#qi-hardware IRC log for Saturday, 2012-05-26

luis_hola gente07:57
pabs3huh, tuxbrain closing their shop08:08
viricoh10:04
wolfspraulyes, tuxbrain was one amazing guy to buy stuff from, now it's in distributor heaven :-)12:45
wolfspraulhttp://www.tuxbrain.com/en/content/we-are-fulminating-stock-cheap-cheap-aka-tuxbrains-reborn12:46
kristianpauloh12:49
valhallathis is sad indeed12:50
wpwrakhe was on #qi-hw a few days ago saying that he had good news, but never said what. so maybe he's not so sad about what he's doing.13:14
kristianpaulhe mention word reborn13:15
wpwrakyeah. also, i think running a shop is 99% terribly boring work. better to focus more on what's actually fun.13:16
kristianpaulboring and full of details and customer support :)13:16
wolfsprauldoes anybody know the state of migen actually?13:19
wolfspraulI should ask in #milkymist of course but it feels stressful sometimes :-)13:19
wolfspraulI hope we can switch to it and it's really superior, who knows. maybe another few months?13:19
wolfsprauldon't know13:19
kristianpauli just see uart and ethernet cores13:19
kristianpaulhavent tried sinthesize it yet..13:20
wolfspraulhmm13:20
wolfspraulmake that 6 months13:20
wolfspraul:-)13:20
wolfspraulkristianpaul: will the m1-gps-expansion board fit on rc3 as well?13:20
kristianpaulsebastien pasted something about ddr but i dint look like a speed test or such13:21
wolfspraulmaybe we can go ahead and just make some of them, once I have my diy process figured out I need some jobs...13:21
kristianpaulsure may be no fit, but can be adapted13:21
wolfspraulyeah. in the schematics it shouldn't matter much13:21
kristianpaulcurrent layout isnt friendlier either for rc2 :)13:22
wolfsprauldo you think you can bring them to a 'kind of complete' state at some point?13:22
wolfspraulI'm not following it closely13:22
kristianpauli havent touched layout yet13:22
kristianpauland after the cut off it should work yes i just removed the fgpa and cypress tranceiver we dont need13:23
kristianpaulonce sch is finished i'll move to bom13:25
wolfspraulyes great13:26
kristianpaulwolfspraul: current design dont fit, it need wires13:26
wolfspraulI think my first board targets will be something like atben clone, combine small xc6slx9 with atrf to configure over RF, m1-gps-expansion board13:27
wolfspraulstuff like that13:27
kristianpaulmaxim ic is very similar to the avr chip used in atben13:27
wolfsprauland sure maybe it's too ambitious. then I try to light an LED with a series resistor instead :-)13:27
kristianpaul;-)13:27
wolfspraulworst case I just press the wall button here to turn on my desk light, if all else fails that needs to give me some confidence about my mastership of *electronics*13:28
wolfspraulfrom there to the ghz circuit should only be a small step13:28
kristianpaulhow etching the PCB went finally? i lost track of the thread13:32
kristianpauli tried that  two weeks ago, wpwrak etching solution is  powerfull, so match than some errors on the tonner transfer ended "eating" all coper13:35
wpwrak;-)))13:35
wpwrakmaybe your peroxide was a bit too strong13:35
kristianpaulmatch/much13:35
kristianpauli need improve that tonner transfer13:36
kristianpaulafaik i dont used that hp paper, instead re-used sheets from a local magazine13:36
kristianpaulmay be the pcb wast clean enought.. i dont know13:37
wpwrake.g., if you use 30% peroxide, it'll etch your board clean in about half a minute. that's an quite exothermic reaction, too. so after a little while, the acid starts to boil and fume. so patience is advised :)13:37
kristianpauldo you use a cloth iron for transfer right?13:38
wpwraki use a "plastificadora"13:38
wolfspraulI haven't started yet, too many other issues. I will take the next step starting from mid next week, together with xiangfu probably13:38
kristianpauloh13:39
wolfspraulbut 100% on track for this, definitely13:39
wpwrakbut a clothes iron would do, too. it's just less reliable than the machine.13:39
kristianpaulthats fancy :)13:39
kristianpaulless reliable indeed..13:39
kristianpaulis the only tool i dislike from this process, i dont feel confident with it13:39
wpwrak(less reliable because it's a manual process. some people are great at it, though. ironically, women tend to be very good at handling toner transfer with an iron. they produce better results than my little machine. maybe there's something right about those stereotypes :)13:40
wpwrakwolfspraul: the first experiment should be a test of the acid. no circuit, just a snippet of the pcb. see how it works.13:42
kristianpaulthat is fun :-)13:42
wpwrakwolfspraul: you can add some patterns with a water-resistant pen, too. that will allow you to find a pen that provides a reliable cover, to repair flaws in the toner transfer.13:43
wolfspraulsure, joerg mentioned that already13:44
wolfspraulsaved in brain13:44
wolfspraulof course I will experiment, I am quite confident that after a few days this is sorted out13:44
kristianpaulwpwrak: ok, so you just put together the PCB and printed layout and let the "plastificador" do its job?13:44
wolfspraulthe main difficulty seems to be the toner transfer, that sounds like it has some unknown variables inside13:45
wolfspraulbut I'm not worried13:45
wpwrakkristianpaul: if you acid is too aggressive (e.g., if the etching finishes in less than, say, 5 minutes), you can add a bit of demineralized/distilled water to dilute it, which slows it down13:45
wpwraks/you ac/your ac13:45
kristianpaulget a plastificadora, it seems make it more easilly and confortable to use13:45
kristianpaulwpwrak: yup i took 3 minutes..13:45
kristianpauls/i/it13:45
wpwrakkristianpaul: you need to modify the plastificadora. 1) they usually don't allow passage of such thick material and 2) they're too fast13:46
wolfspraulah distilled water, good idea to get in advance13:47
wpwrakwolfspraul: the key to successful toner transfer is finding a good paper. expect to spend some time with that. that hp paper is perfect but may be hard to find.13:47
wolfspraulfor the toner transfer, I would feel better if it were a better understood chemical thing13:48
wolfspraulsure, I know [hp paper]13:48
wolfsprauljust what is 'better' and why do some work and some not so well and some not at all13:48
wolfspraulbut ok, no problem13:48
wolfspraulI will try with a variety of papers and something will work13:48
wpwrakdemineralized water should be available at any place that has stuff for cars. just make sure it has no additives (anti-corrosive, anti-freeze, whatever)13:48
kristianpaulhmm13:50
wpwrak"better" fewer flaws. flaws come in many shapes: can be that traces "blur" (-> too hot). can be that the toner doesn't stick to the pcb (so some thing come off when you remove the paper; usually a sign of insufficient heat/pressure, but can also paper chemistry). can be that there are holes/gaps inside of toner regions (-> usually a sign of outgassing)13:51
wpwrakit's imperative to do a careful visual check of the board after toner transfer. you can fix small issues manually. e.g., if a trace is broken or looks too weak, you can fix this with the water-resistant pen. if toner has blurred into an adjacent pattern, you can scratch off the toner with tweezers / knife / etc.13:54
wolfspraulare there flat laser printers where you can just pull a thin pcb right through it?13:54
wolfspraulor if a transfer is needed, maybe we can transfer via some other material that works more reliably and is still easy to source?13:55
wolfspraulfor example I could go through the variety of kitchen papers and foils, some of them have amazing characteristics and easily available13:55
wpwraka very thin pcb, maybe. but most likely, they need to change the drum after a very small number of runs. so that might be a bit expensive.13:59
wpwrakalternative papers are possible, yes. you just need to experiment a bit.14:00
wpwrak(change the drum) because i would expect it to get scratched. but i'm too coward to try that :)14:02
wpwrakyour printer's specification should provide a hint of what thickness of materials it can handle in general14:03
qi-bot[commit] Werner Almesberger: add floor() function (master) http://qi-hw.com/p/fped/9a6c7d217:14
qwebirc89599h19:45
--- Sun May 27 201200:00

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