#avr | Logs for 2013-10-24

Back
[00:00:01] <weilawei> (its now at the end of the .text section)
[00:00:12] <juri_> weilawei: to put what you want as the first thing, use the vectors attribute?
[00:00:17] <weilawei> no
[00:00:20] <weilawei> it needs to be last
[00:00:39] <juri_> whyfor? normally, its between vectors, and program code.
[00:00:45] <weilawei> the start of .text is the reset vector
[00:00:50] <weilawei> i need to store data values, in flash
[00:00:57] <weilawei> so they're in a progmem section
[00:01:01] <weilawei> but this is bare metal
[00:01:03] <weilawei> if i put them first
[00:01:16] <juri_> yea, i ran into this last night.
[00:01:20] <weilawei> im overwriting my reset vector and other interrupt vectors
[00:01:23] <weilawei> so i moved it to the end
[00:02:07] <juri_> the solution for me (in C) was to add __attribute__ ((section (".vectors"))) to my hand written vector table.
[00:02:08] <weilawei> this whole thing works grand, minus the sevenseg_num procedure and sevenseg_numbers table
[00:02:21] <weilawei> problem there is..
[00:02:27] <weilawei> you're not doing bare metal
[00:02:38] <weilawei> if you objdump -H your elf file
[00:02:41] <juri_> you're right. i'm writing C... but thats about it.
[00:02:48] <weilawei> youll see that it has a .vector section
[00:02:56] <juri_> now, yes. before, no.
[00:03:05] <weilawei> in my case, i dont have anything to move stuff around
[00:03:12] <weilawei> my code is *it*
[00:03:15] <weilawei> the *only* thing on that atmega
[00:03:28] <weilawei> so i use objcopy to extract the .text section and write that to the atmega
[00:03:29] <juri_> um. same here.
[00:03:47] <weilawei> you should see how much extra stuff avr-gcc adds in
[00:03:57] <juri_> avr-objcopy -j .text -j .data -O ihex HeadControlm8535.elf HeadControlm8535.hex
[00:04:00] <juri_> um.
[00:04:06] <juri_> its not adding it into my .hex file.
[00:04:43] <juri_> you can write C, without leaning on anything. technically, i should probably remove that .data.
[00:04:46] <juri_> i'm not using it.
[00:05:13] <juri_> somehow, i'm compiling my C such that the code that would normally set up .data isn't getting integrated.
[00:05:28] <weilawei> not a clue
[00:05:30] <juri_> i'm just storing my strings in progmem as a result.
[00:05:44] <weilawei> but none of this explains why its calculating addresses wrong
[00:05:52] <juri_> my code is on gitorious. https://gitorious.org/reprap-head-controller
[00:06:30] <juri_> i haven't had problems with address calculation. my problem was the strings weren't getting there. i had to add them to the progmem section, and learn to get my vector table first in the .hex file.
[00:07:00] <juri_> now, after adding them to the progmem, i ran into address calculation problems.
[00:07:58] <juri_> this is a harvard archetecture, so i had to use PROGMEM and appropriate pgm_read_byte() macro to read them.
[00:08:03] <weilawei> theyre in progmem, as progmem gets coalesced into .text, as per the linker script supplied with avr-gcc. all i did was move progmem from the beginning of text to the end
[00:08:15] <weilawei> juri_, progmem is a macro for what i did
[00:08:23] <weilawei> and pgm_read_byte is also effectively one
[00:08:33] * juri_ nods.
[00:08:41] <weilawei> http://pastie.org/8425925#2,108-115,252-257
[00:08:44] <weilawei> see the highlighted sections
[00:08:54] <juri_> thats why i'm confused about why you didn't use .vectors, and leave the progmem at the beginning of the text segment.
[00:09:42] <weilawei> .vectors appears to tell it to put it VERY FIRST in the output file
[00:09:50] <weilawei> which is the location of the interrupt vectors
[00:10:01] <weilawei> which should not be my table of values
[00:10:28] <juri_> riiight. i meant leave your table alone, so it pops to the front of .text (after vectors) and setting vectors for your vectors.
[00:11:24] <juri_> weird. 1ec.. where is it getting that?
[00:11:36] <weilawei> thats a handcomputed value
[00:11:40] <weilawei> its getting it from me
[00:11:53] <weilawei> its the actual correct address, empirically determined by hexdumping and counting
[00:12:06] <weilawei> luckily, changing it doesnt change the size/location of everything else
[00:12:13] <weilawei> so i can assemble, and then change the source and rebuild
[00:12:55] <weilawei> (a huge pain in the butt, btw, and not at all workable)
[00:13:07] * juri_ nods.
[00:13:36] <weilawei> but pm_lo8(sevenseg_numbers) and the real address of sevenseg_numbers are not the same
[00:14:09] <weilawei> nor does pm_hi8 seem to be right
[00:14:27] <weilawei> it seems that nobody in their right mind uses avr-as directly from all the googling ive done
[00:15:13] <juri_> well, i'm not, but farbeit for me to claim i am in my right mind. ;)
[00:16:06] <weilawei> so any idea how to set vectors for those initial jmps?
[00:16:26] <weilawei> i had to compile a dummy program and decompile it just to get an example of how to set a progmem section in avr-as
[00:16:31] <weilawei> *disassemble
[00:16:54] <weilawei> er nvm, i kept the temp files including the .s
[00:16:56] <weilawei> but anywho
[00:17:31] <juri_> i created a function using asm volatile() to do pretty much the same thing you did.
[00:17:38] <juri_> i then threw it in the vectors section.
[00:17:46] <weilawei> i mean
[00:17:52] <weilawei> how do i declare the first 21 jmps as .vectors?
[00:18:01] <weilawei> gas is notably lacking on useful documentation
[00:18:14] <juri_> that'd be an assembly question. i can't answer those. :)
[00:18:25] <weilawei> >_<
[00:18:29] <weilawei> theres a way to do this in C...
[00:18:40] <weilawei> theres a way to do this using someone elses assembler
[00:18:44] <juri_> we might be similarly close to the metal... but i still write in C, and read the assembly in the .lss file.
[00:19:11] <weilawei> thats what i did to figure out how to get progmem.. do you have an example of what it spits out when you ask for a vectors section?
[00:19:22] <specing> weilawei: make an empty C program
[00:19:28] <specing> compile it, disassemble it
[00:19:34] <weilawei> specing, i did that basically before
[00:19:35] <weilawei> but
[00:19:37] <specing> there, that is how you make 21 jumps in asm
[00:19:40] <weilawei> PROGMEM is a macro
[00:19:46] <specing> ya
[00:19:48] <weilawei> it decompiles to a bunch of stuff
[00:19:52] <weilawei> whats the macro for .vectors
[00:20:07] <juri_> .vectors is not a macro. its an attribute.
[00:20:10] <weilawei> yes
[00:20:14] <weilawei> but how do you set it in C
[00:20:27] <specing> you set it in the linker script
[00:20:33] <weilawei> i dont work with this thing in C and i avoid compiler specific stuff like that plague..
[00:20:37] <weilawei> nah that sets the location of it
[00:20:37] <juri_> void start(void) __attribute__ ((section (".vectors")));
[00:20:41] <weilawei> juri_, thanks
[00:20:58] <juri_> void start(void) { /* vector table here, in assembly */ }
[00:28:02] <weilawei> well it moved them to between the vectors and the text
[00:28:20] <weilawei> problem is now, it doesnt even get to setting up the stack and flashing the LED signalling boot is done
[00:32:18] <weilawei> ah yes
[00:32:28] <weilawei> the reset vector is now pointing square at my first element of data
[00:32:39] <weilawei> it's definitely calculating addresses wrong when it moves things around
[00:33:03] <juri_> Interesting.
[00:33:09] <juri_> i'm not seeing this in C here.
[00:33:20] <weilawei> yeah tell me about it... id like to know what im doing different
[00:33:46] <juri_> the only thing i can tell you is i'm using rjmp, instead of jmp.
[00:34:09] <weilawei> i have this now http://pastie.org/8425965
[00:34:20] <weilawei> that shouldnt be an issue, im so low in the flasg
[00:34:22] <weilawei> *flash
[00:34:27] <weilawei> and everythings pretty close
[00:34:44] <weilawei> to blink an LED in C takes something ridiculous like 4 K
[00:34:58] <weilawei> in asm its around 140 bytes
[00:35:24] <juri_> i'm at less than 530 bytes, and i have serial set up, and an LED flashing. ;)
[00:35:44] <weilawei> i was just going by what gcc gave me when i did it in C originally...
[00:35:58] <weilawei> it leaves a bunch of extra stuff inside thats not strictly necessary
[00:36:12] <juri_> i daresay your implementation was not as efficient as what i'm running here.
[00:36:46] <weilawei> im going to guess thatd be right..
[00:36:54] <weilawei> but i reimplement it in assembler and it drops right down
[00:37:28] <juri_> nod. i'm loving these lss files, as i get to watch the compiler flail, then optimize the assembly by changing the C.
[00:45:46] <weilawei> its not writing the same code for the vectors at the start...
[00:45:54] <weilawei> i compared two programs which are identical at the start
[00:45:55] <weilawei> http://pastie.org/pastes/8425976/text
[00:46:26] <weilawei> so it should be a different offset in the file
[00:46:47] <weilawei> it moved it by 5 bytes?
[00:47:02] <weilawei> so why no boot..
[00:48:48] <weilawei> so my size was wrong... i changed it to 20 and it spits out the same code as if i dont have a .size directive..
[00:49:04] <weilawei> which doesnt make much sense..
[00:49:08] <weilawei> im only storing 10 bytes
[00:49:12] <juri_> whats a .size?
[00:49:38] <weilawei> specifies the size of a section, in bytes according to the manual
[00:49:45] <weilawei> so 10 shouldve been fine
[00:49:51] <weilawei> 5 words, 16 bit aligned
[00:50:24] <weilawei> when it autogenerates progmem code, it specifies a section for it as a subsection of .progmem
[00:50:28] <weilawei> and it gives it a size and such
[00:50:59] <weilawei> not specifying the size and specifying .size 20 gave me the same binary
[00:51:13] <weilawei> .size 10 or .size 5 gave me others
[00:59:26] <juri_> undefined reference to `__prologue_saves__'
[00:59:34] <juri_> any ideas? printf doesn't like me.
[01:00:04] <juri_> its wanted by libc.a, but including libscanf_flt or libprintf_flt doesn't fix it.
[01:08:39] <weilawei> that seems to be on function entry...
[01:08:44] <weilawei> did it compile and link before?
[01:08:55] <hjohnson> hah.. almost forgot the I2C pullups
[01:09:14] <weilawei> i also dont have a libc.a in here?
[01:09:24] <weilawei> must be somewhere.. ive got a libgcc.a
[01:10:22] <weilawei> the only place I found it as text was in lib/gcc/avr/4.7.2/plugin/include/insn-flags.h, but if you run strings against libgcc.a (or w/e), youll see it repeatedly
[01:11:50] <juri_> yea. no printf() for me, for now.
[01:16:03] <weilawei> juri_, http://lists.gnu.org/archive/html/avr-gcc-list/2008-07/msg00044.html try -mno-call-prologues
[01:16:24] <weilawei> http://www.avrfreaks.net/index.php?name=PNphpBB2&file=printview&t=30189&start=0 see also
[01:19:45] <weilawei> also, apparently avr-libc is compiled with -mcall-prologues, might be worth a recompile
[01:21:21] <weilawei> hjohnson, i dont think it would do much without them =P
[01:31:24] <spec__> Anyone familiar with the LM317 in constant current mode?
[01:34:18] <weilawei> http://theparanoidtroll.com/electronics/primers-and-tutorials/other/constant-current-sourceload-lm317/ detailed explanation
[01:40:18] <weilawei> i think im giving up on avr-as... taking my ball and going to avra
[01:40:45] <weilawei> gcc is a special layer of hell for computers that hate people...
[01:55:42] <hjohnson> weilawei: yeah
[01:56:02] <hjohnson> ok... just trying to see if there's anything else missing on my board layout
[01:56:08] <hjohnson> i'm pretty stoked to start building software for this thing
[01:58:34] <spec__> Interesting. The 10W LEDs I got off eBay only consume 3W at their rated voltage
[01:59:09] <spec__> They're still bright as hell. Just wondering if they'll go brighter.
[02:00:14] <weilawei> pop. i had one of those moments earlier.. "why is this so bright?" "oh yeah, im using a 100 ohm resistor, not a 330 ohm resistor.."
[02:05:09] <weilawei> juri_, so i switch to avra.. same code minus the GCC specific stuff
[02:05:14] <weilawei> works flawlessly
[02:05:36] <weilawei> im thinking theres some major bugs lurking in the avr port of gcc
[02:06:26] <weilawei> in fact they compile to a byte for byte exact match of each other until you try to store something in progmem
[02:06:37] <weilawei> s/compile/assemble and link/
[02:06:54] <juri_> interesting.
[02:07:12] <juri_> i've noticed some minor bugs... but its free software, so i consider it a personal responsibility to fix those.
[02:07:15] <weilawei> my little procedure now spits out the right number happily
[02:07:23] <weilawei> im not fixing gcc, not with a 1000 mile pole
[02:07:46] <juri_> pay me to fix it, then. ;D
[02:08:21] <weilawei> if someone wants to PAY me, otoh...
[02:08:42] <juri_> see? didn't that pole get much shorter? ;)
[02:08:47] <weilawei> but im just trying to learn avr assembler, not the intricacies of gcc.
[02:08:52] <weilawei> i like yasm anyway xD
[02:09:12] <juri_> i'm trying to get something done, with a deadline of "before you were born".
[02:09:25] <weilawei> isnt it always like that?
[02:10:06] <weilawei> on /. today, there was an article suggesting that time is an emergent phenomenon (if you do out the math, its said you get a static universe.. but my math skills arent there yet to try it myself)
[02:10:15] <weilawei> so, you can argue that its actually done =P
[02:10:19] <weilawei> you just cant perceive it
[02:14:56] <hjohnson> weilawei: http://i.imgur.com/byZoyQ4.png my board...
[02:15:51] <megal0maniac_afk> hjohnson: That's perdy
[02:15:55] <weilawei> altium?
[02:16:00] <weilawei> indeed, it is pretty
[02:16:36] <weilawei> whatd you use to layout the PCB?
[02:16:47] <spec__> I haven't hit any compiler issues, but the avr chip support needs serious clean up and fixes.
[02:17:13] <spec__> Like there's no PROGMEM support for the Tiny2313A
[02:17:34] <theBear> outrageous !
[02:17:54] <spec__> I *know*!
[02:17:59] <theBear> that's the newer sibling of my favourite general purpose avr, and the one i have many of for my next few projects
[02:18:23] <theBear> and if i EVER get this damned pc back together those next few projects will be closer than ever before !
[02:19:42] <theBear> heh, reminds me of the they might be giants song "your older than you've ever been and now you're even older, and now you're even older, you're older than you've ever been. and now you're older still ! TIME, IS MARCHING ON ... AND, TIME, IS STILL MARCHING ON" heh
[03:29:03] <juri_> does anyone have a recommendation for a usb hub chip, for integrating into an embedded device?
[04:45:10] <nge> Im having trouble setting up TWI with the help of the ASF in atmel studio.
[04:45:24] <nge> Using the write function makes it gets stuck forever waiting on an TWI_SR_TXRDY, or at least TWI_RECEIVE_NACK. What could be the problem here? Even the probe function gets stuck.
[04:45:56] <nge> I mean the master never seem to receive any signal at all, what could this point to?
[04:46:11] <nge> 1.) The device doesn't exist on the I2C bus ?
[04:46:33] <nge> 2.) Could I have done something wrong when starting the card, like missing to do something with a port?
[05:01:55] <megal0maniac_afk> juri_: GL850G
[05:15:55] <juri_> Thanks.
[06:05:19] <hugogee> !gn
[07:11:49] <setra> hello, like to use avr-gdb with avrice on xmega16a4 in codeblocks, any tuts or advices
[07:32:12] <setra> any help from this channel?
[07:40:48] <specing> Absolutely none
[07:40:57] <specing> atleast when Im around
[08:59:52] <theBear> hehe, we'll tell you as soon as specing disappears again :)
[09:00:30] <specing> that could take a whiiiiiiiiiiiiiiiile
[09:14:46] * twnqx squashes specing according to the channel rules
[09:16:07] * specing encourages twnqx's squashing
[09:21:29] * megal0maniac_afk gets the tree...
[09:25:21] * theBear encourages specing as per channel rules :)
[09:27:42] * specing explodes from all the courage
[09:30:32] <nge> is the i2c clock supposed to tick all the time even if there is for example no transfer at the moment?
[09:38:06] <megal0maniac_afk> nge: I don't think so
[09:42:04] <nge> megal0maniac_afk: Oh okay, but what if the master is in transfer mode and waiting for a ready bit from slave forever. would the clock be ticking then? I'm trying to see if my TWI clock pin is active at all and it's not
[09:42:42] <nge> I mean is there some mode I could put the master in to force the clock to be ticking?
[11:55:54] <beaky> hello
[11:56:04] <beaky> how do I get high frequency pwm
[11:56:09] <beaky> above 100 khz
[12:23:31] <N1njAway> beaky: Use external components
[12:25:29] <beaky> what kind?
[12:25:43] <beaky> are there "frequency amplifiers"?
[12:26:00] <beaky> turning a 100khz pwm to a 1Mhz one
[12:37:33] <vsync_> dude you can get 50/60 HZ FROM MAINS!
[12:37:40] <vsync_> just hook it up to xtal and u are good to go
[12:38:13] <Tom_itx> beaky, PLL
[12:38:21] <megal0maniac_afk> vsync_: Shit dude, are you trying to kill him?
[12:38:31] <megal0maniac_afk> beaky: You need to put a resistor in series, at least 10K
[12:38:31] <Tom_itx> i disregard alot vsync_ says
[12:38:33] <RikusW> vsync_: do you wan't to let the smoke out ? :-P
[12:43:19] <OndraSter> want or don't want, RikusW ? :D
[12:43:41] <RikusW> heh
[12:43:53] <OndraSter> wan't is an unknown word
[12:44:03] <RikusW> typo
[12:44:47] <RikusW> So I got an Altera USB Blaster clone working on USB ATmega :)
[12:44:54] <OndraSter> nice
[12:45:00] <RikusW> interested ?
[12:45:14] <RikusW> its still alpha..
[12:45:22] <RikusW> but seems to work now...
[12:45:39] <RikusW> I had a buffering issue that caused failure...
[12:46:47] <megal0maniac_afk> If I had an Altera chip...
[12:47:19] <RikusW> seems the ixo-jtag guy made a patch to let it work with Xilinx
[12:47:28] <RikusW> for openocd anyways
[12:51:19] <RikusW> OndraSter: btw WireShark works great for USB debugging
[12:51:58] <Tom_itx> really?
[12:52:03] <RikusW> yes
[12:52:15] <RikusW> in Linux just do modprobe usbmon first
[12:52:16] <Tom_itx> i've used it on my network before...
[12:52:43] <RikusW> then filter packets by device address
[12:53:21] <N1njAway> Love me some Xilinx
[12:53:49] <OndraSter> RikusW, nice, I just used it to check PXE booting process (I want to do some magic things with PXE at work)
[12:54:53] <RikusW> Its rather easy to see improperly or unhandled requests.
[12:55:09] <OndraSter> are there some requirements for it RikusW ?
[12:55:09] <RikusW> Helped me fix some usb requests..
[12:55:12] <OndraSter> like specific chipset etc
[12:55:14] <OndraSter> OS, ..
[12:55:31] <RikusW> In linux only usbmon driver
[12:55:36] <RikusW> its part of the kernel
[12:55:41] <RikusW> not sure about windows
[12:56:45] <RikusW> I think it was blathijs who told me to modprobe usbmon
[12:56:54] <RikusW> it helped a whole lot
[13:01:22] <megal0maniac_afk> For my fellow Windows users http://desowin.org/usbpcap/
[13:02:35] <RikusW> snoopy pro is nice too
[13:04:04] <OndraSter> thanks RikusW
[13:04:06] <RikusW> ah, WireShark driver for win
[13:04:13] <OndraSter> will look into it.. sometime
[13:04:26] <OndraSter> when I am not busy with work and school
[13:04:30] <OndraSter> or playing games
[13:04:35] <OndraSter> or making games?
[13:04:38] <RikusW> its a whole lot cheaper than hw usb analyzer :-P
[13:04:55] <RikusW> making ?!
[13:04:58] <OndraSter> :P
[13:05:03] <RikusW> what kind of game ?
[13:05:08] <OndraSter> you have not seen my RPG?
[13:05:10] <OndraSter> my awesomest RPG?
[13:05:12] <RikusW> no
[13:05:13] <OndraSter> my semestral project?
[13:05:13] <OndraSter> oh
[13:05:14] <OndraSter> sec
[13:05:22] <OndraSter> will upload two files ..
[13:05:29] <RikusW> how large ?
[13:05:36] <OndraSter> work-related, not game related
[13:05:37] * RikusW don't have ADSL anymore ...
[13:05:43] <OndraSter> I will have to find some screenshots, I lost somewhere already my demo files for it :(
[13:06:13] <RikusW> new place didn't have telco lines :(
[13:06:17] <OndraSter> d'aww
[13:06:48] <RikusW> Telkom is still in the process of deciding whether to install or not ...
[13:06:58] <RikusW> after 2 months...
[13:07:32] <OndraSter> http://pastebin.ca/2395611
[13:07:34] <OndraSter> my screenshot :P
[13:08:35] <OndraSter> it had exactly 0 memleaks and did not crash on wrong config files!
[13:09:42] <RikusW> heh
[13:09:50] <RikusW> put an RPG bot on IRC :-P
[13:09:54] <OndraSter> :D
[13:10:02] <OndraSter> I don't want to deal with sockets in C++ anytime soon :P
[13:10:16] <OndraSter> I still do have the source of my game of course
[13:10:49] <RikusW> make the next one a maze of PCB tracks, then its ontopic for #avr :-D
[13:10:51] <N1njAway> C++ socket programming isn't too bad.
[13:12:31] <N1njAway> Ondra: If you are doing game programming stuff, you may want to look at using RakNet for your if you don't want to deal with the low-level networking stuff on a more intimate basis -- http://www.jenkinssoftware.com/
[13:12:46] <N1njAway> your +implementation
[13:12:59] <OndraSter> I will make IRC RPG game that will run off some atxmega
[13:13:00] <OndraSter> :P
[13:13:07] <OndraSter> with ethernet chip
[13:13:15] <N1njAway> Then there you go. :)
[13:13:31] <N1njAway> Sorry, didn't see the deeper context. :)
[13:13:37] <beaky> so I have a random inductor
[13:13:43] <beaky> how do I find its inductance
[13:13:55] <RikusW> My DMM can measure it...
[13:13:58] <N1njAway> I would suggest WizNet for ease of implementation, but you are too socket-limited if you intend to have more than four players :)
[13:14:17] <beaky> wow how uch does your dmm cost
[13:14:25] <beaky> mine is 50 bucks and it cant :(
[13:14:27] <N1njAway> RikusW: You weigh the coil and cout the turns, multiply by pi, and make and educated guess!
[13:14:32] <N1njAway> count
[13:14:32] <beaky> lol
[13:14:38] <blathijs> RikusW: Good to hear usbmon was useful for you :-)
[13:14:40] <RikusW> mine cost $60 in 1997
[13:14:51] <RikusW> blathijs: very useful
[13:14:56] <beaky> 60 bucks in 1997 is like 100 bucks now
[13:15:03] <RikusW> about
[13:16:01] <RikusW> I have a cheapy DMM for $8 that only does V A R diode and transistor test
[13:16:08] <RikusW> DT830B
[13:16:26] <RikusW> the other expensive one is a Wavetek DM27XT, its not made anymore
[13:16:45] <beaky> wow
[13:16:54] <beaky> its antique you can sell it for a thousand bucks on ebay
[13:17:02] <RikusW> really ?
[13:17:17] <RikusW> not sure I want to
[13:18:14] * N1njAway hugs his Fluke meters.
[13:18:53] <beaky> wow i want a fluke
[13:18:56] <RikusW> my uncle measures his fence energize using a fluke...
[13:19:10] <RikusW> lets just say it doesn't work anymore...
[13:19:21] <RikusW> weirdly it seems Amps still works
[13:23:13] <RikusW> *measured
[13:24:29] <OndraSter> hehe
[13:24:36] <OndraSter> we have.. 2 or 3 flukes at work
[13:24:47] <OndraSter> and then some other "cheap" stuff (not cheap, but "reasonably expensive"
[13:25:37] <N1njAway> There's also the $10 Harbor Freight special
[13:25:41] <OndraSter> :D
[13:25:50] <OndraSter> well we are not a company
[13:25:52] <OndraSter> to start with
[13:25:54] <OndraSter> we are a middle school
[13:26:16] <RikusW> I don't use my Wavetek for HV stuff anymore....
[13:26:33] <OndraSter> I avoid HV stuff to start with :D
[13:26:37] <RikusW> There is something wrong with it that causes components inside to blow...
[13:26:46] <RikusW> so only < 100V...
[13:26:55] <inkjetunito> hv :(
[13:27:05] <RikusW> thats what the cheap DMM is for
[13:27:15] <RikusW> if it blows its a cheap replacement
[13:27:22] <N1njAway> But I wouldn't go sticking that in to a CAT III or CAT IV environment :)
[13:27:26] <OndraSter> :D
[13:27:45] <inkjetunito> a fly swatter turned out to have >= 3kV
[13:27:48] <RikusW> there was an optional 25kV probe for my DMM
[13:27:58] <RikusW> never got it...
[13:28:05] <RikusW> never needed it either
[13:28:20] <inkjetunito> at least the crackling inside the DMM made it sound like >= 3kV
[13:28:27] <N1njAway> RikusW: It's gets fun when you need to deal with arc-flash environments.
[13:28:38] <N1njAway> When using the wrong meter will likely kill you :)
[13:29:26] <OndraSter> that's why I like micros
[13:29:32] <OndraSter> they require 3.3V tops :D
[13:29:33] <OndraSter> 5V*
[13:30:30] <N1njAway> If you ever want some chilling visuals that drive the point home on how dangerous mains can be if not handled properly, there are a LOT of informative arc-flash videos on YouTube. :)
[13:30:31] * megal0maniac_afk attacks OndraSter with a dwarf
[13:31:50] <OndraSter> watttt
[13:32:24] <hjohnson> so yeah... question is... buy a real jtagice, or go with the cheapie fleabay unit
[13:32:31] <hjohnson> (xmega)
[13:32:33] <OndraSter> dragon
[13:32:33] <N1njAway> Get a real one.
[13:32:38] * megal0maniac_afk hugs his jtagice3
[13:32:41] <OndraSter> :D
[13:32:45] <OndraSter> well jtagice3 is now cheap indeed
[13:32:48] <OndraSter> still $99 I presume
[13:32:53] <megal0maniac_afk> Yip
[13:32:57] <hjohnson> the ones I've seen are $120
[13:33:00] <N1njAway> Very, and you have Atmel support for it.
[13:33:02] <megal0maniac_afk> And preliminary SAM support now too
[13:33:04] <hjohnson> where can I find it for $99?
[13:33:12] <megal0maniac_afk> Just the D5 at the moment
[13:33:12] <OndraSter> well $99 + VAT
[13:33:42] * hjohnson has a rather complex project now... the software development on this is going to be big...
[13:33:49] <hjohnson> strongly considering basing it on FreeRTOS at this point
[13:33:56] <megal0maniac_afk> Then definitely the Atmel option
[13:33:57] <N1njAway> How big? Like Monster Truck big?
[13:34:14] <megal0maniac_afk> N1njAway: For us, monster truck is a chevvy blazer
[13:34:17] <hjohnson> http://i.imgur.com/byZoyQ4.png
[13:34:21] <hjohnson> that's the board layout.
[13:34:25] <N1njAway> "BIG BIG BIG! That's right, Sunday Sunday Sunday!"
[13:34:41] <OndraSter> nicely packed
[13:34:47] <hjohnson> feature creep kept coming more and more
[13:34:57] <N1njAway> hjohnson: Doesn't look too bad.
[13:35:22] <hjohnson> nikosapi: 12 serial ports, plus an LCD, 5 items on the I2C bus, CAN BUS, etc...
[13:35:35] <hjohnson> and it's mixed signal, 3 (wel really 4) power rails..
[13:35:41] <OndraSter> nice
[13:35:46] <OndraSter> xmega128a1?
[13:35:53] <hjohnson> xmega128a1u
[13:35:56] <hjohnson> yeah
[13:35:57] <OndraSter> close enough
[13:36:07] <OndraSter> beautiful chip innit?
[13:36:09] <hjohnson> they finally got the 4 port EBI running on these things, from what I'm told.
[13:36:16] <hjohnson> or what i've foudn on avrfreaks
[13:36:17] <N1njAway> It really comes down to what the functionality is and how complex the processing needs to do. A lot of our stuff has been on about an equal complexity and it can go either way with or without an RTOS.
[13:36:51] <hjohnson> N1njAway: there are multiple different tasks that need to go on, trying to manage that in a huge-ass state machine would be a pain.
[13:37:09] <hjohnson> N1njAway: it's basically a central computer for my boat
[13:37:15] <N1njAway> Very nice board layout, BTW. I assume this is a 4-layer?
[13:37:21] <hjohnson> so it's pushing NMEA data around (I'll DMA it mostly)
[13:37:37] <hjohnson> N1njAway: yeah... signal top and bottom, 3v3 on layer 2, GND on layer 3
[13:37:42] <hjohnson> er the other way around
[13:37:45] <hjohnson> but yeah 4 layer
[13:38:30] <hjohnson> the two UARTs i'm using for most of the NMEA data have 64 byte buffers, so for the most part I can just DMA the entire sentence to the buffer in one shot, and not have to spend significant CPU time on it
[13:39:03] <RikusW> hjohnson: AVR Dragon is $50
[13:39:04] <N1njAway> hjohnson: Depending on what needs to be done, a lot of times you can avoid the RTOS by using a bunch of ISRs and timers to do the "background" heavy-lifting, and process through things in your mainline loop. But again without knowing the responsibilities of the applications and what all it needs to do, it's hard to say. FreeRTOS is certainly nice for managing multiple threads.
[13:39:20] <RikusW> megal0maniac_afk: was the jtagice mk3 $50 too ?
[13:39:35] <hjohnson> N1njAway: that's why i'm looking at it.
[13:39:37] <RikusW> ah $99
[13:40:01] <hjohnson> N1njAway: there's also going to be a keypad and graphic LCD, so there's uer interface tasks, there's battery monitoring tasks, satellite telemetry
[13:40:22] <hjohnson> (and/or GPRS) the top right module is set for an Iridium 9602 satellite modem
[13:40:29] <N1njAway> hjohnson: If you're investing that much in having boards done, I'd not skimp on $20 difference - buy the actual Atmel ICE3
[13:40:50] <hjohnson> yeah, that's my thought.
[13:41:04] <hjohnson> my BOM cost (not including the iridium) is ~$300 already
[13:41:09] <megal0maniac_afk> Haha! I wish...
[13:41:37] <megal0maniac_afk> (@ RikusW )
[13:41:44] <N1njAway> hjohnson: All of that could still be handled in ISRs and a mainline, but comes down to your comfort level with implementation of the code. No single right answer. What lets you get it done the most efficiently is the question, and a lot of that is personal preference.
[13:41:53] <N1njAway> Certainly keep us up to date - it sounds like a cool project!!
[13:42:15] <hjohnson> N1njAway: iv'e done it both ways.. the huge switch statements and so forth become really cumbersome to maintain code-wise...
[13:42:17] <N1njAway> I've recently done a bunch with GPS and NMEA data with Garmin OEM units and it was an absolute joy.
[13:42:41] <hjohnson> N1njAway: i'm using a trimble copernicus in my project
[13:43:20] <hjohnson> (there's also an GPS LNA under an RF can, since the iridium antenna isn't quite in the GPS band, bu close enough)
[13:43:25] <N1njAway> I prefer the Garmin 18x OEM units as they're very cheap, IP-67 rated for direct outdoor install, smaller than a hockey puck, and have 5m cables on them.
[13:43:44] <hjohnson> N1njAway: yeah, I've got one of those on my car (use it for APRS)
[13:44:15] <hjohnson> N1njAway: in my case, since I already had to run an L-Band antenna outside for the iridium, it was just easier to piggyback on that antenna. (I dont' want my boat to look like it has warts/be a porcupine)
[13:44:28] <N1njAway> Price is really cheap, too, once you're ordering in quantity, though even single units are still sub-$100 USD :)
[13:44:37] <N1njAway> LOL true
[13:44:46] <hjohnson> yeah, the trimble is $40 I think
[13:44:50] <hjohnson> in qty 1
[13:45:14] <hjohnson> the big score is I landed a scrap Honeywell HMR-3300 that is still good...
[13:45:18] <hjohnson> $450 heading sensor babye
[13:45:46] <bitd> How did you manage that hjohnson
[13:46:13] <hjohnson> bitd: we used to use them at work... customer sent in a system that was upgraded to a design that doens't use the honeywell... didn't want the scrap parts back.
[13:46:31] <hjohnson> we can't sell it, so well, uhm, it wound up in my desk drawer
[13:46:38] <N1njAway> This application was using GPS for timing sync - it's cool having two seperate boards connected to a logic analyzer and seeing timing edge margins line up to <100ns without any physical connection :D
[13:46:44] <bitd> Wow, you actually get to take those home?
[13:47:00] <bitd> Tight asses at my job would never allow that.
[13:47:39] <hjohnson> bitd: it was either they go in recycling or otherwise "idssapear" not used in the current design, can't be used for repairs (since they're used), and the customer didn't want them back.
[13:47:49] <hjohnson> so, well, I "dealt" with the problem. :)
[13:48:18] <hjohnson> and it's in a great little housing that will be perfect for attaching to the bulkhead in the head.
[13:48:20] <bitd> Hah, good for you :P
[13:48:25] <hjohnson> (this is a brain box for my sailboat)
[13:48:57] <hjohnson> mostly because I want to get the instrument data and AIS onto my iPad (hence the wifi module)
[13:49:04] <bitd> I see, well then you really did score.
[13:49:19] <hjohnson> and the sad part of this is, my BOM cost and so forth makes this not much more expensive than commercially available multiplexers that only do half as much
[13:50:18] <hjohnson> thankfully i don't initiall have ot do much parsing of the NMEA data, I just need to verify that it's valid (check the checksums), filter it, and just route it around.
[13:50:23] <megal0maniac_afk> It's more fun to do it yourself
[13:50:30] <hjohnson> megal0maniac_afk: that too
[13:50:54] <hjohnson> our battery monitor died recently as well, and I discovered a beautiful little IC from TI that will let me monitor the batteries
[13:50:58] <bitd> And always better, cause if something breaks, you will probably know what to replace./
[13:51:18] <hjohnson> well, repairing this board will be.. interesting...
[13:51:22] <OndraSter> hjohnson, you need a monitor that will monitor the monitor
[13:51:26] <hjohnson> (I will be reflowing it using hotplate or oven)
[13:52:10] <hjohnson> my actual main worr is the switchmode regulators
[13:52:23] <hjohnson> those are the biggest risk, since I've never designed something like that before.
[13:52:28] <hjohnson> and I have 3
[13:52:58] <hjohnson> followed the directions and reference designs as close as possible so fingers crossed.
[13:53:59] <megal0maniac_afk> Whoo! USB capture working in Windows
[13:53:59] <RikusW> hjohnson: a hint, a friend of mine usually assembles the power supplies first to check it
[13:54:11] <RikusW> otherwise you might fry an expensive module
[13:54:33] <hjohnson> RikusW: good point.
[13:54:44] <RikusW> megal0maniac_afk: so planning on snooping on the ice3 ?
[13:54:49] <hjohnson> RikusW: only problem is that reflowing in two steps is a pain
[13:55:06] <RikusW> he used a heatgun..
[13:55:20] <N1njAway> hjohnson: Good luck! If you were in the Cleveland area I'd say to just come over to our shop and run it through one of our ovens. Those high-mass thermal parts can be tricky.
[13:55:28] <RikusW> though he had problems with 0402's flying around :-P
[13:55:40] <hjohnson> N1njAway: heh
[13:55:57] <hjohnson> the high-mass parts are really the inductors, and I suppose the trimble and the wifly
[13:56:05] <hjohnson> everythign else is pretty low mass
[13:56:18] <N1njAway> Rikus: 0402s and QFNs can be done with a heatgun if you are careful. I had to do this on a prototype a while back, as it was a 30" long PCB that would not fit in our batch oven at the time, before we got our big 5-zone.
[13:56:45] <hjohnson> N1njAway: I do have a proper rework hotplate here.. no oven though
[13:57:00] <hjohnson> the funny part on this whole board is that I kept forgetting how small it actualy is
[13:57:04] <N1njAway> The preheat is exceedingly useful.
[13:57:29] <megal0maniac_afk> RikusW: I wasn't, but I can send you some dumps if you want?
[13:57:43] <hjohnson> N1njAway: i'm sort of thinking heat the board to say 150C for a period to get everything warm, then crank it until I get reflwo
[13:57:53] <RikusW> megal0maniac_afk: don't have use for it right now, but would be nice to have in future
[13:57:55] <hjohnson> (and maybe hit the high mass parts with some extra heat from a reflow gun
[13:58:04] <N1njAway> hjohnson: Yeah, you'll have that. I did extensive layout on a 3-board stackup that is about the diameter of a US nickel, and looking at it for 2 weeks on a 30" monitor zoomed way in... it's a shock when you print a 1:1 out to check the board and see all of your work reduced to the size of postage-stamp!
[13:58:12] <megal0maniac_afk> Cool. Well, I don't plan on blowing it up :)
[13:58:27] <RikusW> megal0maniac_afk: maybe just a read idcode to wet my appetite sometime :)
[13:58:45] <RikusW> when you have time
[13:58:46] <hjohnson> N1njAway: yep... it wasn't really until I dropped the DE9 on the board that I went "oh my god, that connector is huge!"
[13:59:10] <hjohnson> N1njAway: also the mounting holes for 4-40...
[13:59:11] <RikusW> megal0maniac_afk: Wireshark can save gzipped files, that I can then easily open :)
[13:59:13] <N1njAway> hjohnson: That would work. Ideally check the manufacturer's reflow plot for the paste you are using so you can make sure you are in the right realm to get proper flux activation prior to reflow.
[13:59:39] <N1njAway> hjohnson: LOL at DB9
[14:00:00] <hjohnson> N1njAway: eventually I hope to be able to do a bootloader so I don't always need the iCE.
[14:00:08] <RikusW> megal0maniac_afk: you can do a dump of the signature read when you have the m1284p hooked up again sometime
[14:00:17] <hjohnson> (I'm strattling two techs here.. I also wired up the USB, though that's further down the list
[14:00:28] <hjohnson> unless the open source USB libraries come of age.
[14:00:46] <megal0maniac_afk> RikusW: With or without target connected?
[14:00:53] <RikusW> with target
[14:01:00] * hjohnson fell to one last piece of featuritice last night..
[14:01:08] <hjohnson> added a digital barometer.
[14:01:20] <N1njAway> Very nice!
[14:01:55] <hjohnson> (one of the goals here is that every hour, the boat will beacon out its position, heading, speed, depth under keel, wind speed/direction, battery states, temperature, water temperature, and now air pressure
[14:01:55] <RikusW> hjohnson: that board already seems a bit overcrowded to me :-P
[14:02:13] <hjohnson> RikusW: well, i had that big open space to the right of the iridium module...
[14:02:18] <hjohnson> and the I2C run nice and close
[14:02:56] <hjohnson> for $22/mo, I can get a data plan from iridium that will give me 12000 bytes of satellite capacity..
[14:03:07] <megal0maniac_afk> RikusW: I can only do it with the 3.x firmware
[14:03:09] <hjohnson> not much, but more than enough to send out hourly 16 byte packets.
[14:03:30] <N1njAway> hjohnson: Oooh, that's awesome. What model is the module? I've wanted to play with the sat-modem stuff for a while, but only saw some of the stuff from Digi Int'l
[14:03:31] <RikusW> megal0maniac_afk: dumping doesn't work on v2 ?...
[14:03:44] <hjohnson> N1njAway: Iridium 9602
[14:04:03] <megal0maniac_afk> No, mine has V3 and AS6.1 won't accept 2.x
[14:04:16] <RikusW> megal0maniac_afk: thats fine
[14:04:31] <hjohnson> N1njAway: https://www.globalsatellite.us/sites/default/files/IRIDIUM%209602.jpg it's a cute little sucker
[14:05:13] <RikusW> megal0maniac_afk: a hint, select the device number to filter the packets, then export only filtered packets
[14:05:21] <hjohnson> N1njAway: the application is realy nice too... you send out the packet, and then their system emails that packet as a MIME attachement to an arbitrary email address
[14:05:42] <hjohnson> N1njAway: and it's two way, so I could send out a weather request, have the system pull down current weather conditions, and send it back via satellite.
[14:05:44] <N1njAway> hjohnson: That's AWESOME! Supplier?
[14:05:47] <RikusW> otherwise you might just export your USB mouse and keyboard data too ;)
[14:05:50] <megal0maniac_afk> RikusW: It works differently in Windows. USBPcap does the dump
[14:06:02] <megal0maniac_afk> And you select filter before capture
[14:06:07] <RikusW> ah
[14:06:10] <hjohnson> N1njAway: I haven't sourced it beyond seeing that it's available from multiple vendors online
[14:06:12] <N1njAway> hjohnson: Is this direct purchase through Iridium?
[14:06:18] <RikusW> usbmon in linux dumps all of it
[14:06:23] <hjohnson> N1njAway: no, iridium doens't deal in onesies and twosies
[14:06:46] <hjohnson> (I'm in the satcoms biz in my day job... we don't get to deal with iridium directly, and we probably sell $100k a year of their stuff)
[14:07:09] <N1njAway> Nice! I'm going to look in to this, regardless. I want one to play with. I love SatCom stuff.
[14:07:17] <megal0maniac_afk> RikusW: Sent :)
[14:07:23] <RikusW> thanks :)
[14:07:27] <hjohnson> N1njAway: the thing I really liked was the GPS passthrough
[14:07:32] <megal0maniac_afk> I'm supposed to be studying. That's why I'm on IRC, and still "afk"
[14:07:38] <megal0maniac_afk> And finding other things to do
[14:07:41] <megal0maniac_afk> Like USB dumps
[14:07:45] <RikusW> lol
[14:07:51] <hjohnson> so I can just have one antenna that oes both... when the module transmits, it temporarily disconnects the GPS receiver to protect it...
[14:07:54] <hjohnson> very nicely built
[14:08:08] <N1njAway> A friend of mine was involved with trying to purchase Iridium when they were Chapter 11 about a decade ago or so. He has a few of the test antenna assemblies sitting around in his office as decorations.
[14:08:18] <hjohnson> N1njAway: the hardest part was finding the mating connectors that would fit... neither digikey nor mouser carried them... so I had to make a small order from newark
[14:09:07] <hjohnson> it needs a specific samtec mmcx connector which is shorter than most pcb mount mmcx plugs
[14:09:21] <RikusW> megal0maniac_afk: did you start capture before connected using studio ?
[14:09:30] <megal0maniac_afk> No
[14:09:41] <megal0maniac> Started capture, read signature, stopped
[14:10:23] <hjohnson> N1njAway: the one thing I'm going to do is buy most of the parts before I send the pcb out for manufacture.... that way I can test fit the componentst hat I'm worried about to a 1:1 printou to make sure they'll fit
[14:10:48] <N1njAway> hjohnson: Yeah, the wonders of special connectors. Linx Technologies has made up custom RF cables for us for antennas and such in the past, and they have a nice selection of off-the-shelf SMA and RSMA connectors and antennas, too.
[14:11:13] <N1njAway> hjohnson: https://www.dropbox.com/s/cp0etve994id8h0/irsatant.jpg is a picture of that Iridum antenna assembly, BTW
[14:11:37] <hjohnson> for a completley unrelated project, I'm looking at this linkblox? things... I'd love to build some ultra-compact IP enabled, POE powered temperature/random sensors
[14:11:42] <N1njAway> hjohnson: The thing that's amazing is when you start looking at it, and realize the entire frame piece is milled out of a single aluminum billet and is all ONE PIECE
[14:12:03] <RikusW> megal0maniac: seems the Linux version of Wireshark doesn't know network type 249... will figure that out...
[14:12:04] <hjohnson> ahh, that's ku-band, if my scale is right.
[14:12:12] <hjohnson> (the waveguide opening is roughly ku-band size)
[14:12:12] <N1njAway> hjohnson: I like the kapton over the waveguide :)
[14:12:32] <hjohnson> N1njAway: that's normal for me. :P (I deal in C/ku/X band systems in my day job
[14:12:33] <megal0maniac> RikusW: I had the same problem. You running 1.8.x?
[14:12:38] <hjohnson> we go through roll after roll of kapton
[14:12:40] <megal0maniac> 1.10 is latest stable
[14:12:53] <RikusW> 1.8.2
[14:13:07] <hjohnson> N1njAway: that looks like one of the inter-satellite connection antennas
[14:13:14] <megal0maniac> Yeah, you'll need a newer version
[14:13:35] <N1njAway> hjohnson: I just love the precision of all of it. I think these antenna assemblies were somewhere around $900K each for the production version. My friend managed to scoop this up for a couple grand I believe. Still an expensive office decoration, but...
[14:13:58] <N1njAway> hjohnson: Yeah, inter-satellite if I remember correctly.
[14:14:04] <N1njAway> hjohnson: Fascinating stuff.
[14:14:35] <hjohnson> N1njAway: I was doing a job for a guy at a certain organization... sitting on a shelf behind his desk was this wierd looking piece of metal... I ask him what it is and he replied "oh that? that's the prototype I built for the atomic clocks that fly on the GPS satellites... It's isotopically pure titanium"
[14:15:03] <N1njAway> I guess I should go build some PCBs, and I forgot the laptop today. I should really fire up ustream with the USB tap on the one stereoscope and give you guys the URL, so I can entertain you with my soldering madskillz and endless bad jokes XD
[14:15:18] <hjohnson> lol
[14:15:20] <N1njAway> hjohnson: NICE! Bet that was a spendy part.
[14:15:24] <megal0maniac> N1njAway: Yes!
[14:15:28] <hjohnson> N1njAway: i'm sure
[14:15:35] <hjohnson> it was about the size of a football
[14:15:44] <N1njAway> megal0maniac: I'll totally do that soon.
[14:16:59] <hjohnson> anyhow, it's lunch time and I need to get down to work
[14:18:19] <hjohnson> on a completely unrelated note, does anyone have a library or know where I could find one, to easily parse an AT command interface?
[14:18:26] <hjohnson> I'd rather not re-invent the wheel if I don't have to
[14:22:10] <OndraSter> in the name of my brain thinking "I can do it better" I always re-invent all the wheels
[14:22:54] <N1njAway> OndraSter: http://xkcd.com/927/
[14:23:12] <OndraSter> hehe
[14:23:21] <OndraSter> I mean, apply with the standards but re-invent the code
[14:30:14] <megal0maniac> N1njAway: My favourite part is the pop-up text
[14:37:23] <N1njAway> megal0maniac: Very yes!
[15:08:47] <hjohnson> the wonderful thing about standards is there are so many to choose from
[15:21:01] <OndraSter> yep
[15:21:36] <hjohnson> i've started to buy parts now, just to keep myself from adding anything more
[15:21:44] <hjohnson> that board is already about as packed as a board can be :P
[15:22:07] <OndraSter> there are usually at least three: Microsoft, Oracle/Novell/.., Opensource. The Oracle's/Novell's does stuff simply but terribly, Opensource does half of the things or all of them but in a crappy way and MS's is paid and works brilliantly, if it works. If not then you are screwed.
[15:22:31] <hjohnson> heh
[15:22:55] <hjohnson> I've lost most of the snobbery I had in my youth...
[15:23:02] <OndraSter> and sometimes google comes along, takes opensource, turns it into 3rd-nonMS-party, keeps it opensource and makes a moss of it
[15:23:11] <OndraSter> mess*
[15:23:18] <hjohnson> I help out with a non-profit, and have them now completely on Active Directory for all their authentication and account needs
[15:23:22] <hjohnson> makes life so much easier.
[15:23:32] <vsync_> avrfreaks down? ;(
[15:24:05] <OndraSter> hjohnson, I work as an assistant in electrotechnical labs, but right now I am setting up a new network stuff instead :D. WS2012 based network
[15:24:35] <hjohnson> yeah, I'm sticking with 2k8R2
[15:25:13] <hjohnson> but it's a multi-platform environment.. the actual file servers are NASs runnign samba on linux, the caching server is linux, etc...
[15:25:23] <hjohnson> a lot of macs in the mix, wireless is ubiquiti
[15:25:40] <hjohnson> though I now have all the underlying network running on cisco... that was a huge win
[15:25:44] <OndraSter> :)
[15:26:06] <OndraSter> we do have cisco switches, alongside with some planetsomething and nortels. But the actual IT admin is a d.ck so it barely works
[15:26:09] <hjohnson> (I actually really like cisco gear... due to their support policies, there's a lot of really good, high quality equipment available on the used market for not much money)
[15:26:46] <OndraSter> I was supposed to be doing electronics. But I accidentaly replied "kinda" on the question if I know something about servers ... :( :D
[15:27:08] <OndraSter> so instead of fixing electronics and making some awesomesauce measuring stations I just sit at PCs and write powershell scripts
[15:27:35] <hjohnson> my condolances
[15:27:43] <OndraSter> haha
[15:27:50] <OndraSter> well I will get back to electronics hopefuly soon
[15:28:14] <hjohnson> the network I help run is.. interesting... small campus network, interconnects 23 buildings... and has a 1mbps connection to the otuside world
[15:28:26] <OndraSter> 1mbps..
[15:28:28] <OndraSter> we have 20 :D
[15:28:42] <hjohnson> OndraSter: oh, and that's 1mbps/300kbps... via a dedicated satellite link
[15:28:48] <OndraSter> nice
[15:28:59] <OndraSter> you better be running WSUS on the server :D
[15:29:01] <hjohnson> with some 70+ people trying to use it
[15:29:03] <OndraSter> for Windows machines
[15:30:14] <hjohnson> yeah, I'm not so sure what they're doing on that situation... they might have deliberately broken windows update (shudder)
[15:30:44] <hjohnson> the other challenge is there's a fair amount of BYOD, and since those machines aren't in the domain, WSUS won't help
[15:31:29] <OndraSter> if the people would be willing to import stuff into their registry... ]:->
[15:31:36] <OndraSter> (or set a poliyc)
[15:31:38] <OndraSter> policy*
[15:31:50] <hjohnson> yeah, but don't want to break their machines when they leave again
[15:31:54] <OndraSter> back to AVRs, before somebody comes along and kicks us for being off topic...
[15:31:55] <OndraSter> sure
[15:32:21] <hjohnson> (they operate a retreat centre high in the mountains of washington state... there are a lot of 3 week volunteers... )
[15:32:25] <hjohnson> hehe
[15:32:47] <hjohnson> what I'm really looking for is a small gizmo taht would let me build a compact PoE ip enabled sensor/actuator.
[15:33:26] <OndraSter> I actually wanted to build that, but instead of PoE use a battery and build it around NRF24L01 wireless chips
[15:33:56] <specing> OndraSter: 20? we have *atleast* a gigabit per second
[15:34:04] <hjohnson> OndraSter: I'm about to depoy a whack of PoE switches, so doing PoE makes sense.. (I'm trying to get mission-critical stuff off of wireless)
[15:34:38] <OndraSter> specing, only the new (three) servers are Gbit
[15:34:46] <OndraSter> and maybe some cisco switches have uplink ports gbits
[15:35:00] <OndraSter> but the cables are barely catching on 100Mbit with close to zero packetloss :D
[15:35:10] <specing> fibre...
[15:35:13] <OndraSter> haha
[15:35:20] <hjohnson> OndraSter: you can get cisco 2970Gs for like $250 these days
[15:35:21] <OndraSter> the uplink to the internet is 20mbit... no need for more really
[15:35:27] <OndraSter> hjohnson, I don't need it
[15:35:34] <hjohnson> 24 gigabit ports plus 4 independent SFPs
[15:35:44] <OndraSter> the client PCs are 100Mbit.
[15:36:14] <OndraSter> and the servers will be internally connected on their second cards, yes. That is planned
[15:36:17] <OndraSter> on Gbit
[15:36:22] <OndraSter> (replication and backups for now)
[15:37:46] <OndraSter> hmm I am running atmel studio 6 SP2
[15:37:54] <OndraSter> wondering when I launched it the last time
[15:38:49] * hjohnson has done all his atmel development under linux.. vim, gcc-avr, and makefiles. :P
[15:38:57] <OndraSter> pff
[15:39:00] <OndraSter> I like GUIs
[15:42:52] <OndraSter> and gdb of course hjohnson
[15:43:30] <hjohnson> well, i've never had a jtag-ice so i had to debug the hard way
[15:43:36] <OndraSter> oh
[15:43:37] <OndraSter> serial ftw
[15:43:39] <hjohnson> (flash LEDs, dump strings to USARTs, etc...)
[15:47:02] <vsync_> like beaky would say
[15:47:16] <vsync_> i love X... so umm... i love parallel
[15:47:18] <vsync_> and wires
[15:47:26] <hjohnson> heh
[15:47:26] <vsync_> parallel>serial
[15:47:48] <hjohnson> i'm crossing my fingers that the EBI setup i'm depending on will work properly
[15:52:28] <OndraSter> hjohnson, which one?
[15:52:42] <OndraSter> the bitchy thing about xmega is that it cannot do 16bit data bus D:
[15:53:07] <OndraSter> which makes it unusable for DDRs really
[15:53:09] <OndraSter> err, SDRs
[15:53:15] <hjohnson> OndraSter: xmega128A1u
[15:53:18] <OndraSter> yxes
[15:53:21] <OndraSter> but which memory
[15:53:56] <hjohnson> oh, 2x NXP quad UARTs, a kyocera LCD module, and a 128KB SRAM
[15:54:03] <OndraSter> oh, SRAM
[15:54:16] <OndraSter> 128k * 8 I presume
[15:54:17] <hjohnson> I found a couple of references on avrfreaks that the 4 port mode actually works on it
[15:54:23] <OndraSter> ohh
[15:54:30] <hjohnson> yeah, 128KB, 1Mb
[15:54:30] <OndraSter> abcminiuser told me it does not :P
[15:54:40] <OndraSter> I wanted to try out the full 128Mbit SDRAM
[15:54:51] <hjohnson> OndraSter: it apparently works on the A1U
[15:54:53] <OndraSter> but .... 8bit ones are unreachable and 256Mbit are 32bit :(
[15:55:03] <OndraSter> so 8MB SDR is the most I could find
[15:55:23] <hjohnson> my goal was to avoid having the address latch
[15:55:26] <OndraSter> :)
[15:55:28] <hjohnson> maybe that was a fatal mistake.
[15:55:42] <OndraSter> with 4 ports you get how much? 20bit address?
[15:55:51] <hjohnson> yeah
[15:55:51] <OndraSter> upto*
[15:55:57] <OndraSter> that's 1MByte :P
[15:56:05] <hjohnson> OndraSter: more than I need
[15:56:06] <OndraSter> but they get expensive in 1offs :(
[15:56:25] <hjohnson> the 1mbit sram is only $3 or something
[15:56:29] <OndraSter> aye
[15:56:35] <hjohnson> using a 12ns part so I can hopefully run the bus at 64MHz
[15:56:36] <OndraSter> but not 8Mbit
[15:56:42] <OndraSter> hehe
[15:56:53] <OndraSter> you can do for 128MHz for EBI, or is it just for the PWM only?
[15:56:58] <OndraSter> can't remember
[15:57:09] <hjohnson> the EBI is clocked from ClkPer2
[15:57:13] <OndraSter> ah
[15:57:15] <hjohnson> which is usually 2x the system clock
[15:57:18] <OndraSter> I thought you could do for clkper4
[15:57:23] <OndraSter> still, good enough
[15:57:46] <hjohnson> yeah, the SRAM will be 0 waitstate, the UARTs 1 wait state, the LCD? 7 wait states.
[15:57:53] <OndraSter> :)
[15:57:56] <OndraSter> some colour one?
[15:58:09] <hjohnson> naw, 128x64 mono LCD I have kicking around
[15:58:13] <OndraSter> ah
[15:58:16] <hjohnson> (white on blue)
[15:58:21] <OndraSter> I have one of those here as well
[15:58:24] <OndraSter> better than black on green
[15:58:36] <hjohnson> yeah
[15:59:47] <OndraSter> I want to do one of those 240x320 coloured
[15:59:58] <OndraSter> but - you need lots of memory if you want to do a local framebuffer :P
[15:59:58] <hjohnson> I was hoping to find a way to do bi-colour LED for th ekeypad, but couldn't find a compact/cheap way to do it... so red backlight it is
[16:00:05] <hjohnson> yeah
[16:00:26] <hjohnson> in my case for the mono, I'll stick the local frame buffer in the internal SRAM, then I can just DMA it out
[16:00:33] <hjohnson> (hurray for DMA)
[16:00:53] <OndraSter> xmegas are awesome, but the small and cheap ARMs are taking over which makes me angry
[16:07:16] <hjohnson> eh, imho, they have different needs/purposes.
[16:07:40] <OndraSter> well yes
[16:07:44] <OndraSter> but everybody rather gets the ARM
[16:07:46] <OndraSter> instead of xmega
[16:07:50] <OndraSter> even when xmega would be better
[16:08:04] <hjohnson> it's what is in style I suppose
[16:08:10] <OndraSter> exactly
[16:08:12] <OndraSter> it is cool :D
[16:08:15] <OndraSter> to use ARM
[16:08:38] <OndraSter> wow xmega128a1u is quite cheap on mouser
[16:08:57] <OndraSter> 100pcs at 2.49€
[16:09:07] <OndraSter> considering what peripherals it has...
[16:09:24] <OndraSter> oh wait, they have only BGA on stock
[16:09:25] <OndraSter> the TQFP is not
[16:09:43] <OndraSter> same for farnell
[16:09:43] <beaky> hello
[16:09:48] <OndraSter> 'sup beaky
[16:10:37] <OndraSter> only digikey has them on stock
[16:10:42] <OndraSter> but digikey pricing for shipping... NO.
[16:10:55] <OndraSter> I would have to order worth of $200 xmegas :D
[16:11:46] <OndraSter> either 62 for $200 or 100 for 288.48
[16:11:48] <OndraSter> lol
[16:11:51] <OndraSter> which is not bad
[16:11:58] <OndraSter> but then add taxes and what not..
[16:12:04] <OndraSter> and I don't need so much.
[16:12:14] <hjohnson> I dunno, my order from digikey is only $8 in shipping
[16:12:21] <OndraSter> well I am not from the USA
[16:12:24] <hjohnson> well, will be free if I add the jtagice3
[16:12:34] <hjohnson> OndraSter: I'm in .ca (but yeah, they have local canadian shipping)
[16:12:40] <OndraSter> :P
[16:12:44] <OndraSter> I am in the middle of the europe
[16:14:12] <beaky> so only digikey has arms?
[16:14:19] <beaky> do they sell arms to hobbyists?
[16:14:27] <OndraSter> only legs and livers
[16:15:16] <beaky> why use xmega when you can buy arms for $1
[16:16:16] <OndraSter> because those xmegas contain more peripherals
[16:16:48] <beaky> ah
[16:16:52] <beaky> like what
[16:17:06] <OndraSter> $1 ARM won't have dual 2MSPS ADC and 1MSPS DAC
[16:17:13] <beaky> ah right
[16:17:25] <OndraSter> just check the datasheets
[16:17:29] <OndraSter> 128MHz PWM
[16:17:31] <OndraSter> etc
[16:19:32] <vsync_> i heard arm was gud apel uses it n shit
[16:19:41] <vsync_> innit so beaky
[16:19:55] -OndraSter:#avr- lol
[16:20:57] * antto slaps OndraSter with a hot iron
[16:21:27] <OndraSter> :P
[16:23:10] <beaky> yes they dont use avr in appel
[16:23:28] <beaky> but i use it for all my projects :D
[16:23:45] <vsync_> so... did you get the replacement 328
[16:23:57] <beaky> yes
[16:24:05] <vsync_> has it released magic smoke
[16:24:08] <vsync_> already
[16:24:43] <beaky> no
[16:27:09] <vsync_> got one of those raspberry pis?
[16:27:23] <vsync_> i hear they are very good for blink_led.c
[16:28:05] <beaky> yes i have a pi
[16:28:07] <vsync_> chances are you can control the io with some really cool zero overhead stuff like python or ruby
[16:28:26] <beaky> problem with pi is it is not as energy efficient as arduino
[16:28:29] <beaky> and pi has no pwm
[16:28:32] <beaky> no adc
[16:28:49] <vsync_> you can still pwm the pi right? to increase efficiency?
[16:29:18] <beaky> oh yeah
[16:30:28] <vsync_> Did you know, that smps aren't really efficient?
[16:30:34] <beaky> ?
[16:30:35] <beaky> why
[16:30:38] <vsync_> like, they can only get to around 95% or so
[16:30:58] <beaky> yeah that sucks
[16:31:02] <vsync_> Why not power your Pi straight out of mains? I'm sure it has a "reg on board"
[16:31:14] <twnqx> primary line smps?
[16:31:17] <twnqx> side*
[16:31:50] <vsync_> beaky: more efficient when you ditch the smps adapter in between
[16:32:05] <vsync_> just get a nice lil mains socket to dc jack cable and go
[16:32:41] <beaky> ok ill try
[16:32:47] <vsync_> ok really don't
[16:33:09] <beaky> why? mains has embedded 60Hz pwm, and ditching the brick yeilds 100% efficiency
[16:33:22] <beaky> and will cook my pi good
[16:33:48] <vsync_> what?
[16:34:06] <beaky> alternerating current
[16:35:01] <vsync_> well it's just a bit like pwm dc right?
[16:35:08] <vsync_> you can ask N1njAway he knows all about it
[16:37:06] <vsync_> i'm not sure whether you were kidding or not, but i suggest you never try anything you hear me suggest, just for the sake of me being able to fall asleep in the night okay?
[16:37:22] <beaky> :D
[16:37:52] <OndraSter> and for the sake of earth, do not play with anything on mains
[16:37:53] <beaky> the ghosts of all my smoked arduinos shall haunt your dreams ^^
[16:37:56] <Tom_itx> vsync_ don't give out bad information then
[16:37:56] <OndraSter> limit yourself to 24V
[16:38:15] <vsync_> Tom_itx: yeah :)
[16:39:10] <beaky> but an arduino-based smps sounded like a good idea in my head
[16:39:20] <beaky> then you can set output voltage digitally!
[16:39:29] <OndraSter> imagine the possibilities!
[16:39:34] <beaky> (even if the efficiency would be like 60%
[16:39:59] <beaky> can you set output on an analog smps like max638
[16:39:59] <OndraSter> you can do it with dedicated chips and digital pots
[16:40:07] <beaky> ah
[16:40:36] <vsync_> beaky: why on earth, would you want an arduino base smps?
[16:40:52] <beaky> its a good learning experience
[16:41:00] <OndraSter> no
[16:41:12] <beaky> and the stakes are high; do it wrong and things explode in delightful, colorful ways!
[16:41:13] <OndraSter> SMPS is more than a PWM and a comparator
[16:41:31] <beaky> yes there is also capcitor involved
[16:41:38] <OndraSter> that's not what I ment
[16:41:42] <vsync_> :D
[16:42:05] <vsync_> yeah totally. you'd want to add an inductor also...
[16:42:46] <amee2k> i thought you make a digitally controlled PSU by making a normal analog control loop
[16:42:52] <amee2k> and vary the reference with an MCU
[16:43:21] <beaky> aka pwm and a comparator!
[16:43:49] <amee2k> a regulator control loop has an error *amplifier*, not a comparator
[16:44:18] <amee2k> if you have just a comparator then you have MC34063 style "hichup" regulation
[16:44:35] <OndraSter> NOT THE MC34063!!
[16:44:35] * OndraSter runs and hides
[16:44:35] <vsync_> beaky: if you want to do something with power and pwm i have a project for you
[16:44:47] <vsync_> which is also a great learning project
[16:44:53] <beaky> wow what kind of project
[16:44:55] <amee2k> if you want power and PWM you could make a brushless motor driver
[16:45:03] <beaky> (btw why run away from mc34063)
[16:45:05] <vsync_> yeah. that was my idea also
[16:45:13] <beaky> whats a brushless motor driver
[16:45:14] <vsync_> a bl speed controller
[16:45:30] <OndraSter> because MC34063 is a noise generator pretty much :D
[16:45:32] <vsync_> a controller circuit used to drive electric brushless motors
[16:45:44] <amee2k> essentially a variable frequency 3-phase inverter with feedback
[16:45:53] <OndraSter> whenever I see something damaged @ PSU, it contains mc34063
[16:45:56] <OndraSter> *always*
[16:46:11] <beaky> thats because mc34063 is ubiquitous
[16:46:24] <amee2k> the 34063's control loop is essentially unstable by design
[16:46:27] <beaky> open up any cheapo psu, and it has one
[16:46:49] <amee2k> it just pulses the output at a fixed rate until the comparator turns off
[16:47:03] <amee2k> then it does nothing until the comparator turns on again
[16:47:22] <amee2k> and then it flickers on and off in bursts to keep the output somewhere around the setpoint
[16:47:33] <vsync_> but yeah beaky there's a project for you. I'm afraid there'll be some smoke involved
[16:47:51] <beaky> an arduino-powered grill?
[16:47:59] <twnqx> lol
[16:48:05] <vsync_> how about ditch the fucking arduino already
[16:48:07] <amee2k> the problem with that strategy is that you can get high-amplitude ripple WAAAAAY below the main switching frequency
[16:48:09] <twnqx> arduino doing the grilling, no extras needed :P
[16:48:23] <vsync_> grill it. and get on with your life
[16:48:24] <amee2k> so you need to filter like 500mVpp @ 1kHz out of your supply rail
[16:48:48] <beaky> my arduino is already toast :( yesterday i connected a 12v supply backwrds to it
[16:48:53] <twnqx> nothing a capacitor doesn't solve
[16:48:58] <beaky> and it became very hot
[16:49:05] <twnqx> lol
[16:49:17] <amee2k> twnqx: yeah but you need a huge ass filter, which is what an SMP is supposed to avoid
[16:49:17] <twnqx> sorry for your loss.
[16:49:21] <OndraSter> didn't use to be there a diode?
[16:49:37] <beaky> yeah but maybe the diode doesnt block the +5v
[16:49:44] <twnqx> it should...
[16:49:45] <beaky> (which i pluged my batttery into)
[16:49:50] <twnqx> oh god
[16:49:51] <amee2k> fixed frequency controllers have massively lower subharmonics on the output
[16:50:17] <beaky> arduino has fixed frequency if you program it right
[16:50:22] <beaky> so maybe it can be decent smps
[16:50:29] <vsync_> seriously...
[16:50:38] <beaky> ATmega328p*
[16:50:49] <twnqx> you know that sms controllers just are simpler
[16:50:52] <twnqx> smps*
[16:51:04] <vsync_> no. you totally need 20 MIPS for it
[16:51:04] <OndraSter> mc34063 is simple
[16:51:05] <twnqx> two capacitors, one chip, one coil, done
[16:51:20] <OndraSter> proper SMPS controllers from TI that cost $8+ are not simple.
[16:51:20] <beaky> no diode?
[16:52:05] <vsync_> i love coils
[16:52:11] <vsync_> don't you beaky?
[16:52:11] <OndraSter> tesla coils <3
[16:52:22] <twnqx> tps62205 is simple :P
[16:52:36] <N1njAway> I like the L7980 as well. Diode, coil, couple caps, couple resistors. VERY well protected part, too.
[16:52:45] <twnqx> smps for primary side? or secondary?
[16:52:53] <beaky> there are 3 input smps
[16:52:58] <beaky> i think
[16:53:02] <beaky> 3 pin*
[16:53:14] <N1njAway> 3 pins for three-phases.
[16:53:39] <OndraSter> twnqx, yeah, these small sot23-5 tiny controllers are usually awesome
[16:53:42] <OndraSter> cheap and very powerful
[16:54:00] <N1njAway> vsync_: Best smsp ever is the 5 pin device in the SOT23-5 package, takes three-phase power and outputs +5V regulated.
[16:54:14] <twnqx> lol
[16:54:29] <twnqx> 5V, 6kW :P
[16:54:46] <OndraSter> at 99.999% efficiency, why not
[16:55:29] <amee2k> if you want a simple SMP, pull the 494 out of an old computer PSU
[16:55:42] <amee2k> it'll beat the shit out of a 34063
[16:55:53] <vsync_> N1njAway: so. I have readily available three phase mains at work
[16:55:57] <OndraSter> ayay cap'n
[16:56:00] <vsync_> can i drive some duinos?
[16:56:06] <amee2k> the atmega probably *can* work as an SMP controller
[16:56:07] <OndraSter> we have three phase even at home
[16:56:14] <N1njAway> vsync_: It makes the Arduinos run FASTER.
[16:56:17] <amee2k> but firstly, make sure your output caps are rated for the full input voltage :>
[16:56:20] <OndraSter> three times as fast!
[16:56:23] * N1njAway loves his 3-phase here!
[16:56:27] <OndraSter> 60MIPS at 20MHz
[16:56:28] <vsync_> THREE(3)
[16:56:42] <vsync_> nice. and all in sot23-5!
[16:56:47] * specing unearths a three and smacks OndraSter with it
[16:56:58] <vsync_> that's some hella stuff going on in there.
[16:57:02] <N1njAway> OndraSter: Actually it's only 2.12 times faster, since you have to use RMS Hz for 3-phase
[16:57:07] <OndraSter> :P
[16:57:11] <OndraSter> g'dammit
[16:57:21] <amee2k> and secondly, you get to deal with all the control loop theory in software
[16:57:25] <vsync_> well i can dig 40 mips @ 20 mhz
[16:57:27] <vsync_> or well
[16:57:30] <vsync_> a bit more...
[16:57:42] <amee2k> because you'll need to do the PID-style control *and* compensation in the firmware
[16:58:09] <N1njAway> Uses the Atmel opcode called "DCDC" which takes two registers as the operands - the incoming voltage, and the voltage to regulate to, respectively.
[16:58:57] <beaky> wow there is an instruction like that
[16:59:22] <OndraSter> beaky, this is the internet which means that it has to be the truth
[17:00:30] <N1njAway> There is also an instruction called "ACAP" which allows you to specify additional capacitance equivalence in uF without having to add additional off-board capacitors. It's kind of like a software-based gyrator. :)
[17:00:41] <N1njAway> Man, we need to start keeping a list of all these special opcodes!
[17:01:09] <vsync_> N1njAway: How to implement inductance?
[17:02:06] <N1njAway> "ACOIL" I think?
[17:03:06] <amee2k> beaky: well, if it isn't obvious by now... N1njAway is pulling your tail :3
[17:04:05] <N1njAway> Gotta keep people on their toes. It's sophisticated Atmel humor :)
[17:04:16] <amee2k> not entirely, no
[17:05:31] <beaky> wow eevblog has easy tutorial on building smps
[17:05:48] <twnqx> oh god
[17:05:56] <vsync_> this is bad
[17:06:10] <OndraSter> NOT THE EEVBLOG! D:
[17:06:13] <vsync_> on a totally unrelated note, has anyone got experience of abebooks?
[17:06:19] <vsync_> thing seems like a ripoff to me
[17:06:37] <OndraSter> the only thing I watch from eevblog are teardowns of some stuff
[17:06:53] <OndraSter> (NOT the new scopes etc where he is like "watch what I just got - $5k scope!"
[17:07:40] <amee2k> THE SQUISHED TESTICLES VOICE MAN
[17:08:05] <OndraSter> haha
[17:08:08] <OndraSter> exactly
[17:08:15] <N1njAway> High Rising Terminal has never been so.... irritating.
[17:08:28] <OndraSter> and when he has something stuck in his throat.. I have to cough for him myself
[17:08:30] <N1njAway> I want to run eevblog through AutoTune every time I hear it. :)
[17:08:34] <OndraSter> lol
[17:08:48] <vsync_> I think the guy got booted out of microchip just due to sheer lack of credibility because of his voice
[17:08:50] <amee2k> actually
[17:09:02] <amee2k> pop out an EQ and turn down everything over 3-4kHz
[17:09:08] <amee2k> that helped me :3
[17:09:18] <N1njAway> lol his voice just phases in and out
[17:09:22] <amee2k> it sounds like a phone call but at least the headaches get better
[17:09:34] <vsync_> I thought you wouldn't hear anything at all then
[17:09:39] <OndraSter> so did I
[17:10:05] <N1njAway> Even LUFA Dean (an Aussie) commented that his voice is a bit absurd - Australian is known for having High-Rising Terminal as a common trait, but eevblog's is really excessive.
[17:10:33] <OndraSter> LUFA Dean.. I like that name :D
[17:11:00] <N1njAway> It rolls off the tounge/keyboard
[17:11:31] <amee2k> http://www.youtube.com/user/mjlorton?feature=watch
[17:11:35] <amee2k> i kinda like that guy better :3
[17:11:43] <OndraSter> I like mike
[17:11:49] <OndraSter> he does not shit around :D
[17:11:52] <OndraSter> and just tears it down
[17:11:53] <N1njAway> LUFA Dean Breakfast Sausage (for those in the US)
[17:11:53] <hjohnson> has LUFA been ported to work properly on the xmegas yet?
[17:12:03] <OndraSter> hjohnson, I think so
[17:12:06] <OndraSter> I wrote my own stack :D
[17:12:23] <OndraSter> USB + CDC
[17:12:39] <vsync_> center for disease control
[17:12:46] <OndraSter> pretty much
[17:12:50] <OndraSter> it is that awful, yes
[17:12:50] <N1njAway> vsync_: LOL that's what I was thinking
[17:13:15] <OndraSter> USB is a huge, huge, HUGE hair ripper to get started with
[17:13:20] <OndraSter> but once you get hang of it, it is only terrible
[17:13:20] <N1njAway> What was I doing when I wandered in here before getting sucked in to IRC? Oh yeah, printing BOM for all these boards being built right now.
[17:13:45] <vsync_> i thought it was going to be followed with the buttsex analogy
[17:13:49] <hjohnson> OndraSter: that's why it's a bit of a ways down my priorities list... i've just left support htere for it
[17:13:56] <OndraSter> haha
[17:14:07] <OndraSter> LOL vsync_ ... yeah, that could work
[17:14:57] <amee2k> other than that, i search for videos thematically, i.e. reviews of products and shit
[17:15:13] <amee2k> i couldn't give less of a fuck about who shoves it into the camera
[17:15:19] <amee2k> and if he likes the color or whatever
[17:15:23] <N1njAway> "USB: It only hurts the first few hundred times..."
[17:15:43] <amee2k> i just want a view of the product in operation, and an inside shot of it
[17:15:59] <amee2k> not the photoshopped marketing pics that all the places selling them use
[17:16:00] <OndraSter> N1njAway, that depends.. are you "the port" or "the peripheral"? :D
[17:16:06] <OndraSter> (or the cable)
[17:16:35] <N1njAway> lol too true
[17:16:45] <vsync_> "... and then you'll have fecal incontinence"
[17:16:46] * amee2k puts a condom over OndraSter's cable
[17:16:47] <N1njAway> Okay back to SMT assembly
[17:17:01] <OndraSter> N1njAway, reflow ovens ftw
[17:17:10] <vsync_> i think he's going to oven it...
[17:17:26] <OndraSter> I got some oven getting ready to make a simple controller
[17:17:32] <OndraSter> but never made it
[17:17:54] <N1njAway> Oven for sure!
[17:18:06] <amee2k> http://vimeo.com/65763753 "I Hate Myself For OVEN You"
[17:18:18] <N1njAway> But this is a really small quantity, so will probably just use the batch oven instead of firing up the five-zone
[17:18:30] <N1njAway> Though it IS getting a bit cool out, and the 5-zone heats the room nicely.
[17:18:53] <beaky> oven?
[17:18:57] <beaky> baking chips?
[17:19:07] <OndraSter> baking crisps
[17:19:18] <OndraSter> now I am hungry for some chips
[17:19:23] <OndraSter> atmel's chips!
[17:19:43] <specing> AVR cookies
[17:19:49] <OndraSter> mmm
[17:19:58] <OndraSter> cook them on beaky 's chip
[17:20:30] <amee2k> mmh wasn't there an NE555 cake? >_>
[17:21:10] <OndraSter> I want to give my ex-emlab teacher a tshirt with 555 schematics
[17:21:16] <OndraSter> and the opamp one
[17:21:26] <beaky> the ones dave jones have?
[17:21:30] <OndraSter> no
[17:21:33] <OndraSter> better one
[17:23:02] <vsync_> davy jones?
[17:23:08] <OndraSter> and his locker
[17:23:18] <vsync_> didn't know he was into flip flops
[17:24:34] <OndraSter> g'dammit qnet, y u disconnect me
[17:24:40] <OndraSter> err freenode
[18:04:23] <vsync_> so. any audio guys out here? Looking for a headphone amp for desktop usage
[18:04:33] <vsync_> using a fiio e07k atm but i am disappointed in the dac
[18:16:31] <hjohnson> man I hate it when the terse, precise language used by us techs/Engineers becomes a proplem for the powers that be.
[18:17:15] <hjohnson> the trouble ticketing system I maintain, it contains a lot of lingo to be precise... for one project, we have to give a summary of all the tikets to a bureaucrat who oversees the project... he wants us to go back to april and change all the tickets into a format that makes sense to him.
[18:17:24] <hjohnson> which makes it a useless resource to those of us who actually need to use it.
[18:19:47] <vsync_> i heard of this thing called automation. Apparently it's the thing in 2013!
[18:21:37] <hjohnson> also he's complaining about a lack of detail on some tickets "internet problems in <community>" that's because someone from taht community left us a voicemail saying "We have inernet problems" and nothing else, and when we lookged at it remotely there weren't any problems to see...
[18:22:49] <vsync_> may i inquire as to what sort of a job you are doing
[18:23:09] <hjohnson> vsync_: I work in a company that does satellite communications.
[18:23:11] <vsync_> I hope not another IT support *shivers*
[18:23:39] <hjohnson> one of my many jobs is to supervise/support network that supplies internet to 18 native reserves
[18:23:56] <vsync_> cool. Soon you can book easy trips for maintenance jobs on the satellites via virgin galactic
[18:24:03] <hjohnson> haha
[18:24:24] <hjohnson> I wish, but virgin is just doing sub-orbital balistic stuff
[18:24:31] <hjohnson> (I'm also a space geek... I know my rockets)
[18:25:52] <vsync_> sub-orbital? an orbit can be an inch up from the ground level!
[18:26:25] <vsync_> anyway yeah. agreed, they aren't going high enough yet. But you know. Shouldn't be too hard to extend. Maybe you can book Musk then
[18:26:29] <hjohnson> repeat after me: Orbit isn't about altitude, it's about speed.
[18:26:42] <vsync_> it is yes :)
[18:26:49] <vsync_> the speed determines it
[18:26:59] <hjohnson> vsync_: naw, the virgin galactic stuff isn't goign anywhere near fast enough for orbit...
[18:27:06] <hjohnson> orbit is about 10x harder than what they're doing.
[18:27:21] <vsync_> yeah :) well, can always book musk.
[18:27:44] <vsync_> I have been doing a fair share of rocketry. Just the water type, not pyro :[
[18:28:10] <hjohnson> heh
[18:28:34] <beaky> for voltage sensing, is lower better?
[18:28:38] <hjohnson> and the satellites I use are 1/3 the way to the moon too...
[18:28:38] <beaky> or should i aim for Vcc
[18:29:01] <vsync_> hjohnson: it's merely a matter of adding more fuel!
[18:29:23] <hjohnson> which means adding more rocket, which means adding more fuel...
[18:29:26] <hjohnson> (see the rocket equation)
[18:29:38] <vsync_> indeed it does.
[18:29:48] <hjohnson> one of the reasons why SSTO is a bit silly, imho
[18:30:15] <hjohnson> there's no resaon to bring your fuel tanks into orbit with you... that's wasted energy
[18:30:15] <vsync_> single stages are way boring anyway
[18:30:24] <vsync_> yep
[18:30:59] <vsync_> so
[18:31:07] <vsync_> apart from the armchair have you done rocketry?
[18:31:40] <hjohnson> I spent a summer working for the space agencies... does that count? :)
[18:31:45] <beaky> or should i use internal 1.1v vref
[18:31:47] <beaky> is that best?
[18:31:58] <vsync_> i guess it depends, what were you doing?
[18:32:20] <hjohnson> (Admittedly, I was just the camp engineer, fixing vehicles, generators, running polar bear patrol, working the satcoms, and gtting drunk with astronauts)
[18:33:29] <beaky> if i use 1.1vref, my error is only 1 millivolt
[18:33:36] <beaky> but if i use 5v vref, my error is 3 millivolt
[18:33:38] <beaky> right?
[18:34:00] <vsync_> what the f? scale the vref based on wtf you are measuring
[18:34:10] <vsync_> hjohnson: alrightey. well i'm afraid that doesn't count
[18:34:26] <beaky> i am measure a signal between 1 to 4v
[18:34:38] <vsync_> then you probably want 4v ref
[18:34:39] <beaky> so i am thinking of simple voltage divier to use 1.1v internal reference
[18:34:48] <vsync_> yeah sure
[18:34:56] <hjohnson> heh
[18:35:02] <vsync_> hjohnson: i'm more into the design and the aerodynamics part of it
[18:35:05] <beaky> 4 / 1024 = 4 millivolt of error
[18:35:15] <beaky> while with 1.1v vref, 1 millivolt of error
[18:35:17] <hjohnson> vsync_: I don't really care how it gets up there, I just like talking to it. :P
[18:35:50] <vsync_> going to build a slim water rocket this winter, it will be my first of that type
[18:37:36] <vsync_> I don't do pyro so hahah =D My tops have been 180 meters. Not much, but it's quite good for a water one
[18:39:00] <vsync_> as the pressures you work with start to get quite ridiculous
[18:39:00] <beaky> i am measuring 12v signal. how do i bring it down to 1.1v
[18:39:15] <beaky> I am thinking of using a 100k resistor dividor,
[18:39:29] <beaky> 500k / 100k
[18:39:37] <vsync_> dude
[18:39:37] <beaky> should bring it down to 2v
[18:39:46] <beaky> 1M / 100k
[18:39:52] <beaky> now thats 1.1v
[18:40:05] <vsync_> i thought you said you loved ohm's law?
[18:40:16] <beaky> with a lm741 buffer in between to make a nice and smooth voltage for avr
[18:40:33] <beaky> yes ohms law is the grand unified field theorem of electronics
[18:40:42] <beaky> everything can be understood as resistance
[18:41:26] <vsync_> ...
[18:41:49] <vsync_> why don't you use the "DCDC Opcode" ninja suggested?
[18:42:02] <vsync_> you can pass it through that and then feed your adc pin from that
[18:42:04] <beaky> DCDC is low impedance
[18:44:15] <vsync_> throw in resistors and whatnot
[18:44:25] <vsync_> the avr handles the rest. DCDC is awesome
[18:47:31] <hjohnson> god damned fucking shit-ass bureaucrats... first up against the wall when the revolution comes.
[18:49:00] <vsync_> is this US?
[18:49:11] <vsync_> is this me john wayne...
[18:50:28] <hjohnson> naw, .ca
[18:51:15] <vsync_> is this me, mario lemieux...
[18:58:50] <beaky> when you change fuses in avr, what are "fuses"
[18:59:07] <beaky> are they like fuses that pop in high current
[18:59:12] <beaky> or are they like rom
[18:59:21] <vsync_> http://4.bp.blogspot.com/-A90f7akVGX8/UC0ARuFm9uI/AAAAAAAAAMU/96cduoLHD5E/s1600/fuse+types.png
[19:01:24] <hjohnson> beaky: they're really a few bits of EEPROM
[19:01:29] <hjohnson> more than anythign else.
[19:02:13] <beaky> ah
[19:02:29] <vsync_> they are basically settings
[19:02:30] <vsync_> oh
[19:02:38] <vsync_> well mom pulled the plug i guess
[19:11:08] * juri_ tries to move to interrupt driven everything. Wheee...
[19:13:48] <hjohnson> heh
[19:23:53] <N1njAway> Interrupt-driven-interrupts, even? :D
[20:48:31] * hjohnson returns
[22:52:29] <N1njAway> Strippin'
[22:52:32] <N1njAway> Strippin'
[22:52:36] <N1njAway> Strippin' the night away!
[22:52:39] <N1njAway> Cables, that is.
[22:52:57] <N1njAway> At least it's only 10 this time, rather than 90.