#qi-hardware IRC log for Sunday, 2014-04-20

DocScrutinizer05LOL! http://xkcd.com/1357/   00:32
DocScrutinizer05[2014-04-16 Wed 03:32:56] <DocScrutinizer05> natsukao: you've been warned twice to stop00:32
DocScrutinizer05[2014-04-16 Wed 03:33:20] <natsukao> I° Amendament of American Constitution defends Freedom Of Speech00:32
DocScrutinizer05[2014-04-16 Wed 03:33:31] *** You set a ban on natsukao!*@*.00:32
wpwrakso he'll never vote for you to become his governor :)00:53
DocScrutinizer05it is actually strange how xkcd fits in a 3 days later01:10
DocScrutinizer05you might think he#s on #maemo, lurking01:10
DocScrutinizer05if he's not even a she, who knows01:11
DocScrutinizer05ooh, >>Randall Munroe<<01:15
DocScrutinizer05awesome: http://xkcd.com/249/01:19
whitequarkDocScrutinizer05: btw, on VFAT +x is the default06:26
whitequark(unless you set a mount option)06:27
DocScrutinizer05though it shouldn't, and users/owners are hard to specify08:48
wpwraknice: the cost of heartbleed http://www.wired.com/2014/04/cost-of-heartbleed/11:20
wpwraki guess it would almost make sense to add "tags" to certificates, so that one could revoke a tag and with it all the certs that have it11:22
wpwrak(of course, the few sites who weren't affected by heartbleed would complain bitterly if they had to renew their certs due to revocation-by-tag ...)11:23
wpwrakand fireworks at samsung cause the worlds samsung devices to fail: http://www.engadget.com/2014/04/20/samsung-com-outage-sds-fire/11:30
wpwrakthose ads for moving stuff into the cloud are almost writing themselves ;-)11:31
DocScrutinizer05I just came up with a (life) risk assessment unit for missions (like to compare mars mission to jogging): LD50/99, LD for Lethal Duration. A jogger has probably 70(-age)/100(-age), means of a test group of joggers 50% are dead with 70 years old, 99% are dead with 100y12:28
wpwraka common metric is LD<percentage>/<duration>, meaning that, after exposure to ingredient X, <percentage>% of the test subject population have perished within <duration>12:32
wpwraka bit more meaningful than the ordinary LD50 that doesn't take into account time12:33
whitequarkwpwrak: you'd make a good GLaDOW12:33
whitequark"test subject population", hehe12:33
DocScrutinizer05>>The big question left? Why a fire at one location seemed to have such a large affect on the company's devices and services<<12:34
DocScrutinizer05those guys are so damn pretty clueless12:34
DocScrutinizer05about how android works12:34
wpwrak"because we don't know how to design such a thing"12:34
DocScrutinizer05or rather, how smartphone shit in general works12:34
wpwrakand it gets better: the whole thing is completely useless: http://www.sodnpoo.com/posts.xml/spoofing_the_samsung_smart_tv_internet_check.xml12:35
DocScrutinizer05or rahter, how sw licensing of non-FOSS stuff works today12:35
wpwrakit's basically a "ping - can anyone hear me" ? doesn't carry any real information12:35
wpwraki think it nicely demonstrates how product obsolescence will be implemented in the not too distant future12:36
DocScrutinizer05bzzz false: *got* implemented during last few years already12:36
DocScrutinizer05Nokia HARM did it12:36
wpwraki mean we're already getting stories of this and that cloud-based or internet-centric company folding and customers left with mere days to save whatever they can12:37
DocScrutinizer05even Nokia fremantle maps already did it12:37
DocScrutinizer05and seems even jolla does it12:37
wpwraksounds like a must-have feature for neo900 :)12:38
DocScrutinizer05"no registration to our server, no authentication on app startup, ->no app starting up"12:38
DocScrutinizer05"no software update of course"12:39
wpwrakeach time you unlock it, it pings joerg personally. if he's asleep (hah, fat chance for that) or otherwise unreachable, you're denied service :)12:39
DocScrutinizer05hehe sounds like a dream. A nightmare12:39
DocScrutinizer05for +me*12:39
wpwraki kinda wonder why the iGadgets don't do that. perhaps Jobs wasn't that much of a control freak after all ;-)12:39
DocScrutinizer05smarter such schemes include a timer that allows using the app for a certain timespan without connectivity, after a successful authentication and "lease" from the company central scrutinizer server12:41
wpwrakheaven no. they use that in DNS ! no no, we have to find some proprietary concept12:43
DocScrutinizer05basically jobs been too early for that scheme to fly.that's why iOS changed from initial concept which been 100% web apps towards a 99% autonomous architecture that actually stores stuff locally and doesn't need 24/7 connectivity to work12:43
DocScrutinizer05nowadays nobody is concerned about 24/7 connectivity (or the lack thereof) anymore. after all when you got no carrier signal for your phone, you also don't need your contacts and prolly neither your calendar or whatever12:44
DocScrutinizer05so more and more system architectures depend on 24/7 connectivity. For sure stuff like maps apps do, to authenticate to the map server every time, even while the tiles are maybe already in local cache12:46
DocScrutinizer05I bet siri and the similar crap on winphone is also an off-board aka web service12:47
DocScrutinizer05aka client server architecture12:47
DocScrutinizer05we're going back to the times of mainframes12:48
DocScrutinizer05connectivity becoming a single point of failure12:49
DocScrutinizer05when IMAP got invented?12:49
wpwrakNSA and samsung are doing a good job at making people aware that this may not be such a great idea. of course, it'll take a while and more incidents to sink in12:50
DocScrutinizer05whitequark: see, the idea lingers around forever12:51
DocScrutinizer05or rather, for 25 years12:51
whitequarkIMAP is not bad. it surely beats POP3 in a world where I may want to access my mail from more than a single device12:54
wpwrakand you can always use it as if it was pop12:54
whitequarka good IMAP client would cache the data locally, like the android one does12:55
DocScrutinizer05in times where literally nobody even in developer community had the "joy" to work on a terminal connected to a mainframe infra via a pay-per-minute dialup connection, and everybody enjoys 24/7-always-on-connectivity at flat rate, it will take a bit longer until things change for the better again. right now they are changing for the worse, at ever increasing speed12:56
DocScrutinizer05wpwrak: (us like POP) that's astonishingly difficult and cumbersome12:56
whitequarkpointless in the first place12:57
whitequarkI use IMAP because I *want* my mail to be stored remotely12:57
DocScrutinizer05yeah, pointless as long as your mail server offer proper POP312:57
whitequarkI don't have a desktop, and every device I have on myself, i.e. every device I have, is expendable12:57
DocScrutinizer05alas some server admins decide to deliberately not offer POP12:58
DocScrutinizer05in POP I can opt for "leave mail on server" and "delete mail on server after expiry of guard period"12:59
DocScrutinizer05in IMAP moving a mail from IMAP dir to a local dir on one server really creates a cascade of funny events13:01
wpwrakDocScrutinizer05: i remember the modem days very well :) even built my own 300 bps acoustic coupler. used first with the VC-20 (VIC-20), then used the VC-20 as "gateway" to my CPC-464. the CPC had lots of memory and a good resolution, while the VC-20 had a UART ...13:01
wpwrakfetchmail handles imap just fine. that's how i get all my mails. (and i keep a read-only copy of everything on the server)13:04
DocScrutinizer05err wut?13:07
DocScrutinizer05how would you do that for the average web.de or gmx.net server ?13:07
DocScrutinizer05the latest when I have to keep my /trash/SAPM/ dir on the IMAP server and local spamassassin in cooperation with the mail client has to upload mails to that SPAM dir, I lose my humor with IMAP13:08
DocScrutinizer05then there are MLs. I got ML: business, f, g, h, private. I got three mail clients on 3 device. client 1 has business, f,g, h in 4 dirs, client 2 has all 5 in separate dirs, client 3 has f,g,h,private all in one common ML dir13:13
wpwrakwell, i have my own server. my server, my rules :)13:13
DocScrutinizer05real fun with IMAP13:14
wpwrakand yes, if you need to share mails between devices, imap (and leaving them on the server) is one of the few choices you have13:15
DocScrutinizer05nah, POP3 and leaving them on the server is the better choice13:15
whitequarkdoes POP3 even have folders?13:16
DocScrutinizer05leaving them on the server untl either the "slowest" of your clinets downloaded them, or alternatively exoire them after maybe 2 weeks or somesuch13:17
DocScrutinizer05whitequark: huh? 13:17
whitequarkDocScrutinizer05: I get a new client, do I also need to manually organize all my mail on it?13:17
DocScrutinizer05whitequark: luckily it doesn't, it has no such idiotic concept of it's-like-webdav-but-not-really13:18
whitequarkso yes13:18
whitequarkI can't classify your system as anything but "incredibly moronic"13:18
whitequarkso no*. anyway13:18
whitequarkit's unusable13:18
DocScrutinizer05whatever it is for you, for me it works way better than any fucking imap forcing me to have identical dir structures on all my clients13:19
DocScrutinizer05"OMG that's unbearable, I have to *choose* what to do"13:19
wpwrakslavery is freedom :)13:20
whitequarktoo much pointless work. sure, if you like bureaucracy and useless shuffling of documents, it will probably fit well13:20
DocScrutinizer05how about you two guys found a club "slaves of IMAP"?13:20
DocScrutinizer05IMAP for sure is great when you wanna waer another "smart"phone of the day with the right color, every day of the week, and still all shall look identical by their UX13:22
DocScrutinizer05I defitely need a different mail config on my smartphone than on my laptop than on my workstation13:23
wpwrakmy use of imap comes from openmoko. anyway, it doesn't matter to fetchmail. all i want are my mails :)13:24
DocScrutinizer05I know, that's incredibly moronic, since ... who's going to distinguish between work and private nowadays. Particularly on their IT13:25
DocScrutinizer05</sarcasm> (another one, so i really get rid of this huge amount)13:26
DocScrutinizer05wpwrak: if I had only one client (or MTA) then I wouln't even think about IMAP, it's just another protocol to transfer mail then. But then, why would I need IMAP then, at all?13:27
DocScrutinizer05wpwrak: and yes, it's been the damn OM server that never did (and doesn't) offer POP3, and it's the only account where I have to mess around with this friggin IMAP13:30
DocScrutinizer05and the only account that gave me like half a dozen mega disasters during the years I'm using it, from all mails getting deleted from server thanks to an oppsie on one braindamaged client, to all mails which stay on server probably getting leaked to the hacker who rooted the box13:32
DocScrutinizer05no hacker ever rooted any of my local machines, neither any of my exposed boxes (though those don't hold any mails except maybe those of last 10 min)13:33
DocScrutinizer05and whn somebody now says "you should accept encrypted mail only, so you don't care when it gets stolen from the IMAP server which you can't control or administrate" then I gonna wake up my killer bat army and send them out to him as soon as sun sets13:35
whitequarknah, there's a simpler way13:36
DocScrutinizer05don't use mail, yeah13:36
whitequarkI never agree to *collect* my mail from anyone else's server; I just set up a redirect from there to a single point that I own13:36
DocScrutinizer05sure, create a single point of failure. Nicely defeats the purpose of having multiple accounts13:37
DocScrutinizer05and guess what, OM mailserver didn't even offer mail forwarding13:38
whitequarkit's a single point where I set up filters *once*, set up backups *once*, and point all my clients to it *once*13:38
whitequarkand the downtime though the last three years was prolly lower than that of GMail13:38
DocScrutinizer05sorry, can't help but notice you don't get it13:39
DocScrutinizer05I also have a single point of setting up my filters: the clinet to which the filters apply13:40
wpwrakwhitequark: where's the constant struggle to fill your life with purpose ? and see, you can't even get a good rant out of it ! look at DocScrutinizer05, he's been going on about this for an hours or two. that's how the experts do it :)13:41
DocScrutinizer05I however can ask wpwrak to send the mail to my web.de account when OM once more cocks up or his account is messed up and doesn't send stuff to OM, or whatever the problem13:41
DocScrutinizer05well, an hour is a mild punishment for "moronic"13:42
whitequarkDocScrutinizer05: you also don't get it, my mailserver is infrastructure that *I* control13:42
whitequarkand in the unlikely case that Linode fails, it has daily backups to a different country, continent and company13:43
DocScrutinizer05usually I won't stop until the other part admits that their mental capacity is insufficient to judge which of my actions are moronic13:43
whitequarkwhich can be restored in under a hour13:43
DocScrutinizer05whitequark: I well got that, but that tells me what? to move my web.de account to your server. or what?13:44
whitequarkto your own, I guess13:44
DocScrutinizer05well, than thanks for this moronic suggestion, since I can't guarantee the uptime of any such server either13:45
DocScrutinizer05and I can't move my web.de account to somewhere else, since... well I bet you can imagine why13:45
DocScrutinizer05if not, do w a `whois web.de`13:46
whitequarkit reduces the possible range of problems to two: 1. server went down entirely, 2. you broke it. from... (amount of accounts × amount of idiotic decisions owners of servers behind those accounts make)13:46
whitequarkI don't think I touched my mail config in four years or so. that deals with point 2.13:47
DocScrutinizer05this finally gave away that you don't get it13:47
DocScrutinizer05of course a web.de -> my_client is much less stable and more prone to downtimes for whatever reason than a web.de->my_MX_server->my_client13:49
DocScrutinizer05since it seems that's needed to not let a chance somebody thinks this been serious13:50
DocScrutinizer05and you call my management "moronic"??13:51
whitequarkin my case, the server is anything but forwarder. it is an aggregator13:51
whitequarkonce I have mail there, I know that I will be able to access it no matter what web.de decides to do with their boxes tomorrow13:51
whitequarkand I will be able to access it even if I get robbed of every single device I have13:51
DocScrutinizer05so what? what do you think I know about +my* mail, sitting comfortably local on my HDD I can see down there?13:52
whitequarkuntil said HDD chokes and dies13:52
DocScrutinizer05I don't get robbed13:52
whitequarklol, life advice by DocScrutinizer05. "-- how to get rich? -- start by not being poor"13:52
DocScrutinizer05thanks for the talk, had a lot of laughs! no for something more meaningful, like watching TV13:53
wpwrakhmm, ubuntu stopped launching ntpd. how quaint. so i've been living a few minutes in the past for the last weeks.13:53
whitequarkwpwrak: why do you even *use* that13:53
wpwrakbecause i like to have accurate time ?13:53
whitequarkdo you feel like not enough of your searches go to google and now voluntarily send even more to amazon? ;)13:53
whitequarkno, ubuntu13:53
wpwrakoh, it's reasonably painless to use. grew tires of days of compilation of gentoo, and eventually find that, if i didn't upgrade often enough, the chain would be broken and my old systems wuold become completely un-upgradeable13:55
whitequarkwhy not plain debian?13:55
wpwrakand what would go to amazon ?13:55
DocScrutinizer05(robbed) your problem when you rely on a remote box for your sole means of "data security". I use proper *local* and *remote* backup schemes for that, I only seen incidences so far with damn IMAP and damn OM server13:55
wpwrakat least back then debian didn't have so much stuff and what they had was often ancient13:56
whitequarkwpwrak: I find debian testing reasonably stable and up-to-date13:56
whitequarkcan't really recall the last time they broke something for me13:56
wpwrakwell, one of these days ubuntu will pass my pain threshold. but i'm not all that eager anymore to migrate. it does get a little boring after the first dozen times ...13:57
wpwrakwell, half a dozen. i think my sequence was: linus boot+root -> MCL -> SLS -> Slackware -> Red Hat (for a long time) -> (tried Debian on some systems) -> Gentoo (for a long time) -> Ubuntu (for a long time)13:59
DocScrutinizer05gentoo vs buntkuh, seems ther are at least a *few* quite reasonable alternatives in between13:59
wpwraki guess after the mail rant we'll now get the distro tantrum ...14:01
DocScrutinizer05btw to poll OM mailserver from my own box, i had to move account credentials for logging in to a *commercial* mission-critical infra to a box exposed to the net. I guess doing so would be enough of a reason to get fired instantly14:01
DocScrutinizer05while OM infra isn't mission critical anymore, due to lack of mission, it been when I got thta account14:03
Action: DocScrutinizer05 just found a good new way to test qualification of applicants for a sysop position14:04
qi-bot[commit] Paul Cercueil: Allow to map actions to keyboard keys and joystick keys at the same time (master) http://qi-hw.com/p/gmenu2x/5fa891914:09
DocScrutinizer05(gentoo vs buntkuh) I'm far from suggesting opensuse (though I think it's a great distro). There are others too, like fedora, centOS, a 1 billion users on debian...14:10
DocScrutinizer05lol, xkcd also states "fav astronomical object: plejades" - same here18:52
DocScrutinizer05N24 TV (or the station producing this "report") goes nuts now: they not only mute "fucking" with a beep, they actually, beliefe it or not, blurr the picture to hide the mouth of the one saying it19:04
DocScrutinizer05only the small section of picture to cover the complete mouth19:05
DocScrutinizer05that's *beeep* insane. Why the *beep* don't they simply cut out those 300ms completely?19:06
DocScrutinizer05sorry, only trailer http://www.n24.de/n24/Mediathek/Dokumentationen/d/2662996/rettung-in-letzter-sekunde.html19:10
DocScrutinizer05I think it's actually been that guy roundabout in middle of the trailer for 0.5s, who also does the bat suit thing and said sth like "it's fucking awesome" (not in the trailer)19:14
DocScrutinizer05the two vultures taking down the parachuter are incredible19:16
wpwrakso vultures do hunt :)19:22
DocScrutinizer05obviously, only this time it been a kamikaze hunt19:30
DocScrutinizer05the bird got bondaged in the strings of the parachute and went down togehter with the guy19:31
DocScrutinizer05dang, tied?19:31
wpwrakah, vulture failing to keep work (hunting) and pleasure (bondage) apart :)19:32
DocScrutinizer05vulture is a pretty strange word, to me at least. And not only becuase it looks like a typo of culture19:38
DocScrutinizer05it also sounds feminine case19:38
DocScrutinizer05and kinda like one of those "only plural exists" words like fish or sheep19:39
DocScrutinizer05but there's "vultures" which doesn't look that strange anymore, for a word19:44
wpwrakhmm, seems to come from latin, "vultur". basically unmodified.19:44
DocScrutinizer05does this have a "meanig"?19:45
DocScrutinizer05meaning even19:45
DocScrutinizer05most words don't, beyond their own meaning, Some names however actually mean something19:46
DocScrutinizer05miller for example19:46
DocScrutinizer05or kindergarden19:47
wpwrakthe origin seems to be unclear: http://www.etymonline.com/index.php?term=vulture19:55
DocScrutinizer05wow  >>A group of vultures is called a wake, committee, venue, kettle, or volt.<<19:57
whitequarkDocScrutinizer05: http://en.wikipedia.org/wiki/List_of_collective_nouns_in_English19:57
whitequarkthis is *seriously* crazy19:57
whitequark"a shrewdness of apes"19:57
whitequark"a draught of butlers"19:58
whitequark"a glaring of cats"19:58
whitequark"an army of frogs"19:58
whitequark"an array (sic!!) of hedgehogs"19:59
whitequark"a bloat of hippopotamuses"19:59
whitequark(incidentally, I now know how to call systemd)19:59
whitequark"a stud of mares" this doesn't even make sense20:00
whitequark"a cartload of monkeys"20:00
whitequarka fucking cartload, who invents those names?!20:00
whitequark"a parliament of owls"20:00
wpwrak"a congregation of alligators" sounds dignified :)20:02
wpwrak"a herd of asses" well, less so :)20:02
whitequarkthere should be "a list of butterflies"20:03
whitequarkor "a balanced binary tree of gryllotalpas"20:03
wpwraki think the general form is "<adjective-attributable-to-individual> <animal>s"20:04
wpwrakor where no adjective is available, some attribute20:05
whitequarkhow often do you see monkeys in cartloads?20:06
whitequarkor owls in parliament?20:06
wpwrak"a chatter of IRCers", "a rant of DocScrutinizers", "a drunk of russians", "a fraud of banksters", etc.20:09
wpwrakwell, some may have weirder origins :)20:09
sb0wtf http://www.digikey.com/product-search/en?pv316=24&FV=fff40019%2Cfff800ae&k=molex+8867&mnonly=0&newproducts=0&ColumnSort=0&page=1&quantity=0&ptm=0&fid=0&pageSize=2520:46
sb0same kind of shitty pricing at mouser20:46
sb0where can I find ribbon cables?!?20:46
wpwrakare you sure about the 100 mil pitch ?20:48
wpwrakthe normal ribbon cables are 50 mil20:48
wpwrak(normal) IDE and such20:49
sb0oh fuck,20:49
wpwrakhmm ?20:50
sb0well, mixxeo boards have 2.54mm headers ...20:51
wpwrak100 mil dual (2xN configuration) headers use 50 mil cable ...20:51
wpwrakif you have 1xN, you can still use a 2xN connector. it'll hang a bit in the air, though20:52
sb0it's 2x10 2.54mm20:52
sb0ah, so it's good20:52
wpwrakor you do it the traditional way and do the wiring manually (i.e., spread out the ribbon at the end)20:52
wpwrakso, no problem :)20:52
wpwrakgood that they didn't have the 100 mil cable at a less obscene price ;-)20:53
wpwrakthe whole rolls are still pricy at digi-key, but they also have them as "value added item", cut to more reasonable sizes20:54
sb0ah, yes, the *connector* pitch is 2.54mm on a more reasonably priced cable. it's kinda confusing they say "Pitch - Mating Interface 1.27mm" in the doc20:55
sb0I guess they mean the cable's mating interface20:55
wpwrak"connector pitch" ? that sounds like a weird parameter :)20:56
wpwrakespecially if it's different from the cable's pitch :)20:57
DocScrutinizer05problems with usual IDE connector type post arrays?20:59
wpwrakcable data sheet apparently specifying properties of the "far" side of the connector. not really the cable's business ...20:59
wpwraknaw, i think he was just confused about the parameters20:59
DocScrutinizer05btw you can use 50mil for the 100mil single row post connectors, just every second cable unused21:00
DocScrutinizer05wire, rather21:00
wpwrakyes, that's what i proposed in case he had 1xN header. but he has 2xN, so no problem21:01
sb0I'm buying complete cable assemblies - I don't have the tools for IDC, and they're pricy21:01
sb0and the docs of the assemblies mention the cable (wire) pitch, not the connector pitch, which is why I was confused21:02
wpwrakof course, the usual approach for 1xN is to use nible chinese fingers to make the cable ...21:02
wpwrak(or a clumsy western emulation thereof)21:02
wpwraksb0: for the ribbon cable connectors where you just squeeze the top bar onto the body, you don't really need special tools. just anything reasonably hard and flat will do21:04
sb0I've seen people doing that, and the resulting cables weren't reliable21:04
wpwrakadvantage over assemblies is that you can pick the cable size you want and don't need to stock a gazillion of lengths and widths21:05
sb0one of these people has even been a customer and I had to deal with serial ports that mysteriously stopped working etc. :)21:06
wpwrak(not reliable) hmm, never had a problem. well, not sure how often it use the proper tool and how often i'm too lazy to look for it21:06
wpwrakah, nice billable hours :)21:06
wpwrak#s/often it/often i/21:07
DocScrutinizer05nope, spreading the cable into a "delta" actually isn't the best idea. it causes asymmetries and thus signal propagation delay skew21:11
sb0DocScrutinizer05, do those cables work well for LVDS by the way?21:11
DocScrutinizer05which cables?21:11
sb0ribbon IDC 50mil cable pitch21:12
DocScrutinizer05why not21:12
DocScrutinizer05should work21:12
DocScrutinizer05unless you have tuned impedance, in which case the impedance of cable might not match21:12
sb0well, idk about the characteristic impedance of a pair of wires on this cable21:12
wpwrak(delta) usually not a concern at the speeds where you do that21:13
DocScrutinizer05but when the signal is on a post conector, you could bet on the 50mil cable being just fine for connecting it21:13
wpwrakDocScrutinizer05: caveat - the one asking now about the cable impedance is also the one who put the connector into the design :)21:14
wpwrakwikipedia is fun: "Knowledge of the impedance is one step toward understanding and control of interference that may be caused by ribbon cables."21:15
wpwraksuch subtle sarcasm :)21:15
DocScrutinizer05then you should have asked that question earlier in design process ;-)21:15
wpwrak(impedance) the cable data sheet generally has such characteristics, too21:16
DocScrutinizer05sure, should21:16
wpwrake.g., this one is pretty comprehensive: http://multimedia.3m.com/mws/mediawebserver?66666UuZjcFSLXTt4Xf6NXF6EVuQEcuZgVs6EVs6E666666--21:17
DocScrutinizer05generally an impedance mismatch isn't much of an issue when you take it into account21:17
sb0well, IDE seems to reach decent speeds with those connectors :)21:17
DocScrutinizer05I guess they are fine up to maybe 20 or 50 MHz21:18
DocScrutinizer05but don't quote me on that21:18
sb0I only have 27Mbps/pin (and I can halve that with more connectors) over 10cm21:19
DocScrutinizer05usually when you got a LVDS input somewhere, it already gives decent hints on caling you may use on it21:19
sb0I might not even need LVDS or other fancy stuff21:19
sb0one end is FPGA, which has LVDS configurable on all IOs21:20
sb0other end is a fancy OLED display module, which does not have LVDS inputs, but I have the option of adding a LVDS receiver on its board21:20
DocScrutinizer05the datasheet should give recommendations, under "electrical recommended operation conditions"21:21
wpwrakhmm, large OLED ? 'cause the smaller ones don't really need high data rates21:22
sb0320*480 24bpp 60fps21:23
sb0it's kinda fancy21:23
wpwrakah, okay. big one :)21:23
wpwrakindeed. almost the sort they use in cell phones21:24
wpwrakwill this one have live video monitor mode ?21:24
DocScrutinizer05awesome! >> Voltage Rating: USA: 300V   Canada: 300V EU: <50V<<21:24
sb0that's what I plan on using it for21:25
wpwrakkewl :)21:25
sb0after I'm out of mechanical design hell21:25
Action: sb0 hates mech design so much now21:25
wpwrakah yes, ME, always fun :)21:25
wpwrakit's art21:25
sb0it's a goddamn frustrating pain in the ass21:26
sb0standard parts 1) are ugly 2) never fit together21:26
sb0but they have the advantage of being just expensive21:26
sb0not horribly expensive21:26
sb0(and by "standard", I mean "stocked somewhere" - dual source is a luxury here)21:27
DocScrutinizer05print it! ;-)21:27
wpwrakmill it ! :)21:28
wpwrakor just dunk the whole thing in silicone ;-)21:28
wpwrakor encase it in a resin. that could actually work :)21:32
sb0I'm going to have to make 3 PCBs, one for the screen, plus one for each group of buttons, and connect the pots with flying wire21:33
sb0just because all those stupid parts can't have compatible heights21:33
wpwrakif you apply some anti-adhesive to the PCB, you may even be able to remove the resin by cutting it at the height of the PCB21:33
sb0and I fucking hate that21:33
wpwrakpick the same family of connectors for everything ?21:34
wpwrakah, you mean height if things above PCB21:34
sb0I need: buttons, rotary encoder with pushbutton, slide potentiometer, rotary potentiometers, screen21:34
sb0and NOTHING is compatible21:35
wpwrakyou could get parts designed for panel mounting21:35
wpwrakof course, this merely shifts the problem21:35
wpwrakbut you can at least mount them in a very solid way21:36
sb0e.g. after a lot of search, I could find a rotary encoder and a rotary pot with the same shaft and bushing configuration21:36
sb0but OF COURSE, they HAVE to use different heights for their bodies (i.e. distance between the shaft/bushing and the PCB)21:37
sb0and that's just for 2 parts, add buttons, screen, slide potentiometer and you're on the way to ME hell21:37
DocScrutinizer05normal :-D21:38
wpwrakwell, as i said, you can panel-mount. that's designed to solve precisely this sort of problem21:38
sb0of course, attaching the mainboard, HDMI/DVI connectors, etc. is again another story ...21:38
sb0yeah that's what I'm going to do21:38
sb0it's interesting that in their product brochures button manufacturers always show panels with only one type of button :-)21:39
wpwrakor if you feel adventurous, you can mix flexible circuit with PCB and have a suitable support structure for the little PCBs. 21:39
wpwrakhehe ;-)21:39
DocScrutinizer05or you do what they did in ancient times, when the push-pull powerbutton of Hameg scope been at front panel but the actual switch at rear panel21:39
sb0but then I have to build axes of suitable lengths21:39
wpwrakmore messy mechanics21:40
sb0and yes21:40
sb0I don't have so much space in the case anyway21:40
sb0note that there's only ONE way of doing it for less than 500¬ per raw case, and that's using an already made diecast aluminum enclosure21:40
DocScrutinizer05s/at rear/close to rear/21:40
sb0and the choice is extremely limited there21:40
sb0(additionally those cases have surface finish problems)21:41
wpwrakit's all a question of volume :)21:41
sb0and sets the height of the device to a 63mm that costs a fortune to be changed21:41
wpwraki find premade cases too inflexible21:41
wpwrakwhy not mill the stuff out of acrylic ? that may be cheaper in the end. you can buy a lot of machine-hours for EUR 500 ...21:42
sb0acrylic doesn't look good and is fragile21:43
wpwrakif you need Al for strength, use plates or simple profiles21:43
sb0I'm using this right now21:43
sb0and mill it21:43
sb0of course, you will not have missed the idiotic screw in the middle21:43
wpwraklooks pretty bomb-proof ;-)21:44
sb0which is going to bump into any reasonably placed connector in any design21:44
sb0so of course I'm going to have to mill that crap away21:44
sb0but it's the least shitty I've found that doesn't cost 50k¬21:44
wpwrakscrew in the middle ? i see 8 holes and 6 screws ...21:44
sb0the screw in the middle of each side21:45
sb0where you always want to put a connector ...21:45
wpwrakuse smaller connectors :)21:45
wpwrakcentered connectors are boring anyway21:45
wpwrakhow painful will be milling into the sides ? that's kinda difficult with the simpler type of mills21:46
sb0oh, tell me about it...21:47
wpwrakgoing 4-5 axis ? :)21:47
sb0note that the side of the case is sligthly angled, too21:48
wpwrakwell, you could also etch the critter21:48
DocScrutinizer05I'd going to ask details like "what device is this meant to become, what mounting and environmental conditions, which number of builds, and so on... but I feel like having a walk to burger king21:48
wpwrak(angled) ah, didn't notice. more fun for the mill :)21:48
sb0yeah, right?21:48
DocScrutinizer05that can however looks like made for RF, or automotive 21:49
sb0they HAD to do that21:49
sb0and that stupid screw in the middle21:49
DocScrutinizer05and with automotive I mean engine controller, mounted next to engine21:49
wpwrakso have you found a shop that will mill it for you ? or do you now have a nice industrial-grade mill in your living room ? :)21:50
sb0yeah, I found a shop21:50
sb0which was a bit of fun, too21:51
DocScrutinizer05for an exposed visible case it looks terribly ugly21:52
sb0yeah, it needs some surface finish21:52
DocScrutinizer05even then21:52
DocScrutinizer05I guess there's a reason why you don't go for dirt cheap simple double U steel case?21:53
DocScrutinizer05or RF-shielding to the max21:54
DocScrutinizer05this case doesn't look like it was made for any other purpose or usecase21:54
sb0hm, what's wrong besides the surface finish?21:54
DocScrutinizer05well, you named a few points already21:55
sb0but nothing about the look21:55
wpwrakDocScrutinizer05: double-U stell is even more fun to mill ...21:56
DocScrutinizer05I can't comment for the look, it looks assuringly sturdy... for a thyristor ignition in your car21:56
DocScrutinizer05wpwrak: depends21:56
DocScrutinizer05you can always mount it at edge of mill desk, with the huge part of it downwards below the table21:57
wpwrakwell, the set of mills that can do steel is a subset of the set of mills that can do Al :)21:57
sb0I've tried painting one, and it doesn't look like something to be mounted next to a car engine anymore21:57
DocScrutinizer05the get a double U alu case ;-P21:57
wpwrakhmm, in many case, the part that goes below would hit something21:58
sb0of course, that paint doesn't stick well and scratches etc.21:58
wpwrakwhat paint did you use ?21:58
DocScrutinizer05what's this? alu? painting alu is an art21:58
sb0just spray paint21:58
sb0tamiya color21:58
sb0and yes it wasn't painted right21:59
sb0I was just trying out how it would look like with a different surface21:59
sb0and it looks fine21:59
wpwrakacrylic. yeah, that would explain the scratching :)21:59
sb0hammond mfg also have a powder coated black version21:59
sb0haven't gotten my hands on one of them yet22:00
wpwrakfor that sort of thing, you'll probably want epoxy paint. also available as aerosol. (or airbrush it)22:00
DocScrutinizer05powder coating is for sure better than painting22:00
sb0would be nice if it looked acceptably good, so I wouldn't have to deal with powder coating it myself22:00
sb0in addition to all the other expensive and painful mech problems in this project22:00
wpwrakthe powder coating applied before the milling may not survive the latter too well22:00
sb0according to hammond it does22:01
sb0also hammond can do the milling, so it becomes their problem :-)22:01
DocScrutinizer05meh, PE protective film22:01
wpwrak(one-stop shopping) neat :)22:01
sb0also it'll be somehow satisfying to make them mill away their stupid center screw22:02
wpwrakyou'll still have to plug the hole with something ...22:03
DocScrutinizer05honestly, those cases are a PITA, I seen them and dealt with them. E.G they often lack any mounting support inside22:03
sb0I'll just use screws + spacers on the bottom to attach the PCB22:03
sb0that's actually not too hard, for once22:04
wpwrakhow about an acrylic "sandwich" with Al caps ? that may also look pretty nice, easier to mill, and probably sufficiently solid22:05
DocScrutinizer05so you have screws going though the bottom and then you put spacers over them from the other side and then place the PCB on those screws inside that "pool" and then put nuts on the screw, also inside the pool, while holding the 4(?) screws from the outside so the don't fall out?22:05
sb0like the M1 case but with aluminum on top/bottom?22:06
DocScrutinizer05good luck!22:06
Action: DocScrutinizer05 already prepares the mdical emergency packet to send to sb022:06
sb0DocScrutinizer05, the cover is the bottom22:06
wpwrakno, like a cigarette box with Al (with a bit of a sidewall) on the small sides22:06
DocScrutinizer05then good luck mounting all your front panel components22:06
sb0attach PCB to cover with screws and spacers, and mount cover with PCB on it22:07
sb0after the front panels components are attached of course22:07
wpwrakso you'd have to find suitable Al caps, but the rest would be relatively simple. you'd still need a 4D mill, or a total of >= 3 acrylic pieces for the "sandwich"22:08
sb0oh and btw, the spacers are threaded22:08
sb0so they hold with just the screw22:08
sb0and the front panel PCB attaches to the main PCB with the ribbon cable I mentioned earlier22:08
sb0my first idea was to stack it directly, but it's even more mechanical hell, e.g. because all the "standard" parts are incompatible as I said22:09
DocScrutinizer05well, up to you. I get headache seeing such stuff again. Last time I had to touch a similar design to repair sth, I swore to never again go near any of those22:11
wpwrakdid you have a look into the fadefox critter ? that device has a ton of different elements22:11
sb0I was planning to...22:12
DocScrutinizer05what's spritzdruckguss in english? Just to let you kmow what I rather don't want to touch again22:12
sb0financial disaster?22:13
DocScrutinizer05high pressure die casting22:13
sb0yeah, alternative translation22:13
DocScrutinizer05that case is exactly that22:14
sb0what do you recommend then?22:15
DocScrutinizer05and prolly cost an arm and a leg22:15
whitequarkinjection molding is apparently the more common term22:15
DocScrutinizer05whitequark: right22:15
Action: DocScrutinizer05 blames ding22:16
wpwrakdie casting doesn't have to be super-expensive. you can also make die from Al. they last only for about 1000 uses, but that's probably good enough to get started22:16
DocScrutinizer05I mean, there are like a zillion relatively cheap and often convenient cases available22:16
wpwrakthe company from which i bought my mill made little demo pieces of translucent plastic that produced a very good-looking image of Carlos Gardel when held against the light22:18
sb0with the holes that I want?22:18
sb0I doubt it22:18
wpwrakthey said the milled the die (one piece) themselves. then the production cost something like USD 1000 (not entirely sure about the price, but in any case it was pretty low)22:18
sb0but then should any pot etc. change for any reason (eg going out of stock) I have to make a new mold22:19
sb0not to mention mistakes22:20
wpwrakyou can probably make samples at home, e.g., with wax. when things fit, you move to the fab with plastics22:20
wpwrakand yes, if your geometry changes, the tweaking begins22:21
wpwrakbut then, if you've been reasonably careful with your choice of components, that shouldn't happen all that often22:21
wpwrakDocScrutinizer05: this one looks promising: http://www.conrad.de/ce/de/product/750856/ABUS-Leergehaeuse-fuer-Alarmsirene-oder-Blitzleuchte-SG3210;jsessionid=7476732B7BA807B20A7B8A9C03F63162.ASTPCEN18?ref=searchDetail22:22
sb0DocScrutinizer05, so what exactly is wrong with the hammond die-cast aluminum enclosure, that another off-the-shelf case would solve better?22:23
wpwrak(and tweaking can just mean to manually modify a case, e.g., add spacer material or dremel off something. doesn't have to be too fancy.)22:24
DocScrutinizer05it's just clumsy to handle, starts with the middle screw that is only there for RF/IP6722:24
sb0that can be milled away22:24
wpwrak(and spacer material itself can have a complex structure)22:25
DocScrutinizer05then the bulges22:25
sb0what bulges? inside?22:25
sb0they don't bump into anything (except the middle one), my PCB is smaller than the case22:25
DocScrutinizer05then it#s often impossible to mount stuff to the case, into that cup22:25
wpwrakthe middle screw can probably stay. you can always cut the column after the screw. since you'll have a rather flexible mill at hand anyway ..22:26
DocScrutinizer05sometimes it even breaks on closing the screws22:26
sb0those cases break on closing the screws?22:26
DocScrutinizer05I had some that did22:26
wpwrakwhen DocScrutinizer05 forgot to close the mercury bottle ... ;-)22:27
DocScrutinizer05it's 'guss'22:27
wpwrakand yes, not all Al are identical. there are fairly brittle alloys22:27
DocScrutinizer05mutch too hard and brittle22:27
wpwraki suppose you just have to trust them that they did their part properly22:27
sb0that's hammond's problem, yes22:28
Action: DocScrutinizer05 waves22:28
Action: DocScrutinizer05 closes that conrad browser window22:28
DocScrutinizer05I don't even know if you maybe actually build a thyristor ignition for car engine. So why should I contine trying to figure better alternatives for an unknown usecase, just because I think that this case is for exactly two usecases which I dunno if you maybe even have one of those two22:30
wpwrakthey do specify a "recommended torque". that  suggests that they have at least thought about the issue :)22:30
DocScrutinizer05I guess you could build your device even inside a glass bottle. Will it be easy? prolly no. will it be unique? for sure. Will it be expensive or inconvenient? hell yes22:32
wpwrakDocScrutinizer05: he's building a video mixer. digital in, digital out. with a few control knobs and faders and a display22:34
DocScrutinizer05or studio?22:35
DocScrutinizer05then it's actually not all that bad a case22:35
wpwraksb0: oh, and if you don't mind a little plastic, an approach with top and bottom Al plates could actually be nice, too22:36
DocScrutinizer05once assembled those critters are relatively sturdy, and heavy enough to not move around on somebody taking a deep breath22:36
DocScrutinizer05the don't tend to "sing" when you hit them22:37
sb0yeah, and video cables are a bit stiff, so it's good to have weight22:37
wpwrakmake two plastic shells (split at mid-height for connector openings) for the sides, put holes into them for the vertical metal spacers, 22:37
DocScrutinizer05for stage equipment this is actually the right case22:37
wpwrak(move around) maybe glue some rubber mat at the bottom22:38
sb0I was planning to glue rubber feet22:38
DocScrutinizer05glued feet will vanish during weeks22:39
sb0hmm, I've seen some good quality ones22:39
DocScrutinizer05hard to find, but yes, the exist22:39
wpwrakif they vanish, you can sell spares ;-)22:40
DocScrutinizer05mat sounds nice22:40
wpwrakyou can even hide the screws under it :)22:41
DocScrutinizer05since the screws are at front panel aiui22:41
DocScrutinizer05ooh nope22:42
sb0yeah, I still need to figure out a way - if possible at all and somehow reasonable - to avoid screws on the front panel22:47
sb0maybe by welding those threaded spacers, but that sounds difficult and expensive22:47
wpwrakyou could have a thin Al sheet on top, as "front panel"22:48
sb0anyway, bedtime for now22:48
wpwrakthen you don't need to paint the rest of the box so much22:48
wpwraksweet dreams :)22:48
whitequarkI just realized why a (metalworking) mill is called so23:11
whitequark(because back in the days it would be powered by a water wheel or wind and some blades... like a grain mill!)23:12
whitequarkDocScrutinizer05: you can prolly answer my question23:15
whitequarkin file:///home/whitequark/Downloads/327216.pdf they show that an USB 3.0 *cable* (among other parts) will emit RF noise23:16
whitequarkbut how?! the signal is transmitted over a differential twisted pair in shielding23:16
whitequarkthere should be no emission, not even at -95dB that they show23:16
whitequarkactually I'm lost as to whoever genius selected 5GHz as the carrier frequency, knowing that it will interfere with the ISM band23:18
wpwraki suppose no shielding is perfect ?23:21
whitequarkwell, it seems like a rather lot of shielding machinery thrown on the problem23:22
wpwraki mean, it's a plug, so there must be some gaps. and since RF loves to crawl on surfaces, ...23:36
--- Mon Apr 21 201400:00

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