#qi-hardware IRC log for Friday, 2011-09-16

whitequarkok, for the first round I'll tell the truth.00:02
wpwrakalways a good first step towards determining which set of lies they want to hear :)00:03
whitequarkafter all, there is a very nice service here which allows to create a virtual credit card attached to a e-money account, and it even works with paypal.00:04
wpwrakoh, nice :-)00:04
whitequarkthe e-money account is itself attached to mobile phone -- but well, you can go and buy a number for $3, nothing more required.00:05
whitequarkthinking deeply about it, I doubt that such system may have appeared and continues to exist just on its own.00:05
wpwrakas a customer there, you'll probably be in illustrious company ;-)00:06
wpwrakto follow proper protocol, you should use tor to access it00:07
whitequarkI doubt someone really uses illegal accounts to buy stuff on DigiKey (why would you ever want it?..)00:08
wpwrakwhy not ? use somebody else's money. ship to a place where your identity isn't revealed.00:09
whitequarkthis service comes with mandatory SMS confirmation (through digikey obviously cannot know about that)00:11
whitequarkand also, does re-selling electronic parts bought from a retail vendor generate a lot of profit? there are easier and more paying ways to fiddle with criminally-obtained money, I'd say00:13
wpwrakwhy not ? buy with someone else's money. sell to local electronics shops.00:16
whitequarkhm, looks like my bank has declined their transaction, which is strange00:17
whitequarkwpwrak, have you bought something from digikey?00:18
wpwrakyeah, tons of stuff00:18
wpwraki found them quite easy to work with. less paranoid than the rest.00:18
whitequarkthey've first sent me a bill for $1, which confirms that card no., exp. date and cvv is correct00:19
wpwrakmaybe you have to advise your CC company so that they pre-approve the transaction00:19
wpwrake.g., in argentina, they do such things. very annoying. luckily, i kept my swiss credit cards :)00:20
whitequarkthat's not a real bank (well, they are de jure, but not de facto). they're more like a common e-money service provider00:21
whitequarkahh. the taxes + usd-to-rur convert rate.00:22
wpwrakmaybe digi-key had some bad experience with that bank and therefore are more careful00:23
wpwrakof they generate fake credit cards, that would explain a lot of trouble :)00:23
wpwrakmuch of the customer authentication in web-based shops relies on matching the shipping address with the billing address of the credit card. mess with this, and all their alarm bells go off.00:24
whitequarkah yes, that may be the case. in all previous cases, shipping and billing addresses were the same.00:25
whitequarkdarn. they want a _fax_.00:26
qi-botThe build has FAILED, see log here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-09162011-0128/00:26
DocScrutinizerwpwrak: (hatched) 5ct cheaper per board00:26
DocScrutinizerdue to less copper needed00:26
whitequarkDocScrutinizer: isn't it etched away anyway?00:27
DocScrutinizerand even when it is, do you think they dump the copper saturated acid in the sea?00:28
wpwraki thought is was the secret ingredient of coca cola :)00:28
DocScrutinizerI don't know of any other reason to use hatched planes00:29
whitequarkDocScrutinizer: I've heard environmentalists saying that, yep.00:29
whitequarkalmost literally and dead seriously, that is.00:29
DocScrutinizeron >1000EUR per metric ton of copper? you're dreaming00:30
wpwrakin russia ? from what one gets to read, i wouldn't be surprised :)00:30
whitequarkwpwrak: yeah, and that's not the craziest thing they say00:31
wpwrakin china ? i'd be surprised if they didn't dump whatever, but i'd be surprised if there's much complaining about it :)00:31
whitequarkone of the reasons I treat them as sick idiots00:31
whitequarkby the way and on the topic of recycling, green & stuff00:31
wpwrakDocScrutinizer: i guess it depends on how easy/hard it is to extract the copper 00:31
DocScrutinizerdead simple, on 1000°C there's nothing but plain elementary copper left over00:32
DocScrutinizerat very least it's as good as any other ore to melt out copper from it00:33
whitequarkisn't Pb-free solder much more brittle (which leads to more device failures), more expensive to produce and buy (which leads to more profit for manufacturers), makes devices quite hard to repair (from experience and by words of a friend who does that), and quite un-environmentally-friendly to produce (not really sure, but heard something about that)?00:33
whitequarkand it's promoted by government, of course00:34
wpwrakDocScrutinizer: hmm, what you have is a salt. are you sure that comes apart so easily ?00:34
DocScrutinizerwpwrak: I'd say yes. If not, add electrolytic desintegration00:35
wpwrakwhitequark: not sure if it leads to all that many more failures. and a lot of consumer electronics are designed to die young. so the soldering may not be the weakest spot.00:35
whitequarkwpwrak: (salt) it melts at 801 deg. C. and evaporates at ~150000:36
whitequarkwpwrak: (devices) for the portable ones, I'd say yes. the point about the design is valid, through00:36
wpwrakDocScrutinizer: depending on the acid, even that may not be enough. so you may need several processing steps before you've gotten rid of all the extra stuff. would be interesting to see the cost structure of that.00:37
wpwrak(salt) i see that you found wikipedia, too ;-) but the salt in question won't be NaCl ...00:37
DocScrutinizeryou bet they'll use an acid that makes recovering the copper simple00:38
DocScrutinizerI recall e.g. ammoniumperclorate which obviously is all gaseous elements00:38
whitequarkwpwrak: the numbers are pretty similar for others, too. and can you name a non-metal-based acid which does not turn to gasses at 1000 deg. C?00:38
wpwrakdunno. i've hard they use Fe(III)chloride in industry as well. that promises to be messy. all the "what can i do with my spent acid" processes i've read about sounded somehow nasty00:39
wpwrakHCl+H2O2 is much cleaner, though00:39
DocScrutinizeron a large scale copper is "etched" away in electrolytic bathes anyway, most cheap method AFAIK00:40
whitequarkah, wait a bit. does "acid" mean also "etching solution" in this context?00:40
DocScrutinizerof course if you use ferro-III-cloride you'll have a problem with mere thermal copper recovering00:41
DocScrutinizerHCl OTOH...00:41
wpwrakhere's the salt you get with HCl+H2O2: http://en.wikipedia.org/wiki/Copper(II)_chloride00:42
wpwrakboils at 993 C. 1000 C were a good guess !00:42
DocScrutinizeractually large scale PCB production isn't using etching at all afaik. They print the traces with graphite or similar, then copperize it in an electrolytic bath00:44
whitequarkDocScrutinizer: so they don't use photolitography then? I thought it is the fastest method00:45
DocScrutinizerfor small scale it probably is. For 100s of k up to mio of boards you start to think all new about processes00:45
DocScrutinizerI don't think this article is comprehensive00:49
whitequarkwhat's Pertinax?00:50
whitequarksome kind of textolite?00:51
DocScrutinizerpaper with a poliphenol-resin iirc00:52
wpwrakah no, sorry. it should be Copper(I) chloride. boils at 1490 C. odd, though. the crystals i get all look like Copper(II) chloride00:53
DocScrutinizermilling probably is also a commonly used method in industry00:54
wpwrakslooooow ?00:54
DocScrutinizerwho cares, cheap anyway00:54
DocScrutinizerand you recycle almost clean plain copper00:55
DocScrutinizerand the milling machines are not that expensive either00:55
whitequarkwpwrak: why Copper(I)? all etchers are quite potent oxidizers, and you only get Cu(I) with a really weak ones00:55
wpwrak(recycle) if you mill away all the extra. that makes it even slower.00:55
DocScrutinizeryou can have one XYZ actuator with dozens of milling heads00:55
wpwraki see milling more for chemicals-free prototyping00:55
wpwrakwhitequark: dunno. the pages i found that claim to explain the chemistry of the process say it's just CuCl, not CuCl200:57
DocScrutinizerbut honestly I never cared about how industry produces PCB. I'm just using it, and I guess they offer large qty by the raw area/weight of copper used per unit00:57
DocScrutinizerthat's where hatch comes in00:59
wpwrak(hatch) heh, you learn a new bit every day :)01:00
DocScrutinizeractually it's been loooong since I last have seen a PCB with hatched areas01:02
DocScrutinizernowadays PCB are usually too small so it doesn't matter01:02
wpwrakand people may want the shielding anyway01:03
whitequarkwpwrak: (digikey) you were right about the addresses--now they want an ID card. I just hope email is fine for them01:03
wpwrakwhitequark: i'm surprised they even go through with this. the normal procedure would be to bounce you back to other forms of payment (wire transfer or such)01:04
wpwraknice text here, alas missing the formulas: http://rdchem.com/chemistry-of-the-cupric-chloride-etch.html01:04
wpwrak"However, the practical difficulties of using Chlorine for replenishment, not just the actual safety issues, but more importantly, the perceived problems, in the heavily legislated, risk-averse, environment of today, [...]"01:05
wpwrak"Chlorine gas can, of course, be very dangerous to any form of life, even to chemical process engineers."01:05
qi-botThe build has FAILED, see log here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-09162011-0226/01:24
xiangfuwe only have one color in RC3 right? 01:25
xiangfuI mean the milkymist one case01:25
wolfspra1lxiangfu: yes light-blue01:26
xiangfusomeone ask me some question in Youtube channel. 01:30
xiangfuthen I ask here for make sure :)01:30
qi-botThe build has FAILED, see log here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-09162011-0324/02:22
xiangfuyes. seems we needs upload the all modules to linux 3.002:23
xiangfuI temporary disable the kmod-xfs02:47
wolfspraulxiangfu: who would need xfs anyway?02:53
wolfspraulis that the xfs filesystem?02:53
wolfspraulI doubt any current nano user uses it, neither can I imagine where/how it could be used02:54
wolfspraulxfs on a memory card?02:54
wolfspraulI'd say until we hear from someone who needs it in some case, we can leave it out02:54
DocScrutinizerhow do you manage optional kernel modules?02:54
xiangfuDocScrutinizer, just build all of them as packages.02:57
DocScrutinizerI haven't seen or heard a real good concept how user could decide to install optional modules or alternatively kick out unused modules based on usecase02:57
DocScrutinizeryeah, you can probably do this if you do a build-all02:58
xiangfuDocScrutinizer, some packages maybe depends on kmod. that end user may install some kmod. 02:59
xiangfuDocScrutinizer, ender user will never touch /etc/modules :D03:00
xiangfuDocScrutinizer, just grep currect release. there are kmod-fuse, kmod-ks7010, kmod-tun installed.03:01
DocScrutinizeron maemo we got a special kernel that supports (among other things) USB hostmode, and it's a real pain to get modules built and deploy them for e.g iso9660, as that kernel is also just a package03:02
xiangfuwolfspraul, yes. I only know there is 'xfs' never look into it. even when we meet this build error. I just disable it :)03:02
DocScrutinizerso when you upload a new kernel to outobuilder, how shall the autobuilder know which (kernel module) packages need a rebuild03:04
wolfspraulkristoffer: he :-)03:08
wolfspraulyou beat me03:08
wolfsprauljust delete the whole page, not each thread...03:08
xiangfuDocScrutinizer, build them all? :)03:08
DocScrutinizerGRRR [2011-09-16 05:07:10] [Fehler] Verbindung zu Server chat.eu.freenode.net (Port 6667) abgebrochen: The remote host closed the connection.03:09
DocScrutinizerxiangfu: I guess there are too many packages on http://maemo.org/downloads/Maemo5/ so building them all is not an option, even if you'd know what exactly shall trigger this build-all03:11
Action: kristianpaul feels that need a nap03:21
freemorCool, watching CC conference @ on my Nanonote while working on my server (NN conneted to servers USB)14:00
wolfspraulall thanks to David14:01
freemorYes, Mplayer is a big +14:04
qi-bot[commit] kyak: libassuan: update to 2.0.2, add libgpg-error as a dependency (master) http://qi-hw.com/p/openwrt-packages/b964a4018:01
whitequarkwpwrak: hm, I've scanned a copy of requested passport pages for them (they're in russian, and the address is, yes, _handwritten_.) After that, I've noticed that the billing address does not match the one in passport, as I've relocated.19:57
whitequarkmaybe it's one of the reasons they (digikey, btw) charge $120 for shipping to russia.19:59
wpwrakthe address in the passport may not matter. most passports don't have an address anyway.20:10
viricwhitequark: are you in moscow?20:15
ArtyomHi kristian21:23
kristianpaulhello !21:28
ArtyomHow is your project? Do you have any news? ;)21:30
kristianpaulbad news.. :/21:30
kristianpauli'm having not able to initialize namuru from rtems21:31
kristianpaulwich i switched to, so i'll be able to do more advanced stuff like writing to a file ;)21:31
kristianpauli just ported the C code from the hacjed milkymist bios, but accumlators are reading zero :(21:31
kristianpauli dont know what is hapening, it should work. initialization is pretty simple21:32
kristianpaul so i decied to add to some namuru registers the posibility to be read as well, so i can confirm the write was sucefull, but this was not needed when working from the bios..21:34
kristianpaulwhat about you Artyom?21:34
Artyomwell, I've managed to make tracking work. It is unstable, but this is the first step which was very important for me. Now I slowly clean-up the code (remove unnecessary comments, magic numbers and so on). I hope I will prepare new post for my site within couple of weeks.21:37
kristianpaullarsc: actyally there is something else, if i intialize namuru from bios and then try to read accumualtors from rtems system hang :/21:38
kristianpaulArtyom: great !21:38
kristianpaulI'm late now, !!21:38
kristianpaulArtyom: had to modify something in namuru to make it work, or just fix tracking loop algorythm in software?21:39
ArtyomOne stupid mistake. Problem was in vhdl - my mistake, not namuru21:40
kristianpaulmistakes are mistakes no matters the size of it ;)21:40
kristianpaulha look it, now is not hanging21:41
kristianpaulArtyom: when you said unstable your atena is a fixed position, or you decided to take a walk around the square with it? :)21:42
ArtyomWhat is milkymist bios? And why did you decide to use RTEMS?21:42
ArtyomNo, I use simple one-channel simulator now.21:43
ArtyomIt's easier when you have stable signal21:43
kristianpaulmilkymist bios is the code that allow to milkymist soc sdram to be initialized amonng other things like ethernet i guess21:43
kristianpauli decided use rtems cause is the only "OS" the support ethernet, so i will be able to plot accumulator and check later on the desktop computer21:44
Action: kristianpaul miss linux all time 21:45
kristianpaulArtyom: (simulator) wow, you have espensive toys 21:46
Artyomoh, you want to send data from acquisistion/tracking to PC. I use rs-232 now and it's not very good. The speed is too slowwww. I have to record values from correlator in array and then send data to PC. But array size is veryyyyy limited21:47
kristianpaulat 115200 bps still aceptable? well is a single channel you are right21:48
kristianpauli could use tftp but the libry just support reading not psuhing data.. :/21:49
Artyomin my case - no matter what is the speed as I record values in a special array at first. (I use 57600 bps because I couldn't start with 115200 and didn't want to spent time on this question)21:51
kristianpaulwell i have 128Mb of ram i think i could allocate data there and transfer later by serial, not bad idea21:52
kristianpaulyeah, i havent spent time with rtems just because i wanted to use ethernet..21:53
ArtyomI have much less. Only 64kb ;)21:53
Action: kristianpaul hides21:53
kristianpaulArtyom: why you think you tracking is unstable? bad doppler compesation?21:55
ArtyomGood question :)21:55
ArtyomI can see two potential problems:21:55
Artyom1) I'm not experienced C-programmer for ARM7 and I could make some bugs;21:56
Artyom2) Preveously I played only with post processing. And there is one significant difference/21:57
ArtyomWhen I work with real-time receiver I always have some delays: for example delay between the moment when correlators output data and the moment when tracking algorithm completed and new control words for NCOs are ready 21:59
ArtyomMay be it's important. I don't know it now. I write a pc-program that will emulate these delays. I'm going to play with it and find out whether these delay are important or not.22:01
ArtyomBTW did you receive a letter fro me? ;)22:01
kristianpaulno :(22:01
kristianpaulor no yet22:01
kristianpaul(delay) may be you can do some buffering22:02
kristianpaulbut i asumed if you to all processing in less than 1 ms things should be in place, correct?22:02
ArtyomI don't know it...22:04
kristianpaulor do use threads ? (wich i dont know thing ethier)22:05
Artyomno, I don't22:06
Artyomtime to sleep ;)22:08
kristianpaulI'm waiting to read your blog post :)22:08
kristianpaulsure, time here to find what's wrong with initilization from rtems :)22:08
whitequarkviric: yeah22:18
kristianpaulbtw Artyom is also from russia, moscu i think too22:19
whitequarkwpwrak: that's what he asked for, anyway. we'll see if that helps22:19
whitequarkkristianpaul: does he know a secret power which makes digikey's requirements sane? apart from that, we're in pretty much the same position, aren't we?22:21
kristianpauldont you could ask him22:21
kristianpaulwhitequark: his blog http://gnss-sdr.ru/22:22
whitequarkah yes, I've seen that some time ago22:22
whitequarkthanks, anyway22:22
kristianpaulas you can see he made a board, but i havent asked about sourcing history22:22
kristianpaulsame position, well i havent orderen from digikey as the rule seems to be i must make those 50Usd shiping (plus aroudn 20% in taxes) to worth :/22:24
whitequarkkristianpaul, it's $120 to russia, but there is a way to avoid it22:24
whitequarkyou can use one of the USA mail forwarding services, they are pretty cheap22:24
whitequarkthe one I use is $10/month and has no per-package fee22:25
kristianpaulwhat about taxes?22:26
whitequarkfor a $50 order, they've charged me something like $4.33 in taxes22:26
kristianpaulnot bad22:26
whitequarkdigikey isn't any cheaper compared to some good russian retailers (they're on par overall), but they have everything, unlike the latter22:28
whitequarkand, more importantly, they're selling that 'everything' from one piece22:28
kristianpaulvery important indeed22:29
kristianpaulwe're not all sharism ;-)22:29
whitequarkI can buy SMD caps (and similar things) like 10 times cheaper here, but when I have a need for an LVDS serializer, I'm out of luck22:29
kristianpaulthis ail forwarding service, uses USPS?22:30
kristianpaulcause if i found a shop similar to diigikey that ship by USPS, i'll be happy 22:31
kristianpaulzero taxes :)22:31
kristianpaulno matter if i must wait even a month, but if arrives no problem22:32
whitequarkkristianpaul: http://shipito.com22:32
kristianpaulexcelent !22:38
kristianpaulUSPS Priority Mail, right way i can afford :)22:38
kristianpauloh, they even have warehouse in china22:44
kristianpaul"For shipments valued over $20, the sender must have a copy of his or her commercial invoice legalized by a Colombian Consulate, and must prepare 4 copies of a consular invoice, one of which must be sent direct to the addressee with the legalized commercial invoice. "22:48
kristianpaulthast not true, i never had to go consulate and i already received intems declared about 800usd by USPS and UPS22:50
--- Sat Sep 17 201100:00

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