:50
wpwrakthere, they typical minimum quantity would be more like 0.5-1 l05:50
wpwrak1 l in easy.com.ar :) peroxide i can get in bottles of 0.1 l - 0.5 l at a pharmacy06:07
wpwrakand don't be shocked by the prices. all that stuff is cheap.06:11
wolfspraulhah yes, went as expected. I got hcl and h202, but at unknown concentrations :-)07:35
wolfspraulwhen pressed, they would try to come up with the number they thought you would like to hear07:35
wolfspraullabeling of the bottles is chaotic, multiple labels, hand-written, refilled etc. etc. fun.07:36
wolfspraulI was upgraded from "wild blob of duct tape" to a proper rubber knob closure of the bottles only after complaining :-)07:37
wolfspraulso who knows, I will find out the concentration "in the field", china style07:37
wolfspraulthe hcl is of some greenish color, and the seller told me I shouldn't spill it over my hands. that's a good sign, must be the right stuff...07:38
wolfspraulmaybe there is a standardized and easy test which would let me guess the concentration of the hcl and h2o2? I think I will just try etching instead, see whether anything happens.07:40
pabs3green hcl? /me guesses it is contaminated, maybe with copper09:41
whitequarkwolfspra1l: that's some really bad customer service09:55
whitequarkwhen I bought chemicals on a market here in Moscow, behind the counter there were two nice girls _and_ they knew more about pcbs than I do now09:56
whitequarknot to say that bottles etc. were of uncomparable quality. strange. I thought it should be good in China as it's an industrial center and whatnot, no?09:56
whitequarkpabs3: or complex ions of Fe, they're green in some cases too09:58
whitequarkcopper is mostly blue as I've seen it...09:59
mthoutdoor bronze turns green, isn't that oxidized copper?10:03
mthwikipedia says it's a carbonate instead: http://en.wikipedia.org/wiki/Patina#Acquired_patina10:04
mthand other non-oxide molecules10:05
whitequarkyeah, Copper(I) oxide is red: https://en.wikipedia.org/wiki/Copper(I)_oxide10:07
whitequarkand Copper(II) oxide is, erm10:08
whitequarkit's kinda bluish10:08
whitequarkas you can see, sulphate is blue: https://en.wikipedia.org/wiki/Copper_sulphate10:08
whitequarkso will be any other dissolved compound when copper ions are only coordinated by water molecules10:09
LunaVoraxHello everyone!10:30
qi-bot[commit] Adam Wang: js28f256j3f105.lib: pin 43, pin name changed to VCCQ (master) http://qi-hw.com/p/kicad-libs/bbab96311:51
qi-bot[commit] Adam Wang: Merge branch 'master' of projects.qi-hardware.com:kicad-libs (master) http://qi-hw.com/p/kicad-libs/4be16e311:51
wpwrakwolfspraul: green is a bit unusual. should be more like colorless or perhaps a little yellowish. but maybe it's an additive. that sometimes happens. is it still transparent enough to see through a cm of it ?12:07
wolfspraulyes definitely12:08
wolfspraulconcentration no idea12:08
wolfsprauland for the h2o2 I'm not even sure whether it is h2o2 :-)12:08
wolfspraulthe pharmacy had a more trustworthy h2o2, but concentration was listed at 2.5%-3.5% only12:09
wpwrakyou can try this: cut a bit off a PCB. make you mixture. WEAR GLOVES AND GOGGLES. set up the stuff outside, somewhere where spills don't harm. even if it would decide to boil.12:09
wolfspraulis there an easy way to reduce h2o2 to increase the concentration?12:09
wpwrakweark clothes you don't mind discarding. if you have some plastic poncho, raincoat, or such, wear that too.12:09
wpwrakthen toss the copper inside and see what happens.12:10
wpwrak(reduce concentration) yes, add water :)12:10
wolfspraulneed to increase the concentration12:11
wpwrakoh, increase12:11
wpwraknaw, not really12:11
wolfspraulthe h2o2 at the pharmacy is only 2.5%-3.5%12:11
wpwrakthat's probably good enough. lemme check what the one i have is ...12:11
wolfspraulyes I will just try, the main protection being that I will try with small amounts12:13
wolfspraulcome on everybody here uses this stuff all over, I won't make myself too laughable :-)12:14
wpwrakhmm. bottle says "10 volumenes". 10 % ? let's find out what the google scienctists has to say about it ...12:14
wpwrakwith the peroxide concentration known to be low, you're already on much safer ground. when you wrote that it's all weird, i thought, the peroxide was unknown too12:15
wolfspraulit is, I got the one in the market, not the pharmacy12:15
wolfspraulcould be anything :-)12:15
wolfspraulI will get the one from the pharmacy too, good idea12:15
wpwrak(if you mix, say, 35% HCl with 30% H2Os, you get a VERY reaction-happy acid. etches a board clean in just a few seconds. alas, it also starts boiling quickly ...)12:16
Action: wpwrak is still in pre-caffeine state12:16
wolfspraulsure but most likely the stuff I buy here errs on the side of water12:16
wolfspraulnot on the other side12:16
wolfspraulI will try a little :-)12:16
wolfspraulthere's no way I can get high-reliability chemicals anyway here12:17
wpwrakgood :) so the chinese are not all _that_ keen on massively negative population growth ;-)12:17
wpwraksince we don't know for sure what the acid is, you should follow the usual precautions: give peroxide into the basin first, then add the acid. that way, the risk of splashes with high concentration is lower12:18
wpwrakyou should still wear gloves and glasses/goggles, though12:19
wpwrak(peroxide) "10 volumenes" = 3%. so yours is perfect.12:20
wolfspraulah ok12:20
wpwrakso only the HCl is unknown. during etching, a bit of Cl gas should escape. it has a very characteristic smell. so that shuold allow you to tell tht there's indeed Cl in there12:22
wpwrakso, first experiment: make a bit of acid and toss a piece of copper-clad PCB inside. 1 cm^2 will do nicely12:22
wpwrakthen watch what happens. move it around a bit with a stick made of plastic, wood, or glass12:23
wolfspraulsure, I remember from my visit :-)12:23
wpwrakif it's double-sided, you can also turn it over, move it for a while "bottom up", then flip, etc.12:23
wpwrakyou should see that both sides etch at different speeds. that's probably because of differences in the flow of the acid in the basin. i find that usually the bottom side etches faster.12:24
wpwrakthe chemical reaction should produce 1) little gas bubbles on the copper. 2) bluish/greenish traces that come off the board. the blue are your copper ions. the green is them with chloride.12:27
wpwrakpity that your HCl already has that color. but you may still see some changes. eventually, the whole solution turns green-blueish, even if you start with "clean" HCl.12:28
wpwrakcomplete removal of all the copper should take between 10-20 minutes.12:29
wpwrakah, maybe give your pcb a scrub with steel wool first. not sure how much the acid likes the copper oxide. might slow it down.12:30
wpwrak(steel wool and then clean with alcohol. wouldn't want to add iron to the reaction.)12:33
wolfspraulgood good12:35
qi-bot[commit] Maarten ter Huurne: MIPS: JZ4740: SLCD: Fixed sysfs tvout value parsing. (jz-3.3) http://qi-hw.com/p/qi-kernel/f20d89613:28
qi-bot[commit] Maarten ter Huurne: MIPS: JZ4740: SLCD: Ignore sysfs rewrite of current TV-out mode (jz-3.3) http://qi-hw.com/p/qi-kernel/0a96d7213:28
canosohave anyone used gnublin embedded GNU/Linux?21:31
GNUtoo-desktopcanoso, what's that?21:34
canosoit is a embedded linux board with an ARM microcontroller.21:35
canosoIt is open hardware21:35
GNUtoo-desktopis it 100% free software?21:45
GNUtoo-desktopseem so at first sight21:46
GNUtoo-desktopbtw how much does it cost?21:46
canosoabout 50 euros21:46
lindi-wrong language :(21:47
GNUtoo-desktopah ok21:47
canosoit has USB On the Go. Can I connect any usb device to it?21:48
lindi-canoso: I wouldn't call that a microcontroller21:48
lindi-since linux doesn't run on microcontrollers :)21:49
GNUtoo-desktoplinux runs on blackfin for instance21:49
GNUtoo-desktopor stuff without mmu21:49
GNUtoo-desktopbut in that case it's a SOC21:49
lindi-and it has 8 MB of RAM21:50
GNUtoo-desktopI think the minimum specs is 4M21:50
canosolindi-: http://www.nxp.com/products/microcontrollers/arm9/lpc3100/#overview21:50
canosolindi-: it is a microcontroller21:50
canosolindi-: it has many peripherals integrated in the same chip21:50
canosolindi-: and RAM in the chip21:50
lindi-openmoko has ram in the chip too21:51
lindi-(half of it)21:51
lindi-canoso: I guess we just have different terminology :)21:51
GNUtoo-desktopmicrocontroller and SOC are not that different21:51
lindi-for me microcontroller is something without MMU certainly and without cache too21:51
GNUtoo-desktoplinux run on some stuff without MMU21:52
lindi-yeah but that's quite an exception21:52
kristianpaulmilkymist soc dont have mmu btw :-)21:52
GNUtoo-desktoplindi-, an exception that is in mainline?21:52
lindi-GNUtoo-desktop: yeah :)21:54
wolfspraulcanoso: [gnublin] nice link, thanks a lot!23:41
wolfspraulfirst time I see this, and please always feel free to post links to new or potentially new open hardware anything here23:41
wolfspraulwe love to learn23:41
wolfspraulI haven't checked about license, tools, etc. but anyway, nice simple project.23:42
GNUtoo-desktopyes indeed seem nice23:46
wpwraklindi-: linux runs fine on microcontrollers. here's an example: http://www.bit-tech.net/news/hardware/2012/03/30/linux-atmel-microcontrollers/123:46
canosowolfspraul: ty23:48
--- Thu May 17 201200:00

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