#qi-hardware IRC log for Tuesday, 2010-09-21

kristianpaulwpwrak: 13 tall is okay for you?00:09
kristianpaullet me take a pic00:10
wolfspraulwpwrak: 300 mA on SD VDD should be fine, here's a rough calculation done in the past  http://lists.en.qi-hardware.com/pipermail/discussion/2010-May/004414.html00:10
kristianpaulwpwrak: http://en.qi-hardware.com/wiki/File:Pantallazo-6.png00:13
kristianpaulhmm still to big i'll try 8mm00:19
kristianpaulwpwrak: check again is 8mm tall in think i'll split and print it00:23
kristianpauldamm i always hate why heekscad dont let me move in less that one mm00:39
kristianpaulwpwrak: check last version of the pantallazo :)00:57
kristianpaulany way i'll print now00:57
kristianpaulin RED :p00:57
Action: kristianpaul printing the lid01:08
Action: kristianpaul printing the body01:13
Action: kristianpaul done printin now taking pictures01:18
wpwrakkristianpaul: wow, that's quick01:24
wpwrak8 mm is still a lot. do the horizontal surfaces have to be that thick ? in total, you shouldn't need more than about 4.5 mm ...01:24
kristianpaul2mm thick horizontal01:29
kristianpaul1mm vertical01:29
kristianpaulwait some pics01:29
kristianpaulokay i try do smaller tomorrow01:29
wpwrakwolfspraul: 300 mA is better than i expected. nice :)01:31
kristianpaulhttp://en.qi-hardware.com/wiki/File:IMG_2666.JPG  http://en.qi-hardware.com/wiki/File:IMG_2668.JPG01:33
kristianpaulwpwrak: my closed soure gps pull 200mA with no conplain i should add to what wolfspraul  pointed01:34
wpwrakhehe, the simplified pcb is cute :-)01:34
kristianpaulhttp://en.qi-hardware.com/wiki/File:IMG_2670.JPG http://en.qi-hardware.com/wiki/File:IMG_2671.JPG01:35
kristianpauland thats it01:35
wpwraknice how it fits !01:35
kristianpaulokay i'm  off bed i try 4.5 mm tomowor01:35
wpwraktill tomorrow ! looks promising, thanks !01:36
qi-bot[commit] Werner Almesberger: ben-baseframe-bottom-100um is done. http://qi-hw.com/p/ben-scans/2cf011d03:10
tuxbrain_awaywpwrak , kristianpaul : can I suggest a protuberance, hook li in that wpan case? , to "anchor" it to the hole "supposedly thinked to attach a strap", this way we can remove some mechanical stress to the uSD attachment03:34
tuxbrain_awayhook li ->"hook like"03:35
wpwraktuxbrain_away: you mean a little "nose" that goes into the hole ?03:38
wpwrakhmm, that hole is about 4 mm below the plane of the pcb. quite far.03:39
tuxbrain_awaysure it will not remove all , but torsion will be less a risk ...03:39
tuxbrain_awaywell plastic is plastic, and the mould should be as expensive with that nose or not03:40
wpwrakone of my suggestions for the ya i had queued up would be two holes above the sd card plane where one could insert 100 mil headers. that would be pretty solid.03:41
tuxbrain_awaygood thinking :)03:41
tuxbrain_awaybut I'm thinking more on an accesory for actual ben...03:42
wpwrakah, the nose could be somewhat difficult to make. for injection, it would at least need a much deeper mold (you'd have to rotate the bottom plate)03:43
wpwraknot sure about 3d printing. may be possible with a filler. depends a bit on the feature size.03:44
wpwrak(injection) and if rotating the bottom plate, it can't be flat anymore. but that may be okay in this case. could even look better.03:45
wpwrakmakes the mold hard to mill, though. so you'd probably need a 3-part mold.03:46
wpwrak(hard to mill) because endmills that go deep enough would be relatively thick.03:47
tuxbrain_awaywell is matter to see how much it cost and if worth the meaning03:49
tuxbrain_awayif it not protecct to much, maybe is just pointless such effort03:50
tuxbrain_awaybut I'm worried about the length of pcb and the tinny contact surface of attachement03:51
wpwrakyes, the mechanical stability isn't so nice. attaching to uSD is a great improvement upon having to cut up the lcd's pcb and running wires between the shells, but it's not perfect.03:53
tuxbrain_awayany accidental push on the extreme and "clack" another sad  geek crying for his broken toy03:53
wpwrakone more sale for you ;-)03:53
tuxbrain_awayI only want happy geeks on my shop03:54
tuxbrain_awayI agree on the avoid on soldering to hardware hack is a really great improvement in deed :)03:54
tuxbrain_awaythis can remove the Warranty removal advice on my  future posts :)03:55
tuxbrain_awaybtw any one has attempted a soft serial by uSD yet?03:56
wpwrakyou mean rs232 ?03:57
tuxbrain_awayno 3V3 uart03:59
wpwrakyeah, that's what i meant. the rest is just level shifting :) hmm, would be tricky.04:00
wpwrakyou could do it at high speed but tx-only with low duty cycle.04:01
wpwrakor at low speed. the problem is that you'd need to get an interrupt for each bit/sample. that's quite a lot, even at moderate speeds.04:01
wpwrakotherwise, you'd have to busy-loop until the byte is in or out.04:02
tuxbrain_away9600 should be enough04:02
wpwrakah, scratch tx-only. rx with low duty cycle would also be okay.04:02
tuxbrain_awaywhy that difference in writting reading?04:03
wpwrakhmm, that's about 20 kHz interrupts or more.04:03
wpwraknaw, tx and rx should be the same. i didn't think that you could just interrupt on the start bit.04:03
wpwrakyou need extremely tight interrupt latency, though. that's usually not a problem, but sometimes it is. so whenever your interrupt is late, you have some data loss or corruption.04:05
wpwrakof course, if your cpu is dedicated full-time to the "soft uart", then you don't have timing problems. you just busy loop.04:06
tuxbrain_awayI have some C bit banging serial code example from carlos from my early serial attempts, never used it, and actually I have scarce time to investigate, I will try to recover it , and think how adapt it to the uSD pins04:07
wpwraki would just add a cheap microcontroller that takes care of the uart timing :) you can feed it a very accurate clock from the ben, so you don't need a crystal.04:08
tuxbrain_awayhehehe you have by passed my electronic skill by far with this sentence :P04:09
wpwrakah, look at the atusd schematics and you'll see :-)04:10
wpwrakfor quick access, http://www.almesberger.net/misc/ben/demo4/04:11
wpwrakCLK_16M is provided by the SD/MMC clock, which I set to 16 MHz04:12
wpwrakthat way, i don't need a crystal. works quite nicely.04:13
tuxbrain_awayit can reach 56700 with this approach?04:13
tuxbrain_awayalso is that microcontroller aviable in more hobbist friendly format?04:13
wpwrakif you have a chip that implements a uart in hardware, sure. you typically need something like 4x or 16x the bit rate. so that would be at least 1 Mbps, if your circuit plays along04:14
wpwrakerr, what would "hobbyist friendly mean" in this case ? :)04:14
wpwraks/ mean"/" mean/04:14
tuxbrain_awayeasy to put in a breadboard :)04:14
tuxbrain_awayno need to SMD04:15
wpwrakbut where's the fun ? :)04:15
wpwraksure, there are tons of pics and atmels in dip packages04:15
wolfspra1ltuxbrain_away: what is a nice breadboard size in your opinion?04:15
wolfspra1lto go along with the breakout cable...04:15
tuxbrain_awaymmm what you mean by breadboard , the soldering one or the lot of holes plastic pluggable one04:16
tuxbrain_awayin the eschema what is the difference of AGND and DGND grounds?04:19
tuxbrain_awaybtw atusd is the hole schema for the wpan thing? is so "easy" to do?04:21
wpwrakagnd and dgnd aren't actually good concepts here. i copied that from the reference design. however, atmel themselves actually recommend four ground areas.04:22
wpwrakyes, the circuit is very simple :)04:23
tuxbrain_awayThat a good thing :)04:24
wpwrakwhat's a bit tricky is the rf side, especially the antenna tuning. the antenna design i'm using is for a thicker board. so i need to adjust the antenna. unfortunately, they don't say how. so it'll be trial and error ...04:25
tuxbrain_awayyes AFAIK RF is even black magic to experts on such matter :)04:26
tuxbrain_awaythey do maths yes but at the end a lot of tolerance has to be added to results :)04:26
wpwrakyeah. and antenna tuning should be among the best in that regard. even adding a bit of plastic (e.g., a case) can change things quite dramatically04:27
wolfspra1ltuxbrain_away: I mean the "lot of holes plastic pluggable one"04:29
wolfspra1ldo you think a soldering breadboard is better?04:29
wpwrakwolfspra1l: i think this is the kind rikard used: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=923273-ND04:32
wpwraknote the price even for such a small one04:32
wpwrakyou would probably also have to include a set of jumper wires04:33
tuxbrain_awaywolfspra1l: well I think they have very different funtions and on this case the soldering one not have too much sense due they are pretended to have hobbist final results , and due the size it will have will be no sense to attachech on uSD04:33
tuxbrain_awayI have one in stock :)04:33
tuxbrain_awaynot of that model I find thos thinny ones too much tinny04:33
wpwrakhere's a better kit. larger and with wires. http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=438-1046-ND04:34
tuxbrain_awayif you circuit has some more than just one no to much complicated mc04:34
tuxbrain_awayhey cammon don't point to competence :P http://www.tuxbrain.com/oscommerce/products/7504:35
wpwrakah, you have them too. good :)04:36
tuxbrain_awayyes we do a mass buy directly to china manufacturer :)04:36
tuxbrain_awayregarding the tinny model, I try it but when your circuit is a bit more complicated than just plug one microcontroler you miss a lot of holes :)04:38
tuxbrain_awayso I find this model quite confortable and unexepensive04:39
wpwrakyeah, i thought more in the sense of a proof of concept. but you're right, a larger board has almost the same price.04:39
tuxbrain_awaybtw any clue when we could have some breakout uSD adaptors as first Ben NN complement?04:41
wolfspra1lI try to make some, but cannot commit to a date right now04:42
wpwrakwolfspra1l: what cost do you expect ?04:42
wolfspra1lfirst 10 maybe 15 USD/piece04:43
wolfspra1lI will probably do ribbon cable, not fpc04:43
tuxbrain_awaythat make sense.. less stress on the connector04:43
wolfspra1ltuxbrain_away: so I understand you have enough breadboard stuff already, I don't need to worry about it. great!04:43
wolfspra1ltuxbrain_away: wow high or wow low?04:44
tuxbrain_awaywow high04:44
wpwrakyup. seems quite expensive.04:44
wolfspra1lgood thing everybody here is used to work for free :-)04:44
wpwrakwolfspra1l: are you sure you asked factory.cn, not factory.ch ? :)04:44
tuxbrain_awayhehehehe  you know the pourest multinational company qi-hardware :P04:46
tuxbrain_awayno comunism no capitalism no money :P04:47
wpwraktuxbrain_away: they'll have an article about wolfgang in Exploit - The Magazine for the Globalized Entrepreneur, "How to pay less than the Chinese" :)04:47
wpwraktuxbrain_away: ;-))04:48
Last message repeated 1 time(s).04:49
wolfspra1lwpwrak: right.04:50
wolfspra1ldo you want to make microsd breakout cables for 10 USD / piece?04:50
tuxbrain_awayok ok let's focus (a bit) I understand this 10 are prototipes so price is .... "justified" but price expected for a higher number let's say 20004:50
wolfspra1lplease don't forget the fuse. also proper testing of each one. when it's done ship to David.04:50
wolfspra1ltuxbrain_away: come on in large quantities, in electronics, it's all zero.04:50
wolfspra1lall costs are just amortizations of one-time investements04:51
wolfspra1lthe silicon and plastic and bit of metals is all near zero04:51
wolfspra1lso sure, for 100 it's already < 10 USD for sure04:51
wolfspra1lfor 1000 even less, etc.04:51
wolfspra1llook at the total04:51
wolfspra1l10 * 15 = 150 USD04:51
wolfspra1lthink about the amount of work, counting really everything not like us here working 16h/day for free04:52
tuxbrain_awayno no please the 0 asymptote discuse again not :P04:52
wpwrakoh, you're making only 15 ?04:52
wolfspra1lfirst I make 10, see the quality04:52
wolfspra1lsend some people, then go from there04:52
wolfspra1lso for 100, let's say the price is down to 8 USD, that's 8 * 100 = 800 USD04:52
wolfspra1lfor 1 million, I'm sure it can be done for 1 USD / piece :-)04:53
wpwrakfor 1M it should be cents :)04:53
wolfspra1lI just bought some pouches (David will send you some), costing 15 US cents / piece04:53
tuxbrain_awaymmm, I'm worried about selling price of breakout, more than 10¬ will be no joy04:54
wolfspra1lrelax guys04:54
wolfspra1lyou asked about price, I gave you one. I am paying.04:54
tuxbrain_awayrelax you too:) , I'm just doing projections about the data I have04:55
Ornotermeswpwrak: the solderless board on dikikey looks quite overpriced04:55
wpwrakOrnotermes: they're not cheap. well, the bigger one has roughly the same price as tuxbrain's.04:56
tuxbrain_awayOrnotermes: don't think so, this bitches is quite expensive even buyed in large numbers04:56
tuxbrain_awayat least if you want a quite bit of quality04:57
wolfspra1laha :-)04:57
wolfspra1ltuxbrain is reality man...04:57
Ornotermesthe small one i have cost like USD 4.25+vat for two pcs in sweden04:58
tuxbrain_awayyes, I even see the small one cheaper, I was thinking on the larger one04:59
tuxbrain_awaybut as I said I found the smaller less practical04:59
Ornotermesthe bigger looks more renable04:59
wpwrakUSD 3/each sounds more like what i imagined for the small one04:59
wpwraki was thinking of a combination of cheap cable plus small board. of course, if the cable is so expensive ...05:00
Ornotermesi buy most things here: http://swechtrading.se/zencart/index.php?main_page=index&cPath=44 (a USD cost about 7.31 SEK)05:02
Ornotermesand their prices inclute VAT05:02
tuxbrain_awayI think I should offer the breadboard+wires+breakout for 20¬/25¬05:03
tuxbrain_awaymaybe I will add some leds and resistors to the package to complete without extra price05:04
Ornotermestuxbrain_away: try to find suppliers in china if you don't have, shuld be a lot cheaper to buy breadbords from there05:04
tuxbrain_awayyou arrive late at the conversation, I also made a 1K order to china manufacturer05:05
Ornotermesoh, ok05:06
tuxbrain_awayI have a lot of stock :) , luckily they are selling quite good05:06
Ornotermesi noticed i was highlighted when i got out of the shower :P05:06
wpwrakthis took a moment to parse :)05:07
tuxbrain_awaybut as I said, was not ease to find a balance on price/quality, I see really cheap crap with closed holes and bad contacts...05:07
wpwraktuxbrain_away: http://www.theonion.com/articles/chinese-factory-worker-cant-believe-the-shit-he-ma,1343/05:09
tuxbrain_awayand expensive ones than for it's price you should tell the board what to do with natural language and work with wireless energy taked from air , but not it was just a piece of plastic waiting for components to be plugged05:09
Ornotermeshttp://dilbert.com/strips/comic/2010-09-21/ wow, i didn't knew dilbert worked at apple ;)05:15
qi-bot[commit] Werner Almesberger: ben-lcdframe-back-1mm is done. http://qi-hw.com/p/ben-scans/6186a7205:25
kristianpaultuxbrain_away: if you can do a simple stecht of your idea about the case will be asier for me to understand where it goes in the design07:41
qi-bot[commit] kyak: Support two-byte characters in console http://qi-hw.com/p/openwrt-xburst/343b7ed07:49
qi-bot[commit] kyak: Added input (typing) and output (viewing) support for cyrillic in http://qi-hw.com/p/openwrt-packages/a8aff4108:16
-:#qi-hardware- [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp11:35
qi-bot[commit] Werner Almesberger: ben-lcdframe-back-500um is done. http://qi-hw.com/p/ben-scans/d5001c612:26
wpwrakkristianpaul: he's looking for an extension of the case at the antenna end, there it goes lower and then has a "nose" that goes into the little hole the ben has for a lanyard12:27
wpwrakkristianpaul: so if you push from above or below, the force would go to that nose and not act like a lever on the micro-SD slot12:27
wpwrakkristianpaul: not sure if it would really make sense, though. most likely, you'd just shear off the small plastic nose while the rest of the board wouldn't even notice what's happening12:28
wpwrakkristianpaul: that is, unless you have a really strong plastic that resists to shear like steel12:32
wpwrakkristianpaul: btw, what sort of plastic are you using for extrusion ? may have properties that also make it interesting for DIY injection molding.12:33
kyakbartbes: are you there?12:43
kyaki've noticed that though i have libncursesw selected, the CONFIGURE_ARGS (specifically, --enable-widec) is not passed to configure script12:44
kyakdoes this "VARIANT" thing work at all?12:44
bartbesyou should ask larsc12:44
kyaklarsc: do you know perhaps what could be the problem?12:45
kyakat the end i end up with libncursesw libraries as well as libncurses, but there is obviously no wide char support12:45
kyakconfigure says "checking if you want wide-character code... no"12:46
kyak(now that i have cyrillic utf-8 in console, i want to build lynx with multibyte support)12:47
bartbeswell, it worked when I delivered it to larsc, he said he needed to make some more changes12:47
bartbesand now apparently it fails12:47
Action: bartbes doesn't want to be blamed ;)12:48
kyakhm, can you show your original version?12:50
bartbesI don't think I still have it12:51
larscin theory it should work ;)12:51
bartbeswell, it might be the reason why I never got wordgrinder to build either12:53
kyaki think it should build two times:12:54
kyakfirst times it should build libncurses, second time libncursesw12:54
larscthats what it should do12:56
larsc(in theory)12:56
larscand at least did at some point in practice12:56
kyaklarsc: while you are here, can i ask for another favor: feeds/packages/utils/vim/Makefile should have "--enable-multibyte" instead of "--disable-multibyte"12:58
kyakonly this way i can make utf-8 work in vim12:58
kyakfor vi (built-in busybox) it is enough to have CONFIG_BUSYBOX_CONFIG_FEATURE_VI_8BIT=y12:59
kyakalso i think you should disregard my previous concerns about libncursesw... the second time (when it build libncursesw configure says "checking if you want wide-character code... yes"13:01
kyakso i'll continue digging on lynx's side13:01
larsckyak: configure:6814: checking if you want wide-character code13:02
larscconfigure:6824: result: yes13:02
kyakyes. soory for confusion13:02
kyaki only noticed the first time build before13:02
kyakwhere it said "no" for libncurses13:02
Action: bartbes looks at larsc 13:03
bartbestime for payback?13:03
larscno time ;)13:04
kyaklarsc: what about vim? i think i have access to upstream OpenWrt13:05
kyak*you have13:05
larscyes. but i'm not sure whether that change is wanted13:08
kyakwhy not? it won't affect existing users experience13:09
kyakalso, i assume that those who want minimal vi, are satisfied with the one from busybox.. when they install vim-full they should expect multibye support (my opinion)13:10
larsci'll think about it.13:11
larscbetter ask mirko. i think he is more likely to merge such a patch13:11
kyakok, thank you13:11
kristianpaulwpwrak: BAS13:54
kristianpaulwpwrak: ABS is the plastic13:54
kristianpaulwpwrak: i also have some nice balck PLA i think ill try that at last when got better design13:55
kristianpaulwpwrak: are you lanning do you own plastic molding machine? cool :D13:58
wpwraki;m trying to find out what it would take. industrial injection molding has some scary properties, but some of them may exist simply because of specific requirements.14:03
wpwrakabs doesnt't look bad. in many ways similar to polyethylene. now if i could just find the mfr/mfi ...14:08
kristianpaulwpwrak: http://store.makerbot.com/plastic/red-abs-plastic-1.html the RED abs i used14:12
kristianpauli dont see datasheet but i t should be14:12
kristianpauli'll ask makerbot if you consider i should?14:12
wpwrakkristianpaul: i think i found a place with answers already: http://www.matweb.com/search/PropertySearch.aspx14:20
wpwrakLDPE should still flow better than ABS. about twice as well.14:22
wpwrakabs can handle more mechanical abuse, that's for sure.14:26
rafata taaaan!14:33
wpwrakrafa: good morning, too :-)14:34
wpwrakkristianpaul: what's interesting is that these plastics melt flow at pretty low temperatures. only ~105-120 C. i wonder if one could just heat a block of wood and use that as a mold, given enough time. (if the mold is hot, then there's no rush in injecting the plastic)14:36
wpwrakkristianpaul: ABS also uses much higher injection pressure than PE. ABS ~80 MPa, PE ~10 MPa. a DIY injector should be able to produce something around 1-10 MPa.14:47
wpwrakkristianpaul: actually .. i wonder if you really need more than a threaded tube you can heat, plus a large screw. add a motor and you have a servo-controlled injector, the pinnacle of injection techniques :)14:49
wpwrak(injection pressure) oops, that was another plastic. nope, only about 50 MPa for ABS and 80 MPa for LDPE. LLDPE would go as low as 10 MPa. grmbl.14:56
urandom__bah fuck off intel!16:27
wpwraka marketing strategy designed to be controversial ;-)16:35
qi-bot[commit] Carlos Camargo: Adding file-system http://qi-hw.com/p/nn-usb-fpga/911637518:08

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