#avr | Logs for 2012-01-22

Back
[00:19:28] <pingec> Hmm I tried to start a debugwire session via avr studio 5 and it didn't work as planned. Now the chip is stuck in dwen mode. I remember someone saying it's possible to reapir this through an avr studio 4 command line tool ?
[00:52:29] <ziph> pingec: There's an option in the Debug menu to switch it back.
[00:54:35] <pingec> in which studio?
[00:54:46] <ziph> Oh, Studio 4.
[00:54:49] <ziph> I haven't used 5.
[00:55:03] <pingec> this is weird
[00:55:17] <pingec> as5 upgraded the dragon fw
[00:55:21] <ziph> It uses DebugWIRE to do it though.
[00:55:23] <pingec> now as4 wants to downgrade it
[00:55:24] <pingec> :)
[00:55:27] <ziph> Yeap.
[00:55:58] <ziph> You might want to check the DebugWIRE application note and make sure you're not doing something on the hardware side that would stop it from working.
[01:01:37] <pingec> hmm, cant really find it
[01:01:47] <pingec> I have a blank as4 now
[01:02:42] <pingec> I let it downgrade dragon's fw and selected appropriate target under Tools -> program AVR -> Connect...
[01:31:58] <rafasc> hello guys. is there a matrix calculation functions in standard avr C?
[01:38:07] <ziph> raek: No, what do you want to do?
[01:58:39] <rue_mohr> do you want a library for manipulating 3d points in C?
[01:59:23] <rafasc> i was trying to implement the kalman filter
[02:00:27] <Roklobsta> learn how to do fixed point maths so you can just use integers. it will help if you can try and define the dynamic range you need in your computations
[02:01:20] <rafasc> i found an arduino library but it used floating point math...
[02:04:12] <rue_mohr> looks like a great space to build mecha
[02:04:17] <rue_mohr> too bad he left
[02:05:28] <_abc_> hi rue_mohr
[02:05:37] <rue_mohr> hello
[02:05:46] <_abc_> what was that about 3d lib? affine transformer?
[02:06:10] <rue_mohr> someone asked for a matrix lib in C
[02:06:19] <_abc_> ya. Affine right?
[02:06:25] <rue_mohr> I have a point manipulation lib, as actaully doing matrixies are slow
[02:06:29] <_abc_> I wrote 2
[02:06:30] <rue_mohr> uh
[02:06:36] <rue_mohr> mm whoever
[02:06:47] <_abc_> well not released but working fine
[02:07:11] <_abc_> So what do you call point manipulation?
[02:07:31] <rue_mohr> translate, rotate, scale
[02:07:39] <rue_mohr> in singles or groups
[02:07:40] <_abc_> without matrices?
[02:07:44] <rue_mohr> correct
[02:07:59] <_abc_> Did you time your code to see if it is indeed faster?
[02:08:00] <rue_mohr> too much junk overhead in actually doing matrixies
[02:08:10] <rue_mohr> it does less math, its faster
[02:08:26] <_abc_> But did you time it? Sorry for being an ass about this?
[02:09:10] <_abc_> Also did you go further to surface direction vectors and z buffer clipping or is it just enough for wireframe for now?
[02:09:10] <rue_mohr> it does the same thing as the matrixies, without the pointless copying of data
[02:09:14] <rue_mohr> no I didn't time it
[02:09:29] <_abc_> what copying of data?
[02:09:34] <rue_mohr> no surface normals or clipping
[02:09:55] <rue_mohr> its a lib made work forward/reverse kinematics of robots
[02:10:06] <_abc_> yes
[02:10:15] <_abc_> I guessed that much :)
[02:10:24] <_abc_> Nice
[02:10:36] <rue_mohr> http://www.youtube.com/watch?v=TrR8moC3qbA
[02:10:45] <_abc_> You know that an arbitrary affine trasform set can be convolved into just one, right?
[02:10:46] <rue_mohr> thats all done in 3d and projected after
[02:11:16] <rue_mohr> abc yup, not that I'v ever seen anyone do it
[02:11:25] <_abc_> So, eventually, whatever transform you need to do, it ends up with multiplication with just exactly one matrix
[02:11:29] <_abc_> which can be sparse too
[02:11:41] <_abc_> So that's the way I did it
[02:11:59] <rue_mohr> yup
[02:12:19] <_abc_> okay cute
[02:12:25] <rue_mohr> I still ned to work on making a robot dance
[02:12:36] <rue_mohr> I wrote a bvh library
[02:12:38] <_abc_> Why did viewing that avi cause my passkey-agent to generate a lot of cpu load?
[02:12:48] <_abc_> bvh?
[02:12:57] <rue_mohr> I had to, all the libraries I could ind that wored at all, required fixed models
[02:13:14] <rue_mohr> bvh is a model and motion file
[02:13:37] <_abc_> I tried it again, again passkey-agent high load as if something is hammering it
[02:13:47] <rue_mohr> I'm planning to use it for canned motion on a robot dancing
[02:14:07] <rue_mohr> want the actual file?
[02:14:46] <rue_mohr> 1.7M
[02:16:58] <_abc_> erm no thanks
[02:17:11] <_abc_> rue_mohr: so your bot is a humanoid?
[02:17:12] <rue_mohr> part of the reason for the bvh parser was cause I wanted to studdy how the models were done for linking the 12' mecha
[02:17:30] <rue_mohr> no, the file can do any model
[02:17:42] <rue_mohr> its a tree model
[02:18:04] <rue_mohr> most trees start at the hip and work out
[02:18:51] <rue_mohr> the motion files were scraped from the internet
[02:19:04] <rue_mohr> I just made the program that renders them
[02:19:16] <rue_mohr> but my library can create / modify them too
[02:20:46] <_abc_> I see, so it is all virtual?
[02:21:15] <rue_mohr> those files were created with a motion capture package
[02:21:30] <rue_mohr> I dont know what one, many of the motion capture systems output bvh
[02:21:56] <_abc_> I have no idea about that, bvh is new to me
[02:22:14] <rue_mohr> mine would be fully virtual, to the point when they were applied to the robot anyhow
[02:22:20] <_abc_> I should pull up the library and work on it some more. I mean mine. I need it to visualize some cnc stuff
[02:22:25] <rue_mohr> bvh is really old, but its out there
[02:22:33] <rue_mohr> its the only thing like it, so it seems
[02:22:48] <_abc_> Do you have any worthy links to pocketing algorythms?
[02:22:59] <rue_mohr> hell no!
[02:23:20] <rue_mohr> do you want one, or a few?
[02:23:27] <_abc_> https://en.wikipedia.org/wiki/Biovision_Hierarchy this?
[02:23:29] <rue_mohr> spiral cant be hard, can it?
[02:23:33] <rue_mohr> yes
[02:23:47] <_abc_> rue_mohr: spiral is one, line raster another, there are many
[02:23:54] <rue_mohr> yea
[02:24:16] <rue_mohr> I'v cringed when I was working with mastercam and thinking about how its coded
[02:24:19] <_abc_> rue_mohr: the devil is in the details, specifically how to handle non convex surfaces and how to handle narrow areas which will be contoured too
[02:24:31] <_abc_> cringed how?
[02:24:34] <rue_mohr> as I cringle trying to work on how to fill any polygon
[02:24:40] <_abc_> cnc algorithms have been around since the 1950s
[02:24:47] <rue_mohr> yes
[02:25:17] <rue_mohr> iirc even filling a triangle could be a challange sometimes
[02:25:20] <_abc_> rue_mohr: of any convex shape is dead easy, you just raster scan it to intersect lines and build a meander continuous path inside it from the 'inside' lines
[02:25:40] <rue_mohr> its when its not tho
[02:25:49] <rue_mohr> I'v seen lots of software get screwed up
[02:26:17] <_abc_> rue_mohr: trivially you can decompose any non convex shape into convex shapes but this leads to inefficient working as the tool needs to jump. So you end up flipping path directions, subdividing convex scan paths and other horrors which are best coded in LISP imho. At least I did most of what I did in Tcl which is lispish
[02:26:25] <rue_mohr> ok
[02:26:39] <rue_mohr> me and lisp dont get along
[02:26:47] <rue_mohr> or rpn
[02:26:56] <_abc_> And it works okay once you get the hang of it and pay attention to boundary conditions
[02:27:02] <rue_mohr> or at&t assembler
[02:27:31] <rue_mohr> de3stination, source, what kinda goofy operand order is that????
[02:27:31] <_abc_> hey we all have out limitations. I don't like basic although I learned that many years ago and used it well
[02:27:48] <_abc_> rue_mohr: a=2 is like that. They just used it so.
[02:28:00] <_abc_> rue_mohr: see <dest> = <src>
[02:28:11] <_abc_> now replace the = with ,
[02:28:26] <_abc_> because of lambda calculus syntax requirements
[02:28:32] <_abc_> and you have at&t
[02:28:35] <rue_mohr> ... was at&t src,dest or dest,src?
[02:28:51] <_abc_> dest,src I think
[02:29:01] <_abc_> intel has it backwards
[02:29:09] <_abc_> or not?
[02:29:16] <rue_mohr> :
[02:29:19] <_abc_> mov cx, dx has cx as dst
[02:29:19] <rue_mohr> :)
[02:29:23] <_abc_> hmm
[02:29:24] <rue_mohr> I hate this new keyboard
[02:30:04] <rue_mohr> I wrote a formula parser,
[02:30:07] <_abc_> well the other paradigm is to put the 'result' pointer at the top of the stack of parameters since it is the one which is unique if any assignment takes place
[02:30:19] <_abc_> so argv[1] is always the <dst>
[02:30:21] <rue_mohr> its definitly a choar to parse math as we know it
[02:30:23] <_abc_> but hey
[02:30:28] <rafasc> im sorry, i was busy.. i think i'll use the floating point lib. at least for now. thanks.
[02:30:52] <_abc_> rue_mohr: did you see the clever old Fortran expression precedence parser thing?
[02:30:57] <_abc_> It is cool and cheap to implement
[02:31:17] <rue_mohr> fortran, cobol, lisp, rpn --> bucket
[02:31:29] <rue_mohr> :)
[02:31:38] <nevdull> i think i'd agree with rue_mohr in at least i think 'mov eax, [ebx+20h]' is easier to grok than 'movl 0x20(%ebx), %eax'
[02:32:23] <rue_mohr> I should make myself code some assembler sometime soon, I'm uber rusty
[02:34:03] <nevdull> there are some solid assemblers out there now for linux and win besides gas
[02:35:26] <rue_mohr> I'd like to know how it is that they manage to use all of the 168 pins on microcontrollers in printers on subcontroller baords that only have 1 stepper, 1 opto, and a serial line
[02:36:24] <rue_mohr> I just dont understand how they do it
[02:36:50] <rue_mohr> there must be as many nodes in the discrete circuitry as they have pins on the controller
[02:37:07] <rue_mohr> they should be abel to single out failure of any component
[02:37:20] <rue_mohr> 'board 5 reports R9 failed'
[02:37:24] <rue_mohr> geeeesh
[02:38:06] <rue_mohr> nevdull, whats most of your assembler been for?
[02:38:27] <nevdull> ARM and AVR, although i've done some x86_64 on linux
[02:38:43] <rue_mohr> arm eh? hmm
[02:39:05] <rue_mohr> ugh, its only 12:30, I dont want to be tired now
[02:39:06] <nevdull> yah ARM9 (and even ARM11) boards are so cheap now...
[02:39:11] <rue_mohr> MORE TEA
[02:39:23] <rue_mohr> hardware or software stuff for the arm?
[02:39:36] <rue_mohr> (like controlling a robot or running webserver)
[02:39:53] <nevdull> i've done some linux arm as well as baremetal eabi
[02:40:04] <nevdull> mostly for intelligent home automation
[02:40:34] <rue_mohr> user entered kitchen, turn lights off and lock cupboards
[02:40:47] <nevdull> the closest i've gotten to controlling a robot is building a little atmega644p-controlled tread bot
[02:40:49] <rue_mohr> no sugar or salt for you!
[02:40:53] <keenerd> I'm still wringing the last few ccles from my arm4 linux box :-)
[02:41:21] <nevdull> lol@keenerd. i was sad that ubuntu doesn't support arm4
[02:41:35] <nevdull> but fedora and cent os works ok
[02:41:36] <rue_mohr> you familiar with hwo to bootstrap an arm board?
[02:41:43] <rue_mohr> I have a few
[02:41:49] <_abc_> rue_mohr: http://en.wikipedia.org/wiki/Operator-precedence_parser#Alternatives_to_Dijkstra.27s_Algorithm fyi
[02:41:51] <nevdull> rue_mohr: i've mostly used u-boot
[02:42:00] <keenerd> nevdull: Don't let that stop you. ArchlinuxARM does not support V4 either, so I ported it myself.
[02:42:14] <nevdull> gasp@keenerd. that's awesome
[02:42:33] <rue_mohr> oh I did see that when I was building my parser
[02:42:43] <rue_mohr> it was pretty good, but there was a problem
[02:42:50] <nevdull> what kinda parser?
[02:43:02] <rue_mohr> I made a math expression parser
[02:43:15] <keenerd> Compiling everything on the little box with 32MB ram is a whole new level of masochism.
[02:43:16] <nevdull> use an AST?
[02:43:53] <rue_mohr> ast?
[02:43:53] <nevdull> haha yah they chug along. i've lately started having a love affair with a dual-core atom board.
[02:43:59] <nevdull> abstract syntax tree
[02:44:09] <keenerd> AST comes after the parser though.
[02:44:19] <nevdull> yah
[02:44:43] <rue_mohr> I parse out the elements of the input, assign prescedences to operators, and boil it down highest priority operators first
[02:44:49] <nevdull> it's more sparse than a concrete tree...well, that is, it is 'syntax' and not 'semantics' of the source code
[02:44:57] <_abc_> keenerd: haha young whippersnappers. I compiled linux on a 8MB ram 25MHz 386SX once
[02:45:07] <rue_mohr> Iirc I did it as a list of somekind
[02:45:07] <_abc_> the kernel at least
[02:45:22] <_abc_> it took 10 minutes to boot x11 into a xterm on b/w screen
[02:45:26] <nevdull> _abc_: thatwas my first box: 386sx w/ 8MB RAM
[02:45:27] <rue_mohr> abc I think I did that, takes baout 4-6 hours?
[02:45:33] <nevdull> running slackware
[02:45:34] <_abc_> rue_mohr: yea
[02:45:38] <_abc_> exactly slackware
[02:45:42] <rue_mohr> yea it was slakware
[02:45:48] <nevdull> that you had to download in those 1.4Mb archives on a 14.4 modem
[02:45:56] <_abc_> hehe
[02:45:59] <_abc_> those were the days
[02:46:10] <_abc_> when I used to run up $100 phone bills
[02:46:21] <keenerd> _abc_: I could not afford spare hardware to play with linux back then. Though I finally picked up a 16 bit orronco wifi card and I do intend to use it on a system like that :-)
[02:46:23] <rue_mohr> back in the says when installing a package ment manually gathering all the right versions of all the right dependencies and their dependencies
[02:46:28] <nevdull> haha...i had a SLIP account with the uni that i downloaded all my stuff thru
[02:47:03] <_abc_> nevdull: yes but you still had to pay for phone use
[02:47:21] <nevdull> it was just a local call for me. no long distance. flat rate phone bill.
[02:47:22] <rue_mohr> I never got over 14.4 from the local isp
[02:47:30] <rue_mohr> they blamed the telco
[02:47:38] <_abc_> I once left a modem connection to a shell account I had open for a day and a night because it looked exactly the same like the other vc's on linux... aieee phone bill
[02:47:48] <nevdull> hehe
[02:48:02] <_abc_> rue_mohr: I heard lots of NA have seriously old infrastructure
[02:48:05] <keenerd> Yay only having one phone line + dialup :-(
[02:48:18] <_abc_> keenerd: still?
[02:48:23] <rue_mohr> nothing like accidently forgetting to tell the modem to disconnect and having it rack up isp minutes all night
[02:48:24] <keenerd> No, back then.
[02:48:54] <nevdull> ATDT911 was my favorite gag
[02:48:56] <rue_mohr> esp when you have a ping script so the isp wont drop you for timeouts
[02:48:59] <keenerd> I think my modem would melt if I tried to update a distro over dialup nowadays.
[02:49:26] <rue_mohr> the fun of 'ah, only 64 megs of updates, I'll go get a tea'
[02:49:35] <_abc_> keenerd: it's bad enough the suck that windows is tries to update every time it gets a chance, even over gprs or crummy 802.11b wifi
[02:49:42] <_abc_> keenerd: you can imagine what that can cause
[02:49:46] <keenerd> Ew.
[02:50:04] <nevdull> when the only way to get a linux kernel on your system was to compile the 0.99pl version by hand and it took HOURS
[02:50:24] <_abc_> It's not just windows, all the crap installed on windows boxes tries to call home and push updates all the time. Java, antivirus, even visual studio I think
[02:50:24] <rue_mohr> I think I remember trying to cap any one file size at about 10M, and I had to be really desperite to get a 10m file
[02:50:32] <nevdull> _abc_: i've been wanting to play with zigbee. know anything about it?
[02:50:44] <_abc_> nevdull: only that it can jam 802.11 quite well
[02:50:58] <rue_mohr> I went from 3.11 to linux, cause ms never made anything better than 3.1
[02:51:05] <_abc_> heh
[02:51:15] <nevdull> ah, is that so? that wouldn't do around here
[02:51:16] <_abc_> I went from 3.11WfWG to linux slackware
[02:51:18] <rue_mohr> I waited till after 95 and all they came out with was 98
[02:51:33] <nevdull> Windows ME
[02:51:39] <_abc_> nevdull: I mean if you push data hard on zigbee it will jam 802.11 off the air
[02:52:00] <rue_mohr> by ME people were starting to flood over to NT, which they hated
[02:52:22] <_abc_> Well NT was more or less the first windows which actually was workable
[02:52:25] <_abc_> Sort of
[02:52:32] <nevdull> i liked windows for workgroups
[02:52:41] <_abc_> no more winsock.dll, 32 bits and so on
[02:52:44] <_abc_> nevdull: eww
[02:52:47] <nevdull> and their big competitor novell
[02:52:51] <_abc_> EWW
[02:53:08] <_abc_> novell network control setup commands are about the worst command language ever devised imho
[02:53:23] <nevdull> although at the time, my main box was a Sparc SuperStation 20 running Solaris 2.3
[02:53:32] <rue_mohr> I used ARCNET in my room, it was the only adapters I could affordably get my hands on
[02:53:34] * nevdull agrees with _abc_
[02:54:00] <rue_mohr> used the coax with 75R cablevision wire
[02:54:05] <rue_mohr> :)
[02:54:14] <_abc_> heh it does not matter over short lengths
[02:54:29] <nevdull> coax is a forgotten medium *sigh*
[02:55:01] <_abc_> Nah I had 10base2 ether for quite a while. Bought a stash of NE2000 based isa cards and I used that with the little Ts and terminators for a long time
[02:55:02] <rue_mohr> I got one link in ethernet coax and then I got a load of utp cards
[02:55:18] <_abc_> or is it 10baseT
[02:55:23] <nevdull> yah those ne2000's were everywhere
[02:56:12] <_abc_> well strictly they were RTL8039s but still...
[02:56:29] <_abc_> wow that was like 15 years ago or longer
[02:56:48] <_abc_> time flies ;)
[02:57:11] <nevdull> hard to believe. about the same time i was calling my brother up and exclaiming "go to <the local computer store> and pick up a new 420MB harddrive. they're only $1 a meg!"
[02:57:22] <_abc_> heh a lot those who are on irc now here were just playing with dolls 15 years ago
[02:57:36] <_abc_> nevdull: hehe
[02:58:09] <pingec> we are lucky to even know what irc is
[02:58:13] <pingec> :P
[02:58:17] <_abc_> heh
[02:58:46] <_abc_> It's called 'chat' now and you see a fancy java client if you access it in the 'approved' [tm] way. Or a web client
[02:59:05] <_abc_> Either way, it is VERY different from green (or black) terminals
[02:59:17] <pingec> :)
[02:59:17] <nevdull> oh the days of ircII and riding netsplits to get ops on a channel
[02:59:24] <_abc_> I mean from those we still use for irc :)
[02:59:24] <rue_mohr> hehehe
[02:59:41] <_abc_> nevdull: you sound like a naughty one
[02:59:55] <_abc_> 911 calls, riding netsplits.
[02:59:59] <nevdull> _abc_: lol
[03:00:01] <_abc_> 3v11
[03:00:02] <rue_mohr> the evil or it, the first time I'z told about irc, some guys told me to meet them, but didn't say what server or channel
[03:00:12] <_abc_> oh
[03:00:22] <_abc_> and there were 3 servers to choose from?
[03:00:32] <rue_mohr> ~1995 96?
[03:00:37] <_abc_> guessing a nick is much harder
[03:00:38] <nevdull> yep
[03:00:56] <_abc_> I was on compuserve before getting onto the net
[03:01:02] <rue_mohr> by 98-2000 I was on
[03:01:04] <_abc_> I found compuserve very limiting and chaperoned
[03:01:06] <_abc_> and expensive
[03:01:28] <rue_mohr> if they take down the internet, I wonder if all that will come back again
[03:01:46] <nevdull> yeah, i was still spending a gratuitous amount of time on BBSs and playing BBS door games
[03:01:51] <_abc_> Retrospectively I could have bought 2 top of the line computers from what I paid for net access to compuserve and the phone company and isp from about 1994 to 2000
[03:01:53] <rue_mohr> hheeh
[03:02:06] <nevdull> hehe
[03:02:08] <rue_mohr> did a bunch of the bbs' go online?
[03:02:41] <nevdull> yeah, the big ones running, like, Wildcat BBS or that other big one, had internet gateways for UUCP email
[03:02:46] <rue_mohr> I think I used to join up with the paper mills bbs
[03:03:07] <rue_mohr> wow, that takes me back
[03:03:21] <rue_mohr> bunch of downloadable programs I wasn't interested in
[03:04:25] <rue_mohr> 1am! ok, cmon rue, stay awake
[03:04:33] <nevdull> the bbs i frequented served a CD-ROM full of program code
[03:04:35] <keenerd> A bunch? So what, tens of megabytes?
[03:04:38] <rue_mohr> maybe if I go work over the tablesaw or soemthing...
[03:05:03] <rue_mohr> keenerd, definitly 10's of megs
[03:05:16] <rue_mohr> I thinkI started with a 120M drive with dos
[03:05:21] <nevdull> better than downloading with kermit tho
[03:05:25] <rue_mohr> iirc I outgrew that really quick
[03:05:34] <nevdull> or xmodem
[03:05:58] <keenerd> rue_mohr: Whoa whoa whoa, you had a hard drive. At least I can one up someone here :-)
[03:06:09] <rue_mohr> hah
[03:06:45] <rue_mohr> the first machine I had (pc) was a 386-25 I was given, I was gonna strip it for parts but thought it might be workable
[03:06:49] <_abc_> rue_mohr: tablesaw at 1am sounds like a future er client
[03:06:59] <rue_mohr> a buddy helped me round up the rest of the parts
[03:07:26] <rue_mohr> before th pc I had an appleIIe
[03:07:33] <_abc_> ah that's a nice one
[03:07:35] <nevdull> i had a commodore 64
[03:07:40] <_abc_> rue_mohr: did you have a modem for it?
[03:07:45] <keenerd> rue_mohr: Finest machine Apple ever made.
[03:07:50] <_abc_> rue_mohr: were any bbs's active at the time?
[03:08:16] <rue_mohr> as apple decided with the mac that nobody would be developing anymore hardware / software for their machines, the pc was the obvious choice
[03:08:32] <rue_mohr> no modem on the apple
[03:08:38] <_abc_> rue_mohr: that was the biggest shot in the foot Apple ever managed
[03:08:48] <rue_mohr> yea
[03:08:54] <_abc_> rue_mohr: turning away from Xenix towards Windows was the biggest for microsoft
[03:09:02] <nevdull> that's like the reason betamax got creamed by vhs
[03:09:07] <keenerd> Though the IIgs (with built in hard drive and crazy per-mac window system) was pretty good too.
[03:09:15] <keenerd> *pre-mac
[03:09:28] <_abc_> nevdull: between us, betamax had sucky mechanics in all variants of the implementation
[03:09:54] <nevdull> _abc_: vhs had longer play times and the porn industry latched on and the rest is history ;)
[03:09:57] <_abc_> nevdull: vhs was much superior in resistance to bad tapes and mechanical problems caused by consumer use
[03:10:04] <rue_mohr> hahah
[03:10:36] <keenerd> Sigh. Sony used to be a defender of consumer rights back in the Betamax days. How they have fallen.
[03:10:37] <_abc_> nevdull: no it was not that. The entire head drum was rotating with vhs, making the tape 'fly' on an air cushion entrained by it much like the hard disk heads fly.
[03:10:51] <_abc_> That was crucial for being able to cope with crap tapes and dirt
[03:10:56] <nevdull> _abc_: aah, gotcha.
[03:11:28] <rue_mohr> but the beta pushed air out the slot the head ran in
[03:11:40] <_abc_> rue_mohr: yes but it was not the same
[03:11:58] <rue_mohr> personally I think its less prone to dirt
[03:12:12] <_abc_> rue_mohr: also head change and upper drum alignment on beta were a bitch
[03:12:28] <rue_mohr> I'd belvie it
[03:12:36] <_abc_> in general the mechanics were 10x more complex for beta and that sunk their boat in the price war
[03:12:41] <nevdull> i remember using isopropyl alcohol on a qtip to clean my vhs head
[03:12:58] <rue_mohr> and catching lint and chipping the head?
[03:13:05] <_abc_> hehe
[03:13:12] <_abc_> you used a sheet of paper for that
[03:13:20] <_abc_> precisely for lack of lint
[03:13:24] <rue_mohr> sitting there wondering if youjust really broke it or not?
[03:13:32] <_abc_> hehe
[03:13:59] <rue_mohr> yea, but the paper would get all grey, and you got to sit ther wondering if it was dirt or grind-off from the head
[03:14:14] <_abc_> rue_mohr: nah you can't grind off ferrite with paper
[03:14:25] <_abc_> rue_mohr: you kept doing it until the paper came out white
[03:14:33] <nevdull> the days of getting charged a fee if you didn't rewind the tape before you brought it back to the video rental store
[03:14:39] <rue_mohr> then there was the bit of dirt that didn't go away till you broke out the acetone..
[03:14:48] <_abc_> rue_mohr: yea.
[03:15:56] <rue_mohr> I'd really like to get into pal, why did they never make this easy?
[03:16:07] <_abc_> pal?
[03:16:08] <rue_mohr> 16V22 n such
[03:16:11] <_abc_> oh
[03:16:17] <_abc_> rue_mohr: go to Atmel and get Wincupl
[03:16:29] <rue_mohr> the programming algs still seem to be top secret
[03:16:33] <_abc_> pal programming and simulation
[03:16:47] <rue_mohr> hmm, prolly runs on my 98se machine eh?
[03:16:47] <_abc_> rue_mohr: and no, there are several open source pal programmers, hard and soft
[03:17:01] <rue_mohr> can you find one?
[03:17:05] <_abc_> uhh 98se is unlikely but I think it works in a vm
[03:17:06] <rue_mohr> no luck here
[03:17:08] <_abc_> rue_mohr: sure
[03:17:24] <rue_mohr> I found an old scan of a datasheet for otp pals
[03:17:45] <rue_mohr> the flash ones (palce, gal) dont say how to erase them
[03:17:53] <_abc_> http://www.repairfaq.org/filipg/LINK/F_pal_gal.html
[03:18:40] <rue_mohr> well at last you didn't pull up an armory ftp link...
[03:20:01] <_abc_> http://www.armory.com/~rstevew/Public/Pgmrs/GAL/_ClikMe1st.htm
[03:20:11] <_abc_> this one? >:~
[03:20:35] <_abc_> and yes that should work with linux after you port it to your version
[03:21:05] <rue_mohr> this page says what I been feeling, make myself an eeprom programer and forget pals ever existed
[03:21:30] <_abc_> oh pals are used for a different purpose than eeproms...
[03:22:51] <rue_mohr> yea
[03:23:03] <rue_mohr> but I can make a pal from an eeprom + latches
[03:23:21] <rue_mohr> little dead bug biggybacking etc and boom, pal :)
[03:24:10] <rue_mohr> that first article you gave me just references a bunch of paper docs
[03:24:16] <rue_mohr> its that top secret eh?
[03:24:24] * rue_mohr dusts off his libary card
[03:24:53] <_abc_> rue_mohr: even a cheap old pal will have tprop ~= 20nsec and modern ones reach 2nsec
[03:25:04] <_abc_> rue_mohr: an eeprom is like 5 times slower at least
[03:25:23] <rue_mohr> bios eeproms
[03:25:28] <_abc_> rue_mohr: it's not top secret, it predates pdf datasheets.
[03:25:31] <rue_mohr> have tonnes of hem
[03:25:43] <_abc_> rue_mohr: several refrences are datasheets from semi makers from the early 1990s
[03:25:48] <rue_mohr> but you would think that there would be scans or soemthing of the documents
[03:26:12] <rue_mohr> yea, my databooks are all like 1979, just a bit too early
[03:26:17] <_abc_> rue_mohr: why? I had those phone book sized data sheets around for years
[03:26:27] <rue_mohr> the intel book still has 4004 4040 and 8008 in it
[03:26:33] <_abc_> too old
[03:26:35] <rue_mohr> with a prelim on the 8041
[03:26:43] <_abc_> use as doorstep or sell on ebay to collectors
[03:27:02] <rue_mohr> those books are sometimes the ONLY existing ref for some of those chips
[03:27:19] <_abc_> yes... and?
[03:27:40] <rue_mohr> I'm determined to use up the drawrs of them I have
[03:28:06] <rue_mohr> trying to find a good 8042 datasheet is a job and a half
[03:29:36] <rue_mohr> do you still have ~1990 databooks?
[03:30:08] <rue_mohr> if so, for what manufs?
[03:31:13] <rue_mohr> last shop clenup I considered ditching the 8080 + support in leiu of the fact I have some 8085 and piles of z80
[03:31:41] <rue_mohr> the first computer I scratch built was gonna be a 8080 till I found out about the z80
[03:31:46] <rue_mohr> so z80 it was
[03:34:54] <rue_mohr> I'm working on using an 8254 as a 2 channel pwm generoator
[03:35:12] <rue_mohr> to operate off an ide cable of a motherbaord on my lawn mowing robot
[03:35:31] <rue_mohr> its a terribly convoluded chip
[03:35:50] <rue_mohr> badly written datasheets too
[03:36:19] <ziph> Why don't you build it a CPU from vacuum tubes? :)
[03:36:33] <ziph> build it from a CPU from vacuum tubes.
[04:01:49] <RikusW> http://en.wikipedia.org/wiki/Fluidics
[04:02:02] <RikusW> Now lets build an mcu with that ;)
[04:02:23] <pingec> cool
[04:04:12] <pingec> ancient greece style
[04:05:39] <RikusW> seems it can even be used for an audio amp
[04:10:18] <rue_shop> I have here an electrostatic deflection crt with all the amp baords etc
[04:10:37] <rue_shop> I think its way too slow for use as a scope so I suppose I'll junk it
[04:11:11] <RikusW> why too slow ?
[04:11:26] <rue_shop> well it cant do Mhz
[04:11:33] <RikusW> scopes use electrostatic because its faster than electromagnetic..
[04:11:37] <rue_shop> its from a nerve analizer
[04:11:51] <RikusW> maybe its the amps thats the limit ?
[04:12:00] <rue_shop> its big and I think if I want such a display I can use an lcd
[04:12:14] <rue_shop> I'v seen that fluidics stuff beofre
[04:12:31] <rue_shop> iirc the mil was trying to develop autopilots with it cause its emp proof
[04:13:55] <RikusW> it you build new amps it might be able to go much faster.
[04:13:58] <rue_shop> none the less, this thing has been taking up space for over 7 years wtih me
[04:14:04] <rue_shop> I want to make a digital scope
[04:14:07] <rue_shop> and I have a scope anyhow
[04:14:18] <rue_shop> and I have two old heathkit tube scopes
[04:14:23] <jacekowski> do you know how to make one
[04:14:29] <rue_shop> yea
[04:14:37] <rue_shop> drive to do it is another matter
[04:14:50] <rue_shop> I have a good newer 25Mhz scope
[04:14:54] <RikusW> you can still use it as a display
[04:14:57] <rue_shop> I want a digital for really slow stuff
[04:15:19] <rue_shop> its as big as a small microwave...
[04:15:36] <rue_shop> a large toaster oven
[04:15:53] <rue_shop> its volumetically larger than my good scope
[04:16:13] <RikusW> my scope is a tektronix RM561A quite heavy...
[04:16:13] <rue_shop> I think its pots are worth more to me than it is
[04:16:24] <rue_shop> hmm
[04:17:36] <RikusW> don't use it much
[04:18:05] <RikusW> but some things just can't be done without it
[04:18:44] <rue_shop> see, I just found a 20msps 8 bit adc
[04:18:59] <rue_shop> ok I'm ditching the crt, I'll appologize to it first
[04:20:44] <amee2k> you're throwing away a perfectly good scope? o.O
[04:21:28] <RikusW> not a scope.
[04:22:14] <amee2k> oh, i just thought by crt you meant one of the crt scopes
[04:30:15] <pingec> poor crt
[04:30:21] <pingec> Don't you have a heart?
[04:32:38] <RikusW> If you do, you'll end up with a electronics and computer museum ;)
[04:34:20] <amee2k> a few years ago i got some free 22" CRTs... awesome stuff. bright and razor sharp picture
[04:34:58] <pingec> I know :(
[04:35:00] <amee2k> panasync pro p110i, belinear err something-8080
[04:35:52] <amee2k> after 3 months of trying to find someone who wants them, i gut them for parts
[04:36:29] <RikusW> vga monitors ?
[04:37:17] <amee2k> um... these things do nutty resolutions like 2048 x something
[04:37:40] <amee2k> at >100hz... faster than my 8800gts does lol
[04:38:24] <RikusW> so what screen do you currently have ?
[04:39:04] <amee2k> a 19" 4:3 and 22" widescreen
[04:39:10] <amee2k> both flat :(
[04:39:22] <RikusW> whats wrong with flat ?
[04:39:50] <amee2k> backlights are getting a bit on the dark side and contrast on the 22" sucks
[04:40:22] <RikusW> thats the problem with them
[04:40:33] * RikusW have a philips 17" crt
[04:40:39] <amee2k> i had the 19" one for ages, since i was in school and it is still in pretty good shape
[04:41:16] <amee2k> the 22" is only like 5 years old and the brightness setting is up to 86% now
[04:41:18] <RikusW> my brothers new widescreen lcd does have pretty good contrast
[04:41:53] <RikusW> you'll probably have to replace its cfl's sometime
[04:42:00] <amee2k> yeah
[04:42:28] <amee2k> i've seen replacement tubes for tfts on ebay but never tried replacing them
[04:46:36] <rue_shop> the museum is overflowing
[04:47:02] <rue_shop> the pile in the attic and the box in the shop are fine, half the second livingroom is being a problem
[04:47:33] <pingec> that just means you need a bigger place
[04:47:34] <pingec> :P
[04:48:00] <amee2k> i've got a small cardboard box with some old drives and CPUs as my private little museum
[04:48:28] <amee2k> these CRT monitors would have been just fine to use for production
[04:51:56] * RikusW still have some 14" vga crt's around
[04:55:06] <RikusW> and a heap of old motherboards
[04:55:38] <amee2k> i only have one old P3 box around. just in case i get another retro gaming spree lol
[04:57:23] <RikusW> mine is dead mb's
[04:57:34] <RikusW> plan to use the power mosfets on it
[04:57:42] <RikusW> and whatever else I can
[04:58:03] <amee2k> i only pull the voltage regulators from these and maybe other stuff that looks interresting like oscillator cans
[04:58:14] <amee2k> then throw the rest away
[04:58:45] <RikusW> and then there is gd75232 for rs232 level translation
[04:58:47] <amee2k> the caps are all too often the cause of the failure anyway and even if not they're pretty much finished
[04:59:09] <RikusW> only problem is it wants +-12V
[04:59:21] <amee2k> oh, and i usually take the d-sub, usb and mini-din connectors
[04:59:38] * RikusW have replaced some mb caps and got it to work again
[04:59:49] <amee2k> rs232 line driver without charge pump is like <50 cent new
[05:00:07] <amee2k> me too
[05:00:12] <RikusW> how do you remove the connectors ? solder sucker ?
[05:00:13] <pingec> damn, leave something for the Indian
[05:00:29] <RikusW> or heatgun ?
[05:00:33] <amee2k> mb caps kinda suck because the multilayered boards with huge ground planes and plated holes are a poor man's heatsink
[05:00:45] <RikusW> yeah
[05:01:04] <amee2k> heat the pin with an iron until i get the cat lose, then heat from one side and use the solder sucker to clear the hole
[05:01:08] <RikusW> I heat one pin, pull a little, other pin, and so on
[05:01:21] <RikusW> I do the same in reverse to put back the new one
[05:01:28] <amee2k> once i couldn't heat the whole pad so i got a small drill bit an carefully drilled the solder out
[05:01:52] <RikusW> thats for caps
[05:02:00] <amee2k> some people just crush the cap's can with pliers and pull it off, then remove the leads individually
[05:02:06] * RikusW removed entire PCI connectors intact ;)
[05:02:16] <amee2k> hehe
[05:02:24] <RikusW> heatgun is the only way
[05:02:35] <amee2k> i tried pulling off some pci and isa slots with hot air just to see if i could do it
[05:02:45] <amee2k> no fucking clue what to do with a pci slot lol
[05:03:04] <RikusW> uneven heat == left behind pins...
[05:03:36] <amee2k> after a while i found out just how exactly i have to wiggle the slot around to see which pins are lose and which are still stuck
[05:04:01] <RikusW> without burning your fingers ? ;)
[05:04:05] <amee2k> yep :)
[05:04:59] <amee2k> i put thumb and index finger on the top ends of the slot, then slightly rock it back and forth
[05:05:22] <amee2k> anyway, gotta run now
[05:11:12] <OndraSter> mornin
[05:58:21] <OndraSter> hmm how am I supposed to fit into this
[05:58:21] <OndraSter> http://clip2net.com/s/1vEOV
[06:00:38] <OndraSter> smaller pad :P
[06:06:53] <HKCMD> how can i show increasing values from 1 to 30 on an LCD display when i continously press a button ?
[06:09:51] <OndraSter> make a counter
[06:09:55] <OndraSter> and each time the button is pressed
[06:10:02] <OndraSter> jump to the 1st position on the display and rewrite the number
[06:10:07] <OndraSter> oh and increase the counter ofc
[06:10:08] <HKCMD> k
[06:11:03] <HKCMD> yes, inside loop counter and update the value, isnt it ?
[06:11:10] <OndraSter> ye
[06:11:15] <HKCMD> thanks
[06:11:47] <OndraSter> np
[06:13:49] <HKCMD> if I want to change if i press continously, means without taking the finger from the button ?
[06:15:05] <OndraSter> set up internal timer in the MCU
[06:15:11] <OndraSter> that will fire up every X msec
[06:15:22] <HKCMD> k
[06:15:27] <OndraSter> and check if the button is pressed (or set up the button press as interrupt and it will be enabling/disabling the counter)
[06:15:38] <HKCMD> k
[06:18:58] <theBear> you should have a timer/counter doing the switch-checking anyway, just make a note if it hasn't been UN-pressed, maybe incorporate with the debouncing.... pressed for X cycles (maybe 1/10th second) = normal press/debouncing, pressed for > 3/10ths of a second, start repeating every 1/10th
[06:19:55] <OndraSter> theBear, if he sets it up as interrupt, on rising (falling) edge he can start the timer, on the other edge he can disable the timer
[06:20:06] <theBear> OndraSter, sounds like a solid plan :)
[06:20:59] <OndraSter> :P
[06:27:50] <HKCMD> what can be the problem with the below code ?
[06:28:06] <HKCMD> http://pastebin.com/Rq5vBKwS
[06:28:56] <HKCMD> sorry , but this is a code for pic
[06:29:08] <HKCMD> in pic channel all are sleeping
[06:31:20] <OndraSter> whats the prob?
[06:31:37] <HKCMD> is it correct code ?
[06:34:08] <OndraSter> well does it work?
[06:34:50] <HKCMD> no some syntax problems
[06:35:04] <OndraSter> at which line?
[06:35:56] <HKCMD> 39 1508 Implicit conversion of int to ptr Counvalinc.c
[06:36:33] <OndraSter> actually
[06:36:42] <OndraSter> for (...) {
[06:36:44] <OndraSter> i++
[06:36:51] <OndraSter> char *output;
[06:36:59] <OndraSter> IntToStr(i, output);
[06:37:00] <OndraSter> I suppose
[06:37:06] <HKCMD> void IntToStr(int input, char *output);
[06:37:08] <OndraSter> I am C n00b
[06:37:12] <OndraSter> I know
[06:37:14] <HKCMD> k
[06:37:15] <OndraSter> try that
[06:37:31] <OndraSter> not sure if you have to alloc the size in memory before running IntToStr
[06:37:38] <OndraSter> or if it does internally
[06:37:40] <OndraSter> check docs
[06:39:14] <HKCMD> k
[06:39:58] <grummund> since you pass a pointer to the storage then the function will assume it is already allocated
[06:42:15] <OndraSter> check manual for the lib, HKCMD
[06:42:34] <HKCMD> s
[06:44:06] <HKCMD> see u later
[06:44:09] <OndraSter> bb
[06:44:12] <HKCMD> here power failure
[07:45:51] <Sgt_Lemming> home now, freaking long day
[07:46:22] <Sgt_Lemming> We went and flew the UAV for about an hour total, plus a few other RC planes.
[08:13:34] <dekroning> hi
[08:14:13] <dekroning> I haven't used AVR for along time, I still have a STK500. I just downloaded AVRStudio and only found example projects for STK600. Now I found a lot of *.bsd files for STK500, how can I use those in AVRStudio ?
[08:14:46] <dekroning> AVR BSDL files
[08:15:50] <ziph> They're descriptions for JTAG testing, I can't imagine you'd find much use for them on an STK500.
[08:16:19] <dekroning> ahok
[08:16:42] <dekroning> so no examples for STK500 , so i'll need to buy new STK600 to use new hardware ?
[08:17:45] <ziph> Why would you need an example project?
[08:18:29] <dekroning> because it's long time ago since I have used AVRStudio, would like to see the basics working, e.g. blinking leds, getting the usage of input buttons etc
[08:18:51] <ziph> If you want to buy something get an AVR Dragon.
[08:19:05] <ziph> Although you'll also then need some hardware for the AVR to do things with.
[08:19:10] * amee2k <3 avrisp2
[08:27:06] <dekroning> in the Tools menu, I should see my STK500 right ?
[08:27:28] <dekroning> I only see "Add STK500" and I just did that, however I don't see any option to "program" my device via the tools menu
[08:49:16] <OndraSter> http://clip2net.com/s/1vGI6
[08:49:23] <OndraSter> took me about 2 hours to make
[08:50:21] <OndraSter> DRC doesn't complain
[08:51:20] <karlp> what's plugging into those 2x4 blocks?
[08:51:37] <karlp> is this a revised board to connect to all your matrix led modules?
[08:52:07] <OndraSter> this is just the "serial in, parallel current driven out"
[08:52:12] <OndraSter> board
[08:52:14] <OndraSter> I split it into more boards
[08:52:36] <OndraSter> and it is cheaper (and even safer, if I fuck up) to make 5x10+10x10 instead 10x15 :)
[08:52:45] <OndraSter> by like $4 it is cheaper
[08:52:51] <OndraSter> but if I fuck up
[08:52:52] <OndraSter> I can change it
[08:57:15] <dekroning> OndraSter: which pcb tool did you use ?
[08:57:20] <OndraSter> Eagle
[08:57:25] <OndraSter> all manually routed
[08:57:35] <dekroning> must have been lots of work then :)
[08:57:38] <OndraSter> yap
[08:57:43] <OndraSter> like I said, about 2 hours
[08:57:57] <OndraSter> actually
[08:57:58] <OndraSter> 1 hour
[08:58:03] <dekroning> your gonna build that pcb your self?or have it build by some manufacturer ?
[08:58:08] <OndraSter> iteadstudio
[08:58:27] <OndraSter> doing both sided + vias at home = not a good idea :P
[08:58:28] <OndraSter> for me
[08:58:31] <OndraSter> with my skills
[08:59:25] <Bushman> ave
[08:59:36] <dekroning> is it expansive to create this board ?
[08:59:41] <dekroning> at iteadstudio
[08:59:43] <OndraSter> dekroning, check iteadstudio
[08:59:52] <OndraSter> I think it is about $22 or $25 for 10 pieces
[08:59:55] <OndraSter> I need two though
[08:59:59] <dekroning> OndraSter: and then you just do the soldering your self?
[08:59:59] <OndraSter> but still cheaper than local stuff
[09:00:01] <OndraSter> ye
[09:00:03] <OndraSter> will be pain
[09:00:15] <dekroning> because looks like SMT devices on there right ?
[09:00:18] <OndraSter> ye
[09:00:25] <OndraSter> parts list is long as dump after you've been 10 days in nature without toilet paper
[09:00:39] <dekroning> $25 per board, if you can order 2 doesn't sound to bad actually
[09:00:45] <OndraSter> $25 per 10 boards
[09:00:59] <OndraSter> I hope I will find use for the remaining 8 though :P
[09:01:08] <OndraSter> it is cheap as hell in the end
[09:01:17] <OndraSter> but you get too many boards at once for prototyping
[09:01:23] <OndraSter> but they are asking $25 for ONE board here
[09:01:26] <OndraSter> 1dm2
[09:02:04] <OndraSter> plus you pay etest separate I think
[09:02:10] <OndraSter> just as 2nd layer HAL and such
[09:03:47] <dekroning> OndraSter: for 10 boards, that's even cheaper :)
[09:03:51] <OndraSter> yap
[09:03:56] <OndraSter> $2.5 per board
[09:03:59] <OndraSter> cheaper than doing it at home
[09:04:03] <OndraSter> one layered
[09:04:11] <dekroning> for the soldering, what kind of station are you using ?
[09:04:22] <dekroning> also got these euhm forgot what the call it, stencils ?
[09:04:36] <OndraSter> nope
[09:04:41] <OndraSter> I will probably buy flux with Sn in it
[09:04:52] <OndraSter> put it on the pads manually (I know, it will be work)
[09:04:59] <OndraSter> put the chips there
[09:05:00] <OndraSter> and use hotair
[09:15:19] <dekroning> and how will you produce hot air, special tool ?
[09:15:32] <OndraSter> yes
[09:15:35] <OndraSter> I got (de)soldering station
[09:15:44] <OndraSter> wish I had infrared... but that is out of normal prices lol
[09:16:23] <OndraSter> I've got this
[09:16:23] <OndraSter> http://www.eres.alpha.pl/elektronika/fusion_images/articles/pt803s1.jpg
[09:16:51] <dekroning> well I'm also thinking about what stuff I need to minimally buy to produce somethings you are doing now, so it's interesting for me :)
[09:17:05] <OndraSter> I often buy stuff I don't ever use
[09:17:11] <OndraSter> just as I buy games that I will never play LOL
[09:17:17] <OndraSter> Steam sales are baaaaad, mmmkay
[09:20:30] <amee2k> steam is pretty fail imo
[09:20:51] <amee2k> i wouldn't buy games anymore even if i was still into gaming
[09:38:38] <inflex> hrmm... if I use a reasonable sized inductor, do you think it'd be possible to 'turn on' an AVR via an ext-interrupt using the BEMF from a magnet passing over it?
[09:39:04] <inflex> basically I want to make a magnetically operated on/off switch without relying on a hall switch or reed
[09:39:17] <amee2k> whats wrong with a hall sensor?
[09:39:25] <inflex> quiescent is way too high
[09:39:37] <amee2k> whats wrong with reed then?
[09:39:48] <inflex> expensive as hell for whatI want to do
[09:39:56] <amee2k> hmm i see
[09:39:59] <inflex> I was thinking of a SMD inductor (non-shielded)
[09:40:14] <inflex> basically put the inductor on one of the GPIO's, set up PCINT to wake up from sleep
[09:40:23] <amee2k> hmm
[09:40:34] <amee2k> not sure if thats enough tbh
[09:41:20] <amee2k> never tried anything like that but i'd say you need a strong magnet and a high value inductor for that to work
[09:41:56] <amee2k> unless you can rig some mechanism that quickly pulls the magnet acros the inductor, sort of like a magneto ignition
[09:44:06] <Sgt_Lemming> https://picasaweb.google.com/115277388924201849210/FlyingPhotos220112?authkey=Gv1sRgCNbovNLjm4PMKg <--- UAV/Rc Plane pics from today for anyone who is interested
[09:44:22] <amee2k> inflex: i'd make sure to have extra clamping diodes on the pin you're using for it, just in case something generates a strong spike on it
[09:44:28] <inflex> amee2k: it'll be a human 'striking' the magnet past the inductor
[09:44:51] <inflex> I guess I'll do some experiments tomorrow and find out how plausible it is.
[09:45:02] <amee2k> also, analog comparator comes to mind. iirc that can create an interrupt and you can arbitrarily set the threshold
[09:46:02] <inflex> I think for controlling the spike, I'd consider a cap - but then you have ringing issues... still, something like a 1nF cap + 1uH inductor shouldn't impose a massive load
[09:46:17] <amee2k> you can probably get a larger voltage swing that is enough for a normal interrupt if you put a transistor on it
[09:47:14] <inflex> not sure what you're meaning?
[09:47:41] <amee2k> the avr input is high impedance and voltage sensitive. i'd say any cap that would safely absorb a higher energy event is too much loading to get enough swing to trigger the pin
[09:48:20] <inflex> mm...maybe just a 1M resistor in parallel with the inductor
[09:49:35] <inflex> Anyhow, I'll experiment tomorrow
[09:50:51] <amee2k> inflex: http://ompldr.org/vY2U0Ng/inductor.png << this is what i'm thinking
[09:51:10] <OndraSter> hmm do I need 100nF at EACH chip on board that is powered only from USB?
[09:51:19] <amee2k> maybe a *very* small cap across the inductor if you find out you're getting noise issues
[09:51:45] <amee2k> OndraSter: for supply bypassing?
[09:51:48] <OndraSter> yes
[09:51:59] <OndraSter> there are not going to be any spikes
[09:52:01] <OndraSter> or shouldn't be
[09:52:06] <OndraSter> this is powered directly from USB
[09:52:07] <amee2k> there is not much difference between USB power and "normal" power
[09:52:21] <karlp> apart from the max cap load you can appear to the host as
[09:52:23] <amee2k> but never assume power to be clean
[09:52:25] <OndraSter> all the power spikes are going to be on different board
[09:52:35] <karlp> you need to limit inrush current
[09:53:05] <amee2k> i'd put a small to medium electrolytic next to the usb connector, then do the "usual" bypassing
[09:53:41] <amee2k> 100n is a fairly large bypass cap for low-powered parts. chances are you can get away with a lower value in many places
[09:54:14] <OndraSter> that is not my point.. myp oint is if I can get rid of some of them :P
[09:54:21] <OndraSter> board is 8x10cm max size
[09:54:27] <OndraSter> I will fit it
[09:54:41] <amee2k> depends entirely on how much noise the parts generate and how noise sensitive they are
[09:55:18] <OndraSter> atmega, 74573, 128kB SRAM, few 74154, and ft232rl
[09:55:22] <OndraSter> that's all there is on the board
[09:55:25] <amee2k> bypassing works at extremely small energy levels so parasitics are more important that raw size
[09:56:03] <amee2k> the 74xxx are HC* series?
[09:56:25] <OndraSter> ye
[09:56:37] <OndraSter> plus one 74hc595
[09:56:41] <amee2k> whats the 74573 do?
[09:56:45] <OndraSter> if I find AS (AC?) version, then that one
[09:56:48] <OndraSter> 8bit buffer
[09:56:53] <OndraSter> for xmem
[09:57:17] <amee2k> my gut feeling is the avr and the ram chips will want bypassing
[09:57:24] <OndraSter> yes
[09:57:28] <OndraSter> as they are the main core
[09:57:33] <OndraSter> and maybe the 573 to be sure
[09:57:35] <amee2k> maybe the buffer as well. a small one can't hurt either way
[09:57:39] <OndraSter> the others, if fail for a us, doesn't happen anything
[09:58:01] <amee2k> the 595 is used as a port expander?
[09:58:23] <OndraSter> no, it actually outputs row and is sent to MOSFET to another board
[09:58:28] <OndraSter> through the 154s
[09:58:53] <OndraSter> let me make a pic
[09:59:16] <OndraSter> http://dl.dropbox.com/u/3936936/logicboard.png
[09:59:20] <amee2k> extremely low speed stuff will probably work just fine. if you can position them close to each other they can probably share a single bypass cap
[09:59:53] <amee2k> i always try to run the traces from the supply bus right through the cap pads and on to the actual parts
[10:00:31] <amee2k> coming up which a good bypassing scheme purely based on a schematic is kinda hard sometimes
[10:01:10] <OndraSter> I have just one limit: 8x10cm board :P
[10:02:01] <amee2k> i'd say the 154s and the 595 can share one or two 100n caps, if you can accomodate that in your routing
[10:02:19] <amee2k> if not, values as low as 10n should be fine
[10:02:21] <OndraSter> well, I can fit them all in the end probably
[10:02:35] <OndraSter> the board is plenty big enough
[10:02:44] <OndraSter> it's just I am lazy
[10:02:53] <ziph> What package are you using that you can't just drop 100nF's in like candy?
[10:03:23] <OndraSter> normal SMT
[10:03:29] <OndraSter> and 0805 or 1206 for caps
[10:03:33] <amee2k> i'd point at SOIC from the schematic
[10:03:35] <OndraSter> I don't enjoy soldering <0805
[10:03:38] <OndraSter> yes
[10:03:48] <OndraSter> it will fit
[10:03:51] <OndraSter> I don't have any doubts
[10:03:52] <karlp> laziness now, vs odd relibility flakiness later....
[10:03:56] <OndraSter> :P
[10:04:00] <OndraSter> yap
[10:04:13] <amee2k> yep
[10:04:36] <amee2k> eliminating bypassing to save money is very ... chinese >_<
[10:04:49] <OndraSter> haha
[10:04:55] <OndraSter> the boards will be made in china
[10:04:56] <OndraSter> so... :P
[10:05:04] <amee2k> :P
[10:05:06] <ziph> The AVR probably doesn't need a bypass for every VCC.
[10:05:23] <karlp> they say to do so, but you normally get away without it.
[10:05:30] <OndraSter> "get away"
[10:05:31] <OndraSter> sounds chinese lol
[10:05:34] <ziph> They do?
[10:05:45] <ziph> Isn't one of them AVCC?
[10:05:46] <OndraSter> I need to do a bit of shooting
[10:05:49] <amee2k> if you don't need the adc, i don't think you'll need to bypass avcc. just connect it so it isn't left floating
[10:05:53] <OndraSter> they have twice Vcc
[10:06:00] <OndraSter> on bigger packages
[10:06:08] <OndraSter> amee2k, ADC is for possible future
[10:06:16] <OndraSter> I am bringing out all unused pins and ports outside
[10:06:25] <amee2k> i see
[10:06:42] <ziph> VCC pins 52 and 21 should probably have separate bypasses.
[10:06:44] <amee2k> then you should keep it. 100n on the aref pin is definitely a big one
[10:06:47] <ziph> Is this a 4 layer board?
[10:07:36] <amee2k> i'd put a cap footprint on aref, but not populate it at first. large caps like that can cause problems with some voltage reference ICs
[10:08:16] <amee2k> TL431 for example has a way of starting to oscillate with lots of capacitive load on the output
[10:15:54] * RikusW just fixed a lcd screen with a matchstick ;)
[10:16:39] <RikusW> loose feedback connection on cfl pressed into place with it
[10:21:19] <OndraSter> I have here older LCD
[10:21:22] <OndraSter> that when you connect one data ribbon
[10:21:27] <OndraSter> it won't power on :P
[10:21:31] <OndraSter> it "loses power"
[10:27:58] <RikusW> mine is samsung 920nw
[10:39:13] <OndraSter> dang, I have moved some part in mm spacing
[10:39:14] <OndraSter> rather inch
[10:39:20] <OndraSter> and now I can't properly connect the pads
[10:39:21] <OndraSter> damnit!
[11:53:35] <amee2k> hmm
[11:53:57] <amee2k> i've been scratching my head all morning long if bit-banging DVI is practical
[11:54:15] <OndraSter> bit banging DVI?
[11:54:16] <OndraSter> wow
[11:54:18] <OndraSter> with what?
[11:54:54] <amee2k> minimum clock is only around 25MHz. i was thinking an external ram chip and counter, clocked by an oscillator can
[11:55:02] <OndraSter> hmm only 25MHz?
[11:55:07] <OndraSter> I was expecting like 100MHz :)
[11:55:22] <amee2k> maximum is 160 apparently
[11:55:47] <amee2k> not sure if that means the clock signal on the wire or the bit clock though
[11:56:21] <amee2k> it apparently only transmits a sync clock at 1/10th the bit clock on the wire
[11:56:57] <amee2k> i was thinking two ram chips as frame buffers and bank switching to sync updates
[11:57:27] <amee2k> 8b10b encoding and everything done in software. optionally with a CPL or FPGA
[11:58:22] <keenerd> "Each TMDS link carries binary data at ten times the pixel clock reference frequency" Good luck.
[11:59:03] <amee2k> yeah, but with the external ram chips, the pixel clock doesn't have to get anywhere near the MCU
[12:00:23] <amee2k> i've got some old SIMM and DIMM ram sticks here. would have to check the ICs on them to see if they're fast enough
[12:01:38] <amee2k> i can handle 25MHz i'd say. 250 would be kind of a problem though, so not sure if the minimum frequency is meant for the bit clock or for the pixel reference signal
[12:02:03] <keenerd> So the FPGA would alos handle the XOR logic stuff?
[12:02:50] <amee2k> my idea was to prep the 10 bit frames in RAM so the discrete stuff just has to clock out bits to the LVDS bus driver
[12:03:51] <amee2k> i'm not entirely sure how this "invert for dc balance thing" is intended, but maybe i can implement it in hardware with just another RAM chip as lookup table
[12:04:48] <amee2k> "In the second stage, the first eight bits are optionally inverted to even out the balance of ones and zeros and therefore the sustained average DC level; the tenth bit encodes whether this inversion took place."
[12:05:30] <amee2k> ^^ if that decision doesn't require knowledge of the previous frames, a look-up table would work
[12:05:35] <keenerd> So you have no net DC current along the wire. But the key word here is optional :-)
[12:06:06] <amee2k> oh, the entire second step is optional?
[12:06:39] <amee2k> i read that as the inversion is optional for each frame, and it is mandatory to decide which frames to invert in order to maintain DC balance
[12:07:43] <keenerd> Possibly. Doubt it will hurt anything unless you are going to a 15m cable run.
[12:09:48] <amee2k> from what i can see the line driver uses some unusual differential current loop scheme
[12:10:58] <amee2k> this is just a weird idea at the moment, but i found it quite interresting that it doesn't look much more difficult than generating composite video with like 80s-era processors and a pile of TTL stuff
[12:11:20] <keenerd> I'd guess a lot of that fancy stuff is to minimize noise at high bandwidths. See how much of it you can ignore :-)
[12:11:36] <amee2k> hehe yeah
[12:12:33] <amee2k> resolution is going to be kinda shitty, and it'll need a pile of fine pitch SMT parts which will be a massive pain in the ass to prototype with
[12:13:22] <amee2k> mmh, there was some dude who used a VGA card to bit-bang DVB-T from a computer lol
[12:34:41] <amee2k> what the fuck
[12:35:09] <amee2k> "pre-assembled and calibrated kit for solder-less assembly"
[12:35:16] * amee2k facepizzas repeatedly
[12:36:58] <amee2k> they didn't seriously have it working, then just take out the board and sell it packaged separately so they can call it kit
[12:53:49] <ali1234> with avr-gcc and avr-as, how do i access variables declared as PROGMEM from assembler code?
[12:54:12] <ali1234> yes, with lpm, i know that
[12:54:21] <ali1234> but how do i calculate the address to put in Z?
[12:55:41] <OndraSter> LOW(SOMEADDR) into ZL
[12:55:46] <OndraSter> HIGH(SOMEADDR) into ZH
[12:55:47] <OndraSter> .-
[12:55:54] <OndraSter> you usually put some kind of name before the text
[12:55:57] <OndraSter> like
[12:55:58] <ali1234> well hey, i tried that, and it didn't work
[12:56:33] <OndraSter> SOMETHING: .db "something"
[12:56:35] <OndraSter> and then
[12:56:47] <OndraSter> LDI ZL, LOW(SOMETHING)
[12:56:51] <OndraSter> LDI ZH, HIGH(SOMETHING)
[12:57:50] <ali1234> that's how you access variables in data memory
[12:58:05] <ali1234> it doesn't work for variables in flash
[12:58:12] <OndraSter> oyesitdoes
[12:58:25] <timemage> ali1234, create something like: const int g = 7; int main() { int temp = g; return temp; } compile with -S or --save-temps and look at the resulting main.s.
[12:58:28] <OndraSter> LDI ZL ZH etc
[12:58:35] <OndraSter> and then LPM
[12:58:41] <OndraSter> and you get the data in R0 I think
[12:58:43] <OndraSter> let me check code
[12:58:51] <ali1234> here is my code http://pastebin.com/R3r5c9qp
[12:58:58] <OndraSter> ops
[12:58:59] <ali1234> well, cut down anyway
[12:59:06] <OndraSter> LPM Rx, Z
[12:59:11] <OndraSter> and Rx contains the byte
[12:59:49] <OndraSter> oh
[12:59:53] <OndraSter> I thought you are doing raw ASM
[13:00:00] <ali1234> timemage: const doesn't put the variable into flash
[13:00:16] <ali1234> OndraSter: when did i say that?
[13:00:23] <timemage> ali1234, odd.
[13:00:25] <OndraSter> well you said "from assembler code" in the end
[13:00:26] <OndraSter> xD
[13:00:37] <ali1234> OndraSter: .S file is assembler cone
[13:00:41] <ali1234> *code
[13:00:59] <OndraSter> ye
[13:01:22] <timemage> ali1234, so it does what just sticks it in ram?
[13:01:36] <ali1234> timemage: yes
[13:01:43] <timemage> ali1234, any idea why they did that?
[13:01:43] <ali1234> see http://www.nongnu.org/avr-libc/user-manual/pgmspace.html
[13:02:10] <ali1234> gotta go afk for a bit, i'll be back though
[13:03:00] <timemage> ali1234, the "A Note On const" section is numb.
[13:08:08] <amee2k> keenerd: okay, i think i figured it out. the transmitted sync clock is the specified frequency, and the baud rate is 10x that much
[13:09:42] <amee2k> so minimum clock frequency is 25 7/40 MHz, and the minimum raw transfer rate thus is 251.75Mbaud
[13:11:02] <keenerd> Hrm. Shouldn't the baud rate equal the clock rate, but there are 256 symbols?
[13:11:45] <amee2k> not quite. for emi reduction apparently the clock signal on the wire is 1/10th the baud rate
[13:12:09] <amee2k> and the bit clock has to be recovered by the receiver with a PLL
[13:12:34] <keenerd> I think you are confused as to what baud means...
[13:12:47] <amee2k> my definition is "bit time"
[13:13:21] <amee2k> one bit on the wire, which doesn't neccessarily have to translate into a payload bit but may be part of physical layer framing
[13:17:35] <amee2k> at the for 640x480 @ 30Hz i would need a link clock of 27.6MHz, memory clock of 83MHz and raw bit clock on the wire of 276.5MHz
[13:17:47] <amee2k> 133MHz sram might just cut it
[13:20:41] <amee2k> keenerd: or what am i misunderstanding here?
[13:31:24] <_abc_> amee2k: stripe the sram and the sram will operate at much lower frequency.
[13:33:34] <amee2k> _abc_: including control bits i'd need at least 9 data bits for one frame anyway, even if i do 8b10b encoding externally
[13:33:51] <_abc_> so?
[13:34:04] <_abc_> stripe the ram on A0 and halve the data rate
[13:34:11] <amee2k> so i'd need two chips anyway which would give me 1Mx16 worth of memory
[13:35:14] <amee2k> you weren't around earlier, eh?
[13:36:31] <amee2k> this morning i idly stumbled over the wikipedia article for DVI and noticed that generating DVI-D / HDMI compatible video signals doesn't look so bad with some external ram and a bunch of discrete logic
[13:37:46] <amee2k> at least not much worse than VGA or composite video circuits from the 80s that i've seen
[13:38:32] <amee2k> some SRAM chips and a small fast FPGA might just be enough
[13:57:09] <_abc_> yes
[13:57:31] <_abc_> There's a guy who generated DVB signals using his video card programmed with custom registers
[14:00:02] <amee2k> yeah, i've seen that one i think
[14:01:31] <amee2k> the idea that DVI/HDMI isn't really that expensive occoured to me when i read the minimum clock rate on DVI is just over 25MHz, resulting in a raw symbol rate on the wire of "only" 251MHz
[14:02:06] <amee2k> i think i know what new toy i want this summer >_>
[14:02:15] <amee2k> FPGA dev board
[14:10:13] <keenerd> Once I get code toe load on the darn GA144, I will have to try bit banging DVI.
[14:10:19] <keenerd> *to load
[14:11:05] <keenerd> 3 clocks per bit and no real synchronous clock might make that near impossible though.
[14:13:38] <amee2k> what makes it somewhat easier imo is the fact that you can do most of the 8b10b encoding with two lookup tables i think, one 8bit->8bit and one 8bit->1bit
[14:14:44] <amee2k> clocking out the raw data should be manageable, what i'm worried about is the video level framing, i.e. h/v sync stuff
[14:14:52] <amee2k> i haven't read much on that yet
[14:31:21] <karlp> I thought you didn't need the sync stuff anymore, as it could just count the bits.
[14:46:36] <Kre10s> hello could someone quickly tell me what device the device signiture 0x1e9584 corresponds to?
[14:48:11] <raek> maybe you could grep through the avrdude definition file
[14:48:30] <raek> I don't have it available myself currently
[14:50:59] <Kre10s> hmm. its supposed to be an atmega32m1... I have one sitting on my desk and it responds with 0x2480a5
[14:53:29] <raek> I might as well search through the avrdude db while my code's compiling
[14:54:13] <raek> does not seem to be there
[14:54:36] <Kre10s> its working now. must have been some fuzz. its a fresh chip
[14:55:08] <raek> found it, though: http://lists.gnu.org/archive/html/avrdude-dev/2011-12/txtV7hzp5DS1J.txt
[14:55:13] <raek> ATMEGA32M1
[16:03:19] <OndraSter> how fast does iteadstudio produce PCBs?
[16:03:23] <OndraSter> and are they good quality? :P
[16:14:31] <keenerd> Time can be two or three weeks. Quality is pretty nice.
[16:34:04] <Roklobsta> man, if it's 2 layers you can make your own in a hour I reckon.
[16:38:36] <Casper> the problem is to actually DO make them
[16:41:54] <Roklobsta> i used to make double layer boards 15+ years ago as an undergrad using laser printer, exposing positive film then exposing the PCB, fixing the resist and then etching. Took too long. I want to make some boards soon so I will try the laser toner transfer to the PCB and etchant on a sponge trick.
[16:42:04] <keenerd> Oh, and their thinnest board is a little too thin for hot air work. Was too slow with one and it started to do all droopy on me (-:
[16:42:42] <Roklobsta> i am waiting for Aldi to sell a cheap hobby drill too.
[16:44:50] <Casper> my printers are of too good quality...
[16:44:51] <amee2k> my toner doesn't want to transfer :/
[16:44:53] * amee2k sobs
[16:44:55] <Casper> they don't pour enought toner
[16:45:04] <Casper> amee2k: hp?
[16:45:08] <amee2k> yeah
[16:45:14] <amee2k> laserjet 2200dtn
[16:45:49] <Casper> same here, they used high temperature toner
[16:45:55] <amee2k> i love the printer, but i couldn't get even a half decent transfer after a few dozen attempts
[16:46:13] <Casper> same..
[16:46:26] <amee2k> i tried all heat settings until the copper plating popped off the board substrate >_<
[16:47:18] <amee2k> after a week i got a single board etched for 0.1" and i can do that in the same time with a sharpie and rub-off symbols
[16:48:14] <amee2k> thats why i want to get into ordering boards too now
[16:48:35] <Casper> I have a color laserjet 2605dn
[16:48:43] <Casper> and a 1012
[16:49:03] <Casper> both are high resolution (read thin toner coat) and high temperature toner
[16:49:04] <amee2k> i used to have a lj4 but one day it jammed and i couldn't get it cleared
[16:49:33] <Casper> the brother at work is also high resolution, so also have the same issue
[16:49:36] <amee2k> then i had some shitty okipage that left grey mush all over the page
[16:50:01] <Casper> I don't know what to do anymore to make boards
[16:50:05] <amee2k> until i traded the oki and lj4 and 50 eurobucks for the refurbished 2200dtn and lived happily ever since
[16:50:23] <Casper> the worse part is that the first board I did... I somehow succeded to do 8/8 mils
[16:50:31] <Casper> now I can't even get 50/50
[16:50:35] <amee2k> 0.0
[16:50:56] <amee2k> i don't want to have to get a new printer just for making boards
[16:51:11] <amee2k> but ordering is so expensive that it isn't much fun either
[16:51:20] <Casper> I for a while considered to get an used epson printer
[16:51:44] <amee2k> before i get another laser for transfer, i'm thinking about the inkjet mod
[16:51:48] <Casper> and use some mispro ink
[16:51:59] <amee2k> but haven't found a suitable ink peeing machine in the trash yet
[16:52:02] <Casper> yup that's what I wanted...
[16:52:07] <Casper> lol
[16:52:09] <amee2k> hehe
[16:52:13] <Casper> ink peeing machine lol
[16:52:18] <amee2k> what. :P
[16:53:18] <Casper> I need to order some toner
[16:53:38] <Casper> I wonder if I can order some toner bottle...
[16:53:59] <amee2k> a few weeks after i got my first laser printer i was like "wow, almost a month and the cartridge still works, no stripes on the printout and not a single jamming in like 100 pages 0.0"
[16:54:26] <OndraSter> haha yap
[16:54:43] <OndraSter> (considering you have dtn) plus duplexing, ethernet card in it, ..
[16:54:53] <Casper> amee2k: I'm on my 3th black cartridge, still on original color one
[16:54:55] <OndraSter> and I found inkjet printers to fail every time you need it the most
[16:54:56] <OndraSter> :P
[16:54:57] <amee2k> yeah, the ethernet card is really handy
[16:54:59] <keenerd> Yeah, lasers are almost as reliable as a dot matrix.
[16:55:17] <OndraSter> that's why I have black laser (hp laserjet 2300dn) and color multifunction inkjet
[16:55:21] <OndraSter> haha dot matrix
[16:56:00] <amee2k> the problem i have is that i don't print on a regular basis, so the ink cartridges were almost guaranteed to have dried out by the time i had something to print again
[16:56:24] <OndraSter> I print at least once a week on inkjet
[16:56:26] <OndraSter> or my sister does
[16:56:32] <keenerd> B/W laser + color printer at the drugstore is a perfect combo.
[16:56:37] <amee2k> i sometimes don't print for months
[16:56:48] <OndraSter> laserjet... <3 it, 6k pages, paid about 30€ for refilling and new roll
[16:57:16] <OndraSter> got it few months, still 90% toner left
[16:57:19] <amee2k> i think i paid less for all my laser time together than i used to on any single ink peeing machine i ever had
[16:57:25] <OndraSter> I got it actually for free, the printer itself LOL
[16:57:36] <OndraSter> some dude bought toner... but wrong one
[16:57:47] <OndraSter> and he didn't want to spend anything on it anymore
[16:57:51] <OndraSter> so he gave it away for free
[16:57:51] <amee2k> i'm still on the cartridge that came with the refurb'd lj2200 after almost two years
[16:57:54] <OndraSter> I was the 1st one lol
[16:58:03] <OndraSter> I even got the wrong toner
[16:58:05] <OndraSter> but nobody wants it :(
[16:58:07] <amee2k> the dude said it has one but can't tell how much toner is left
[16:58:08] <OndraSter> it is for hp lj2400
[16:58:45] <OndraSter> lol povray
[16:58:51] <OndraSter> http://clip2net.com/s/1vMIF
[16:58:53] <OndraSter> missing parts
[16:58:55] <amee2k> OndraSter: i got a pair of old lj4 cartridges for free during an internship in school
[16:59:01] <OndraSter> haha
[16:59:27] <amee2k> the printer they bought them for was long gone but noone bothered to get rid of the spares they had in the storage room
[16:59:50] <OndraSter> I have got here one of the first LED printers eve
[16:59:50] <OndraSter> n
[16:59:52] <OndraSter> 300 DPI
[16:59:53] <OndraSter> OKI
[16:59:55] <OndraSter> parallel port only
[17:00:07] <OndraSter> maybe there is some toner left in it even
[17:00:26] <OndraSter> and I even have oooold dot matrix
[17:00:28] <OndraSter> from Epson
[17:00:29] <amee2k> i never got around to use them myself and when i gave away the lj4, i sold them to a local computer store
[17:00:32] <OndraSter> LX800 or something liek that
[17:01:03] <amee2k> mmh, i had an oki once too
[17:01:20] <OndraSter> http://images04.olx.com.ph/ui/2/27/51/18322651_1.jpg
[17:01:20] <OndraSter> yap
[17:01:23] <OndraSter> this is the one I have
[17:01:27] <amee2k> when i got my laptop it came with a docking and a small oki laser
[17:01:29] <OndraSter> infinite paper ftw
[17:01:34] <OndraSter> oki LED?
[17:01:39] <amee2k> yeah
[17:01:54] <OndraSter> I took HP laser printer apart few days ago
[17:01:56] <OndraSter> supposedly not working
[17:02:00] <OndraSter> not worth getting toner for it
[17:02:03] <amee2k> at least it didn't look like any conventional laser system i've ever seen inside
[17:02:09] <OndraSter> I finally found out how does it work actually
[17:02:18] <OndraSter> yap, there is on the top/bottom row of LEDs
[17:02:25] <OndraSter> should be 300 of them for 300 DPI lol
[17:02:34] <amee2k> http://di1-3.shoppingshadow.com/images/pi/fd/99/b4/20157902-260x260-0-0_OKI+OKIPAGE+6ex.jpg
[17:02:42] <amee2k> fuck me if i remember the model, but it looked like that one
[17:02:43] <OndraSter> yap, exactly this one lol
[17:02:57] <OndraSter> 6e or 6ex
[17:02:59] <amee2k> but something was wrong with the toner or the drum or both
[17:03:18] <amee2k> it kept having some weird grey mush all over the page and vertical stripes
[17:03:38] <amee2k> got a bit better when i took it apart and cleaned everything, but it didn't go away
[17:03:40] <OndraSter> I remember when we had one of the first consumer copyprinters (just copy, no printing I think... but I might be wrong)
[17:03:44] <OndraSter> it was imported from USA
[17:03:47] <OndraSter> for really high prices
[17:04:01] <OndraSter> and we still have the 220V -> 110V transformer in the basement lol
[17:04:12] <OndraSter> I took the device apart like 15 years ago
[17:04:17] <OndraSter> it wasn't working anymore :(
[17:04:37] <OndraSter> and I somewhere lost the control panel.. it had like 200 LEDs on it... ofc with nowadays ebay prices it is worth nothing
[17:04:40] <amee2k> haha cool
[17:04:41] <OndraSter> but back then... lol
[17:04:59] <amee2k> yeah
[17:05:15] <amee2k> now i can buy a 2k leds on a belt for like 50 bucks on ebay
[17:05:19] <OndraSter> ye
[17:05:28] <OndraSter> I was looking for RGB LEDs cheaply
[17:05:39] <OndraSter> the ones with internal circuitry for flashing? price is almost 0
[17:05:47] <amee2k> lol
[17:05:49] <OndraSter> for normal RGB with common anode/cathode? prices are absolute trash
[17:07:39] <OndraSter> $20 for 100pcs
[17:07:39] <OndraSter> http://www.ebay.com/itm/100x-5mm-RGB-LED-Common-Cathode-Diffused-Lens-Manual-Control-4-Pin-Tri-Color-/150721551100?pt=LH_DefaultDomain_0&hash=item2317b45afc
[17:07:44] <OndraSter> 24 with shipping
[17:07:58] <amee2k> nice
[17:08:10] <OndraSter> not really
[17:08:19] <OndraSter> you can get 100pcs flashing RGB for one third
[17:08:34] <amee2k> well, i'm used to european prices
[17:08:47] <amee2k> and wtf do i want with flashing LEDs anyway
[17:08:56] <amee2k> i'm happy if they just stay on these days >-<
[17:08:58] <amee2k> >_<
[17:09:36] <OndraSter> haha
[17:09:40] <OndraSter> and this is trash too
[17:09:40] <OndraSter> http://www.ebay.com/itm/1pc-8x8-Matrix-RGB-LED-USA-Common-Anode-Diffused-Arduino-Full-Colour-Arduino-/150672913445?pt=LH_DefaultDomain_0&hash=item2314ce3425
[17:09:48] <Steffanx> Trash?
[17:09:53] <OndraSter> I can get 25 3mm RG ones for $35
[17:09:55] <OndraSter> Steffanx, price
[17:10:20] <OndraSter> prices of RG are like 10 times lower than RGB
[17:10:21] <OndraSter> but why!
[17:10:49] <izua> no B
[17:10:59] <izua> blue leds are inherently expensive
[17:11:11] <OndraSter> hmm
[17:11:12] <izua> less internal routing
[17:11:33] <OndraSter> and why are flashing RGB LEDs one third of the price of simple RGB LEDs with 4 pins?
[17:11:41] <OndraSter> is it cheaper to make few transistors than two extra wires?
[17:11:52] <amee2k> flashing looks shitty
[17:12:02] <OndraSter> plus you can't control it
[17:12:10] <amee2k> who would buy that except for sexually frustrated civic owners
[17:12:10] <OndraSter> haha
[17:12:12] <izua> much more products require plug and play flashing lights
[17:12:19] <izua> than intelligent control of flashing lights?
[17:12:43] <OndraSter> I can get 1000 blue LEDs for $26
[17:12:49] <OndraSter> now 1000 red and 1000 green
[17:12:52] <OndraSter> and I will build myself a wall
[17:13:00] <OndraSter> (and few FPGAs to ride them)
[17:13:19] <izua> you'll need either 100 or 10000
[17:13:36] <izua> or you know, 900.
[17:13:51] <amee2k> why not 961?
[17:14:07] <amee2k> or 1024
[17:14:12] <OndraSter> okay
[17:14:13] <Steffanx> or just one
[17:14:16] <OndraSter> 1024x768
[17:14:19] <OndraSter> :P
[17:14:19] <izua> 1024 is much nicer
[17:14:30] <izua> but you'd get a mere 32x32
[17:14:38] <izua> uVGA
[17:14:40] <OndraSter> that's not nice picture with 32x32
[17:14:41] <OndraSter> haha uVGA
[17:14:45] <amee2k> lol
[17:14:48] <OndraSter> LED wall
[17:14:50] <OndraSter> instead projector
[17:14:52] <izua> more like wail
[17:14:54] <OndraSter> who's in
[17:15:05] <Steffanx> As long as it's my wall :P
[17:15:10] <OndraSter> :P
[17:15:12] <izua> sounds like something i'd gouge my eyes out for
[17:15:13] <izua> literally
[17:15:15] <amee2k> make it 1056 LEDs
[17:15:24] <amee2k> 48x22 and call it WuVGA
[17:15:26] <OndraSter> why izua ?
[17:15:27] <Steffanx> I go for full hd
[17:15:31] <OndraSter> haha fullHD
[17:15:33] <OndraSter> I already counted that
[17:15:37] <OndraSter> how much LEDs would that be
[17:15:42] <OndraSter> I was at really high numbers
[17:15:44] <OndraSter> few milion that is
[17:15:44] <izua> OndraSter: try standing in front of those advertising panels powered by LEDs
[17:15:46] <OndraSter> per color
[17:15:51] <amee2k> and how many arduinos to control them
[17:15:56] <OndraSter> izua, that's why I wouldn't be shoving amps through each LED
[17:15:59] <OndraSter> haha arduino
[17:15:59] <izua> 2073600
[17:16:01] <izua> * 3
[17:16:06] <izua> yes yes
[17:16:07] <Steffanx> RGB
[17:16:09] <izua> arduino powered
[17:16:10] <izua> definitely
[17:16:17] <izua> cuz arduinos are cool like that
[17:16:22] <izua> and write the code in ubuntu too
[17:16:26] <Casper> . . . tonight on tv.... "vlog" then "lol :-)" ... what's next?
[17:16:38] <Casper> and yes, it's 2 real show...
[17:16:38] <Steffanx> Uh?
[17:16:39] <amee2k> lol
[17:16:43] <izua> "vlog" is a show?
[17:16:46] <Casper> yup
[17:16:53] <Steffanx> Welcome in the US?
[17:16:57] <amee2k> followed by vomit?
[17:16:58] <izua> we have this epic show
[17:16:58] <Casper> nope, canada
[17:17:04] <izua> where some guys come out and protest
[17:17:04] <Steffanx> Ah, same place
[17:17:08] <OndraLappy> :D
[17:17:14] <izua> and the ruling party officials hide
[17:17:15] * amee2k just realized why again he quit watching tv like 5 years ago
[17:17:18] <Casper> vlog is basically showing youtube video
[17:17:23] <OndraLappy> TV.. what is TV?
[17:17:24] <izua> been running for 10 days now
[17:17:33] <izua> i've been seeing it irl quite a few times, heh
[17:17:40] <amee2k> emergency measure to preserve my insanity
[17:17:41] <izua> sort of dejavu though
[17:18:15] <Casper> my bedroom tv don'T even have any antenna connected to it
[17:18:26] <OndraLappy> I have bedroom laptop :o)
[17:18:35] <izua> last time this show started, the people's leader ran away in a helicopter, later, the official bodyguards caught them, \they got a closed door trial, and they've been executed
[17:18:36] <amee2k> i have a bedroom bed
[17:18:45] <Casper> I have a WDTV live, which I should wire up, wireless isn't best
[17:18:59] <izua> i've seen several anonymous masks too
[17:19:03] <Steffanx> wtf izua ?
[17:19:26] <izua> Steffanx: there are seom epic protests going on, against the current party
[17:19:30] <izua> all major cities, etc
[17:19:39] <OndraLappy> LED WALL MODULE!
[17:19:39] <Steffanx> Yes but "and they've been executed" ?!
[17:19:39] <OndraLappy> http://www.necdisplay.com/p/l-d%20-walls-w/led-06af1
[17:19:44] <keenerd> Do you live in Poland?
[17:19:47] <OndraLappy> who's got 7.3 grand
[17:19:49] <izua> yeah, when the commies fell, in 89.
[17:19:55] <izua> it sort of happend just like this, lol
[17:19:58] <Steffanx> Aah, that's ages ago
[17:20:04] <OndraLappy> 80x80 dot
[17:20:05] <OndraLappy> lmao
[17:20:13] <amee2k> izua: they do that show in the US too, they just call it elections there
[17:20:14] <izua> epic protests everywhere, president & co hiding
[17:20:28] <izua> it's a neat show, isn't it? :D
[17:20:45] <Steffanx> It's called .. the news?
[17:20:46] <amee2k> when in rome do like the romans do
[17:20:58] <izua> well lol
[17:20:59] <amee2k> and stab your politicians
[17:21:13] <izua> it's been running for 10 days or so straight, and every time i got near a tv, i hear about it
[17:21:32] <amee2k> you should stop and fap
[17:21:35] <Steffanx> Arabic spring in Romania :P
[17:21:38] <izua> all i wanted was a batman movie or a cheesy animation
[17:21:39] <amee2k> err, by that i mean watch of course
[17:21:50] <amee2k> izua: how about you've got internet?
[17:22:01] <izua> yeah, only that the arabic spring led to nothing, except more censorship
[17:22:09] <izua> my thoughts exactly
[17:22:11] <Steffanx> Same will happen there?
[17:22:18] <amee2k> izua: was it an rgb one at least?
[17:22:23] <izua> if they don't stop, probably internet and telephony will be cut off
[17:22:24] * amee2k runs
[17:22:32] <izua> considering that this thing started as a flash mob
[17:22:33] <Steffanx> Yay
[17:22:44] <amee2k> sounds more like lynch mob
[17:22:52] <amee2k> not that there is anything wrong with that
[17:22:57] <izua> yeah, the banners were sort of like that
[17:23:13] * Steffanx <= ZZzzz
[17:23:22] <izua> "up with the president", near a makeshift gallows, etc
[17:23:37] <OndraLappy> http://www.youtube.com/watch?v=T88ej64aXUM
[17:23:40] <OndraLappy> china powah
[17:23:59] <izua> they even came with makeshift crosses with the ex-commie leader's name on it, and a coffin
[17:26:21] <timemage> OndraLappy, i was just looking at all the CN IPs in my fail2ban logs.
[17:26:32] <OndraLappy> am I there too?
[17:26:57] <timemage> OndraLappy, in my logs?
[17:27:00] <OndraLappy> ye
[17:27:04] <timemage> OndraLappy, i hope not.
[17:27:09] <amee2k> lol
[17:27:09] <OndraLappy> or is it response to the "china powah"
[17:27:20] <timemage> OndraLappy, more to that.
[17:27:40] <amee2k> http://www.youtube.com/watch?v=50HXHcl8rbE << afghan power > china power
[17:28:28] <timemage> OndraLappy, i was slightly entertain the idea that there should be legal to practice attacks on chinese IPs.
[17:29:58] <amee2k> OndraLappy: someone can't use a camera there
[17:31:01] <amee2k> also, as far as wall warts go that one isn't too bad
[17:31:12] <OndraLappy> no?
[17:31:19] <OndraLappy> 700mA output (rated supposedly upto 1A)
[17:31:23] <OndraLappy> and the voltage is at almost 0?
[17:31:31] <Roklobsta> hrm, i have a samsung ml1710 an apparently the toner transfers well. we'll see. i used to print onto semitransparent drawing paper in the old days. i want to try baking paper now.
[17:31:53] <amee2k> the cable is pretty shit but the rest is kinda standard
[17:32:00] <amee2k> and pretty good considering they didn't use an IC
[17:32:29] <amee2k> i've got a phone charger with a shitty one transistor circuit with primary side sensing
[17:32:53] <amee2k> unloaded the voltage goes up to like 8V (rated 5V)
[17:33:13] <Roklobsta> a friend reckons he designed the original small nokia charger. got $10000 for his efforts and had to be makable for under US$2.
[17:33:22] <amee2k> also, when you buy the power supply from china, always oversize by a factor of three if you want it to work
[17:33:49] <amee2k> or why do you think people put 500W PSUs into 150W computers. guess what, they're only 200W PSUs irl
[17:33:56] <OndraLappy> I know
[17:33:59] <karlp> I love the "very carefully remove the ferrite core. <BANG>"
[17:34:04] <OndraLappy> that's why my 525W PSU cost more than chinese 1.5kW :P
[17:34:08] <timemage> OndraLappy, i like the "carefully remove the ferrite core" bit.
[17:34:10] <OndraLappy> haha
[17:34:13] <karlp> hehe
[17:34:30] <timemage> creepy =)
[17:35:54] <amee2k> i once modded a cheap ATX PSU into an SLA battery charger
[17:36:19] <amee2k> the sticker said 250W. when i load-tested it, turns out duty cycle reaches 100% at just over 100W
[17:38:07] <keenerd> I think I got this link from someone in this channel: http://www.cloudfrom.com:8080/showItem/showDetail/10680033.html Good read if you can get past the bad translation.
[17:38:58] <amee2k> http://ompldr.org/vOTRzbg/wall-wart-board-small.jpg << this is the board from a wall wart i fixed. and pretty much the worst soldering job i've seen like ever
[17:40:06] <amee2k> http://ompldr.org/vOTR0OA/cimg7614b.jpg << same device. check out how the board is bent to fit because they shoved it in next to the guides in the case
[17:41:38] <OndraLappy> lol
[17:41:44] <OndraLappy> such a cool soldering job on the bottom
[17:43:03] <OndraLappy> I have Enermax Pro82+
[17:43:04] <OndraLappy> 525W
[17:43:06] <OndraLappy> it cost me a lot
[17:43:11] <OndraLappy> and now it costs even more lol
[17:43:21] <OndraLappy> I remember some people testing it at about 550W
[17:43:34] <OndraLappy> and still giving 12V +- 0.05V
[17:46:56] <amee2k> R5 has an identity crisis and pretends to be a cap, R4 can't decide which pad it likes better
[17:47:22] <amee2k> C5 is going for a swim in a huge blob of solder
[17:48:24] <OndraLappy> gn
[17:48:30] <amee2k> nighty :)
[22:36:38] <hexorg> hey all
[22:37:37] <Essobi> sup
[22:37:54] <hexorg> thoughts lol
[22:38:13] <hexorg> do you have a preferred way to get data from PC to uC?
[22:39:57] <LoRez> serial usually works.
[22:40:40] <hexorg> would RS232 or LTP have higher bandwidth?
[22:41:02] <LoRez> LTP?
[22:41:16] <hexorg> parallel printer port
[22:41:34] <LoRez> LPT
[22:41:46] <hexorg> oh >_< sorry, tired lol
[22:42:10] <LoRez> parallel ports require more IO.
[22:42:34] <hexorg> yea, I have plenty IO, I just worry about transfer speed a bit
[22:43:05] <LoRez> besides the 8 bits of IO, it takes at least some sort of clock pin.
[22:43:35] <LoRez> and that's only if it's unidirectional. another bit or two if you want to do bidirectional.
[22:44:08] <LoRez> serial, you can do 2M with a proper clock speed.
[22:45:36] <hexorg> hm... it'd be perfect if I could transfer at 9.6megabytes / second
[22:46:12] <LoRez> you're going to have a hard time doing that either way while doing anything else.
[22:46:14] <Kevin`> hexorg: you are not likely to be able to handle that much data in the cpu
[22:46:37] <hexorg> hm... or maybe I can implement some compression...
[22:46:54] <LoRez> even at 20MHz, every other clock would be transmitting a byte
[22:46:59] <Kevin`> hexorg: also, the parallel port probably can't clock data that fast. no technical reason, but it's not really designed to
[22:47:14] <hexorg> ok, from the beginning then... I have 100x200 pixels... I'd like to output video on it
[22:47:25] <hexorg> each pixel is 3 bits
[22:47:29] <hexorg> (R G B)
[22:48:20] <hexorg> so 7.5kb for the whole image
[22:48:22] <Kevin`> I wonder if you could get something like that working with dma
[22:49:44] <hexorg> well... I have segments with 5 pixels per segment, each segment has ATmega328 and a constant current driver
[22:50:11] <hexorg> I use those uCs as memory pulls
[22:50:13] <hexorg> *pools
[22:50:50] <hexorg> I can talk to those uCs by SPI
[22:51:01] <Kevin`> does atmel make any devices that have a microcontroller core (cpu and memory) surrounded by programmable io? who's the best for something like that, from an ease-of-use standpoint?
[22:51:02] <LoRez> hexorg: that's only 225kb/s
[22:51:46] <inflex> Kevin`: the xmega perhaps?
[22:51:51] <hexorg> ... *scratches head
[22:52:06] <hexorg> for 30 FPS... Yea... 225 kb/s
[22:52:15] <inflex> Kevin`: but not entirely sure how what you're asking differs from normal AVRs
[22:52:19] <LoRez> so where'd you come up with 9.6M/s?
[22:52:22] <Kevin`> inflex: xmega has dma, but I think all of the models are normal microcontrollers, no fgpa sections in them
[22:52:26] <hexorg> I think I multiplied by 20 one extra time
[22:52:39] <inflex> ooooh, FPGA section
[22:52:42] <Kevin`> inflex: erm, programmable LOGIC, not programmable io. mental typo :)
[22:52:55] <hexorg> nevermind then lol I can do 225kb/s easily with RS232
[22:54:22] <hexorg> Another question then, if my SPI master runs at 14.7Mhz and slave at 8Mhz, can I set master clock for anything above 4Mhz?
[22:54:57] <Kevin`> hexorg: sure you can, it just won't work
[22:55:16] <hexorg> lol thanks :p
[22:55:31] <hexorg> with anything below 4Mhz "work" ? :p
[22:57:05] <Kevin`> as far as I know you can run the slave interface as slow as you like, down to dc
[22:57:35] <hexorg> but whats the top limit? Half of slave's frequency?
[22:58:47] <Kevin`> that sounds right, but the datasheet says it
[22:58:58] <Kevin`> that or 1/4
[22:59:13] <hexorg> alright thanks :)