Back
[00:16:18] <inflex> mmmm
[00:16:25] <inflex> I should start trying the 8U's one day
[07:07:01] <zoobab> hi
[07:07:01] <tobbor> Hello zoobab
[07:07:29] <zoobab> anyone using avrdude with FTDI FT2232 adaptors?
[07:08:45] <RikusW> I don't think its supported
[07:08:54] <RikusW> you might be able to add support....
[07:09:19] <RikusW> avrdude provide raw SPI commands which could be sent to the FT2232
[07:09:26] <RikusW> but you'll need to add some code
[07:10:29] <zoobab> avrdude 5.11 says it supports MPSSE
[07:11:16] <RikusW> thats the FT2232 protocol ?
[07:11:50] <zoobab> that's the engine inside FT2232
[07:16:42] <RikusW> I see there is an avrftdi.c file in the 5.11 source
[07:16:52] <zoobab> yes
[07:16:59] <zoobab> but I don't know how to use it
[07:17:13] <RikusW> static int ftype; /** is from FTDI. Use TYPE_2232C, TYPE_2232H, or TYPE_4232H*/
[07:17:27] <RikusW> you might need to edit avrdude.conf
[07:18:34] <RikusW> strcpy(pgm->type, "avrftdi");
[07:20:08] <zoobab> http://www.lura.sk/?&LP=113SK&MP=4&ML=3&PO=''
[07:20:19] <RikusW> programmer
[07:20:20] <RikusW> id = "avrftdi";
[07:20:20] <RikusW> desc = "FTDI x232x";
[07:20:20] <RikusW> type = avrftdi;
[07:20:20] <RikusW> ;
[07:20:27] <RikusW> add this to avrdude.conf
[07:20:31] <zoobab> ok
[07:20:32] <RikusW> it might work
[07:20:43] <zoobab> I am trying to flash an atmega32u4
[07:20:57] <RikusW> using ISP or jtag ?
[07:21:08] <zoobab> I don't know which pins of the JTAG it matches to
[07:21:57] <RikusW> I think avrdude supports ISP mode
[07:27:39] <zoobab> http://wiki.seabright.co.nz/wiki/AvrdudeAndFtdi.html
[07:29:12] <RikusW> thats for 5.8....
[07:29:20] <RikusW> seems 5.11 does have the code
[07:29:37] <RikusW> only avrdude.conf is missing a few lines...
[07:30:32] <RikusW> I see there is source you could try if 5.11 won't work...
[07:37:32] <RikusW> Seems its sleepy sunday time again :-P
[07:38:42] <RikusW> zoobab: another thing, you'll need to have libftdi installed when building avrdude....
[07:39:32] <zoobab> i installed it
[07:39:41] <zoobab> checking the cables now
[11:33:03] <mapee> hi
[12:12:08] <sabesto> heh, my head is not working. I am building a line follower, the main loop needs to sample 7 ADC channels, then do some calculations. any tips on how to do this?
[12:13:37] <sabesto> right now ive set it up to "free run", do i simply have to mux one and one channel in the main, then wait for a flag to be set, read the data register and set the mux to next chan?
[12:13:57] <ziph> sabesto: You're better off setting up an interrupt to do it.
[12:14:23] <sabesto> yeah, i was looking at sample code, all seems to use interrupts
[12:14:40] <ziph> sabesto: Or using an MSP430, which can sample a bitmapped set of channels and then call an interrupt only when all channels are sampled. ;)
[12:16:11] <sabesto> thats overkill for this project
[12:16:30] <ziph> Yeap, hence the wink.
[12:17:08] <sabesto> so, in free run mode, each time a conversion is done i will get an interrupt
[12:17:11] <ziph> Set up an interrupt to flip between the channels.
[12:17:25] <ziph> I wouldn't do it in free run mode with channel switching.
[12:17:38] <sabesto> let that interrupt run 7 times, then run the main
[12:17:56] <sabesto> http://www.avrfreaks.net/modules/FreaksFiles/files/383/DN_021.pdf
[12:17:57] <ziph> If you do that the ADC may be starting a conversion on the same channel again by the time your interrupt gets serviced.
[12:18:09] <Guest99775> sabesto: what I personally was doing is interrupt based
[12:18:26] <ziph> Well, your main can spin waiting for the interrupt to set a flag saying all 7 conversions are done.
[12:18:27] <sabesto> yeah, thats what i was thinking, but that code does just that
[12:18:47] <sabesto> ziph: yeah, what i meant
[12:18:59] <Guest99775> I was simply doing a single conversion, the interrupt change the channel, start a new conversion and update a global variable
[12:19:17] <Guest99775> the main code just read that global variable whenever it want, and always have the last one available
[12:19:49] <Casper> WOW
[12:19:50] <sabesto> i need a fresh set of 7 channels before the main can run one time
[12:19:56] <Casper> -NickServ- 450 failed logins since last login.
[12:19:56] <Casper> -NickServ- Last failed attempt from: MixZhare!Ratu@209.115.237.125 on Dec 04 18:06:50 2011.
[12:20:10] <sabesto> hah
[12:20:13] <Casper> that is in like... 12 hours
[12:20:22] <ziph> sabesto: You could just spin on the flag then and manually start each conversion.
[12:22:32] <sabesto> that will take longer time though
[12:22:47] <sabesto> free running is faster i think
[12:23:17] <sabesto> i will read some more, give up 1 or 2 times, then do it!
[12:29:34] <mark4> where on my drive does as4 put all the .def files?
[12:30:24] <mark4> i just did a major tidy up of rickusw's jtag ice code, modularized it etc etc. made a new project for it and added all the files and when i go to compile it cant find <m32U2.def>
[12:31:23] <mark4> erm m32u4def.inc or whatever its alled
[12:35:28] <sabesto> can i disable global interrupts from an interrupt?
[12:36:08] <mark4> i would hope so :)
[12:45:18] <mark4> #include <m32U2def.inc> <-- where is that file located? my build cannot find it at all
[12:53:47] <karlp> locate -i m32u2def.inc or <win>-f m32u2def.inc....
[12:54:54] <mark4> i have a gentoo virtualbox thats not able to locate or find . -name that file
[12:55:39] <mark4> the windows partiton is mounted as a vbox share in the vm. updatedb will have scanned it so locate SHOULD have worked if the file had existed
[12:55:47] <mark4> i want to know where its SUPPOSED to be.
[12:58:02] <allgood38> So I am developing an autonomous sailboat for a university, could someone critique my code?
[12:59:13] <Steffanx> Who knows allgood38
[13:00:07] <allgood38> Steffanx: ? I'm working with a team, I just need to make sure it isn't too confusing
[13:00:34] <allgood38> Our current codebase is a mess, so I've taken to rewriting it
[13:00:54] <allgood38> Its written for the Arduino, which is why I'm here
[13:01:38] <allgood38> www.github.com/allgood38/newmast
[13:02:17] <mark4> WHY the hell would anyone write drag and drop, cut and paste object OBFUSCATED c++ for a fucking arduini
[13:02:34] <mark4> artuino
[13:02:45] <mark4> ugh i still cant type
[13:02:48] <mark4> arduino
[13:03:01] <allgood38> What do you mean? I'm new to this and open for all criticism
[13:03:31] <mark4> your developing for a VERY minimalistic/SIMPLISTIC micro controller
[13:03:45] <mark4> yu want your project to be 800 times more bloated than it should be? use c++
[13:03:59] <mark4> ypu want the code to be orders of magnitude more complex and UNREADABLE? use c++
[13:04:32] <allgood38> So it would be better to implement this in ASM?
[13:04:48] <mark4> if your going to develop for a microcontroller and are using a training wheels device like the arduino c++ is probably the worst choice you could make
[13:04:50] <mark4> asm is simple
[13:05:02] <mark4> asm is VERY easy to understand (thats why its called a LOW level language)
[13:05:23] <mark4> asm would result in code thats 100 times smaller and probably executes faster
[13:05:35] <sabesto> ziph:
http://pastebin.com/0ZYD0p7W
[13:05:44] <sabesto> ziph: thoughts?
[13:05:45] <allgood38> Okay, but how is asm at dealing with strings, I'll need to do some parsing
[13:05:45] <mrfrenzy> if you want to be finished with any decent sice project in a realistic time frame I would recommend C
[13:06:05] <mrfrenzy> ASM is great when you want do low level stuff very fast and have limited resources
[13:06:06] <mark4> why the hell do yu want STRINGS on a autonomus boat?
[13:06:15] <mrfrenzy> for a big project like this C is the best way IMO
[13:06:23] <mrfrenzy> I agree to stay away from C++
[13:06:26] <allgood38> The wind sensor we use outputs NMEA style strings
[13:06:32] <mark4> and how asm deals with strengs is entirely dependant on how YOU code that handling
[13:06:38] <mrfrenzy> but certainly go on the forums and read the different opinions on this
[13:06:42] <mark4> so write a parser
[13:07:01] <allgood38> Okay, cool, I'll take a look at it
[13:07:05] <mrfrenzy> if you use C you just include string.h or whatever it's called
[13:07:17] <mark4> mrfrenzy, developing embedded applications in C is like opening a can... with a ROCK
[13:07:21] <mark4> -- Ron Oliver --
[13:07:33] <mrfrenzy> I disagree, as do many others
[13:07:47] <allgood38> But you get the benifits of a higher level language
[13:07:51] <mrfrenzy> but sure, if you want to get the very most out of your attiny with 2kB flash, then asm is the only way
[13:08:11] <mrfrenzy> if you have a big enough device, and time is a limited factor asm is not feasible imo
[13:08:16] <allgood38> We're actually using an ATMEGA2560
[13:08:31] <allgood38> Plenty of memory for our needs thus far
[13:08:39] <mark4> mrfrenzy, if you need to make 200 thousand gizmos a year and NEED to keep the price below $10 then you need the cheapest device you can use to solve the problem
[13:08:52] <mrfrenzy> if you want to make a million devices, and you can save $2 dollars per device by choosing a smaller micro, then it might be worth 2 months extra developing time to do it in asm
[13:08:53] <mark4> c is going to dictate a larger memory requirement and thus a more expensive device
[13:08:56] <mark4> c is NEVER the best choice
[13:09:22] <mrfrenzy> read up mark4, he is building a device for a sailboat costing tens of thousands of dollars
[13:09:26] <allgood38> Thats the thing though, we only need to make one. It doesn't need to be the best, as long as it is rock solid
[13:09:29] <mrfrenzy> doesn't matter if the micro costs $2 or $20
[13:09:32] <mark4> not in 25+ years of doing embedded development professionally have i ever seen a single project that benefited in any way shape or form from having been developed in c
[13:10:10] <allgood38> lol would your head explode if I told you I was thinking of doing the sailing logic and high level math in python on a seperate board?
[13:10:20] <mrfrenzy> as I see it, benefits of asm: cheaper microcontroller. benefits of C: a lot shorter development time
[13:10:33] <mrfrenzy> allgood38: python rocks! there actually is a python implementation for avr
[13:10:54] <mark4> actaully you r head might explode if i told you what language i think it really should be coded in but c/c++ would and SHOULD be removed from all embedded applications
[13:10:55] <mrfrenzy> with python we are talking really high level
[13:11:00] <allgood38> Really? That would be amazing since I could eleminate the second board
[13:11:03] <mrfrenzy> but actually, I would only use python on avr for a toy
[13:11:07] <mrfrenzy> it is a bit too high level
[13:11:11] <mrfrenzy> I would say stick to C
[13:11:21] <mrfrenzy> gotta go, laters
[13:11:21] <mark4> asm would be preferable. you might actualy learn something if you actually had to code it yourself instead oif just #include blah.h
[13:11:21] <allgood38> thanks
[13:11:40] <allgood38> Oh for sure, I'm going to have to learn it next year anyways
[13:12:10] <mark4> actually i wouldnt have used an arduino, i would have used an atmega32u4 or something along those lines
[13:12:22] <mark4> and done it in asm
[13:12:57] <mark4> tho asm would be my second choice i have refrained from stating my primary choice because you would ALL go screamin for mamma :P~
[13:13:06] <mark4> allgood38, when is this due by?
[13:13:41] <allgood38> I want to get it done by the spring, at least most it anyways
[13:14:00] <allgood38> So whats your primary choice?
[13:14:22] <mark4> ok well. DO get rid of c++. DO stop using object obfuscation on an embedded device, go with C because thats what everyone will understand
[13:14:30] <karlp> he's going to say forth...?
[13:14:50] <mark4> my primary choice would be FORTH. if you had a forth compiler for your device you could complete the entire project from sratch within a week if you new forth
[13:14:53] <karlp> hehe
[13:14:54] <karlp> knew it
[13:15:14] <allgood38> I've been recommended that before actually, using C, since its easier to wrap your head around the language
[13:15:17] <mark4> you could learn forth enough to complete the project within a few weeks
[13:15:40] <mark4> allgood38, actualy forth is orders of orders of orders of magnitued SIMPLER to understand than c
[13:15:41] <allgood38> FORTH? hang on, I'm off to wikipedia
[13:16:30] <mark4> you would need a forth compiler for your target or you would need to create one. if you KNEW forth you could create a minimal forth compiler over night
[13:16:35] <mark4> no bs. ive done it before
[13:17:12] <karlp> that's like saying if I KNEW russian, I could read war and peace over night.
[13:17:25] <karlp> if you don't know _any_ russian, it's going to take a hell of a lot longer
[13:17:27] <mark4> and this damned definition include file does NOT exist on my system!
[13:17:37] <karlp> write it yourself :)
[13:17:48] <karlp> don't just #include <someotherdef.inf>
[13:17:49] <karlp> remember?
[13:17:52] <mark4> karlp no war and peace would take FOREVER to read
[13:18:02] <mark4> its not MY code
[13:18:13] <mark4> and its a definitions file for the selected processor
[13:18:25] <mark4> thats NOT the same as #include <usbcdc.h>
[13:18:49] <mark4> i have ZERO def files available for include after installing as4
[13:18:51] <mark4> wtf over?
[13:19:34] <mark4> can someone tell me where on my C drive the <m32U2def.inc> is supposed to be?
[13:20:17] <allgood38> sudo find / -name "m32U2def.inc"
[13:20:43] <mark4> dont need sudo and i told yu, it does not exist
[13:20:50] <mark4> ive used locate AND find . -name
[13:21:00] <mark4> its not anywhere on my system
[13:21:05] <mark4> i want to know where it SHOULD be
[13:21:11] <allgood38> Also, FORTH looks pretty awesome, I'll take a better look at it when I have some time
[13:21:23] <allgood38> I like the fact that it uses Reverse Polish Notation
[13:21:43] <mark4> if that does not scare you away from it then there is hope
[13:22:09] <allgood38> lol, well I'm Comp Eng, second year, so its better I know these things
[13:23:01] <mark4> btw. one week of serious study of forth will teach youmore about forth AND about good programming practices than 284659246598264 years of hard core C will ever teach you about c
[13:23:43] <allgood38> well thats the key right? Good programming practise, everything else is syntax
[13:23:55] <mark4> forth has user defined syntax :)
[13:24:01] <mark4> and no data types
[13:24:08] <mark4> types are a crutch for poor programmers
[13:24:11] <mark4> -- chuck moore -
[13:25:29] <allgood38> Hardcore. I've gotta go, but thanks for all the ideas
[13:38:47] <mark4> can someone PLEASE tell me where the FUCK my assembler include files are supposed to be? i have NO def.inc files anywhere on my entire fucking sytem
[13:38:51] <mark4> and its pissing me the fuck off
[13:43:32] <soul-d> locate def.inc
[13:43:32] <soul-d> /usr/share/avra-1.2.3/1200def.inc
[13:43:32] <karlp> it's a sign from god.
[13:43:37] <mark4> i
[13:43:37] <mark4> cant
[13:43:38] <mark4> fucking
[13:43:39] <mark4> locate
[13:43:40] <mark4> a
[13:43:41] <mark4> file
[13:43:41] <mark4> thats
[13:43:42] <mark4> not
[13:43:44] <mark4> fucking
[13:43:46] <mark4> there
[13:43:47] <soul-d> yum info avra*
[13:43:52] <karlp> he says you shouldn't be using asm
[13:44:01] <soul-d> Summary : Atmel AVR assembler
[13:44:01] <soul-d> URL :
http://avra.sourceforge.net/
[13:44:16] <mark4> all i want to know is
[13:44:18] <Kevin`> mark4: apt-file search def.inc
[13:44:19] <Steffanx> I LOVE the ignore function of my client
[13:44:22] <mark4> where is it on Y O U R system
[13:44:31] <karlp> he wasn't telling you to use locate, he was saying where it was on his system
[13:44:35] <soul-d> usr/share
[13:44:36] <karlp> under /usr/share/avra...
[13:44:41] <mark4> ugh
[13:44:44] <mark4> yea it doesnt help me
[13:44:51] <mark4> thats in linux. im using as4 in windows
[13:44:55] <Steffanx> Aaah
[13:44:56] <mark4> and i have ZERO definition files
[13:44:58] <soul-d> i pasted
[13:45:01] <mark4> none what so fucking ever
[13:45:03] <soul-d> the priogram
[13:45:11] <Steffanx> Then it's in C:/Program Files/Atmel/AVRTools or whatever :P
[13:45:18] <Steffanx> It's really somewhere there :P
[13:45:31] <mark4> yea i have that directory. there are NO def.inc files anhywhere in that tree
[13:45:32] <mark4> none
[13:45:32] <mark4> zero
[13:45:34] <soul-d> he tried searching his pc so probaably missing stuff
[13:45:35] <mark4> zip
[13:45:36] <mark4> zilch
[13:45:45] <karlp> have you tried unplugging it, and plugging it back in again?
[13:46:10] <mark4> u mean my computer or my bad attitude :(
[13:46:14] <mark4> sorry
[13:46:17] <karlp> if you say you don't have the files, how is knowing wher eit is on someone else's computer ging to help anyway?
[13:46:19] <mark4> im just getting a little frustrated
[13:46:29] <karlp> you seem to know how to use the file search methods available to you, but you don't have them
[13:46:33] <soul-d> mark
[13:46:37] <soul-d> i gave you the link
[13:46:40] <soul-d> to the program
[13:46:49] <karlp> try reinstalling it, but knowing where they are on another computer isn't going to help much if you really don't hve them locally
[13:46:49] <mark4> soul-d erm sorry my frothing at the mouth i missed it
[13:46:53] <mark4> scrollback
[13:47:09] <mark4> im not using avra
[13:47:21] <mark4> im using avr stupidio 4
[13:47:26] <soul-d> did you read
[13:47:28] <soul-d> what it uses
[13:48:34] <mark4> i dont see anything in avra that will help me no
[13:48:53] <mark4> i dont want a kludge fix. i want to know why my as4 install is so fucking broken
[13:48:54] <soul-d> maybe you should read on what avr studio is based on
[13:49:06] <soul-d> it uses gcc
[13:49:41] <Steffanx> Maybe you should take a break mark4
[13:49:44] <mark4> i know. so does the pic development system. everyone is so fucking lazy or stupid they cant write their own compilers
[13:49:50] <mark4> Steffanx, probably :(
[13:50:02] <Steffanx> Or is 'gordon ramsay' your second name?
[13:50:28] <mark4> i installed as4 again recently. fresh download from atmel. why is my install broken?
[13:51:12] <soul-d> good reason to switch to linux
[13:51:23] <mark4> linux is my primary os
[13:51:29] <soul-d> then why not use it
[13:51:32] <soul-d> for avr ?
[13:51:34] <bram_> mark4: on linux you need avr-gcc
[13:51:55] <bram_> mark4: on windows you need winavr (which installs avrgcc)
[13:51:56] <mark4> because i dont want 3434875683458367465734 different versions of gcc corrupting my CLEAN linux install
[13:52:03] <mark4> i have winavr
[13:52:17] <soul-d> er how would avr make that a mess ?
[13:52:21] <mark4> winavr is installed. whats not installed anywhere is *def.inc
[13:52:27] <soul-d> also i just use a rpm based distribution
[13:52:34] <Steffanx> bram_ .. afaik the assembler is included in AS4
[13:52:40] <Steffanx> C:\Program Files\Atmel\AVR Tools\AvrAssembler2\Appnotes that's where the file should be
[13:52:51] <mark4> soul-d, i dont use rpm or deb based distros. i use gentoo which has superior dependancy handling
[13:53:24] <mark4> nope wtf they ARE there
[13:53:32] <mark4> err why is that not in my include path then?
[13:53:36] <soul-d> so then it should be easy to do avr-gcc on your linux
[13:53:37] <Steffanx> Just try a re-install?
[13:53:43] <karlp> so, you completely failed to use <win>-f like I told you in the beginning?
[13:53:44] <mark4> and why can linux not find . -name or locae ANY of them?
[13:53:58] <soul-d> did you updatedb afther install
[13:54:15] <mark4> i didn an updatedb right before i did locate
[13:54:23] <mark4> that failed so i did find . -name
[13:54:41] <mark4> why is this directory not in my include path blargh
[13:54:51] <soul-d> never used avr with windows
[13:54:55] <soul-d> thats liek trying pic
[13:54:57] <soul-d> on linux
[13:55:02] <Steffanx> No way soul-d
[13:55:06] <Steffanx> Thats bullshit
[13:55:08] <soul-d> well maybe
[13:55:09] <soul-d> almost
[13:55:14] <mrfrenzy> avrstudio is quite nice
[13:55:15] <Steffanx> No way
[13:55:45] <bram_> mark4: On my windows, the .inc files are in: program\ files\atmel\avr tools\avrassembler\appnotes
[13:55:47] <Steffanx> Install AS4/5 and WinAVR that's all you need on windows
[13:56:28] <mark4> bram ya thay are here, my find . -name skills are lacking i guess. and locate doesnt seem to updatedb my /home/c virtualbox windows drive share
[13:56:39] <mark4> as5 is an abomination
[13:56:45] <Steffanx> You tried to reinstall as4?
[13:57:00] <mark4> its a fresh install. im going to just add the include path manually for now
[13:57:06] <mark4> but ya i need to reinstall
[14:25:15] <mark4> ok NOW all i need is rickusw to show up! }:)
[16:01:45] <Steffanx> mark4 you mean, rikusw :P
[16:02:14] <mark4> thats what i said! }:)
[16:02:20] <mark4> no c bleh
[16:02:23] <Steffanx> No you said rickusw :P
[16:02:26] <mark4> i know lol
[16:02:26] <Steffanx> He probably won't show up anymore today (today for you)
[16:02:29] <mark4> hush u
[16:02:42] <mark4> yea i know he is in SA
[16:03:01] <mark4> do you have his jtagice project?
[16:03:09] <Steffanx> No i don't
[16:03:44] <mark4> k i made mods to it, i think i reduced it by a small amount but the sources are not all in one file now
[16:03:46] <mark4> modularized it
[16:03:53] <mark4> way more maintainable now lol
[16:04:01] <Steffanx> Hehe
[16:04:22] <mark4> was bored what can i say.
[16:04:38] <mark4> but his stupid label names choices conflicts with my 32u4 device lol
[16:07:40] <mark4> whats the difference betwen an atmega32L and an atmega32U4?
[16:07:53] <mark4> the U implies usb?
[16:07:55] <Steffanx> Yes
[16:08:38] <Steffanx> Xu2, Xu4 etc. are all AVR's with usb
[16:25:19] <amee2k> up to what voltage roughly can i count on solder stop being insulating?
[16:25:37] <keenerd> Depends on thickness.
[16:26:05] <amee2k> if i can't find it specified by the boardhouse?
[16:26:18] <amee2k> what i'd need is ~30V tops
[16:26:47] <keenerd> You'll probably be fine.
[16:28:35] <amee2k> hmm maybe let me ask the other way around. is there an easy (better) way to insulate a board area of around 20x5cm that is bolted flat against a heatsink? the planes covering the board area in question are not electrically neutral
[16:29:19] <amee2k> i think the panes will get a decent heatspreading effect, but thermal resistance should be reasonable (say <10C/W)
[16:30:08] <keenerd> There are insulated bushings/pads for connecting heatsinks.
[16:30:47] <amee2k> i know, but sil-pads are ass expensive in small sizes allready and all the larger sheets i've seen for sale were kinda ridiculous
[16:31:05] <amee2k> i don't think i've ever seen mica pads in that size
[16:31:39] <keenerd> Mica comes in that size.
[16:31:45] <karlp> are you trying to make sure that a heatsink is thermally connected, but not electrically connected?
[16:31:55] <amee2k> karlp: exactly
[16:32:28] <keenerd> Mica is probably not at all what you want though.
[16:32:45] <amee2k> why not?
[16:33:51] <Tom_itx> isn't mica a thermal insulator?
[16:36:33] <amee2k> its a pretty good electrical insulator so a relatively thin pad will do. thermally i think it still beats rubber pads
[16:37:49] <amee2k> right now my plan is to just have the back covered with solder stop, then put thermal grease on and bolt it straight to the heatsink
[16:39:39] <karlp> and you're relying on the solder mask to stop the heatsink from being electrically connected to the signal?
[16:39:50] <amee2k> yeah
[16:40:10] <amee2k> 30V isn't much, but i'm still having my doubt. hence the question
[16:40:55] <karlp> why can't it just be electricall connected?
[16:41:14] <karlp> leave off the solder mask, and you'll get pretty excellent thermal conductivity
[16:41:35] <amee2k> because there are four planes on the back of the heatsink that are at different electrical potential
[16:42:05] <amee2k> and i can't see a way to have four independent heatsinks on there
[16:49:16] <karlp> You're putting a 5cm by 20cm heatsink on, rather than 4 separate ones?
[16:49:23] <karlp> this is just to put ona single heatsink?
[16:49:44] <amee2k> no
[16:49:46] <amee2k> well, yes
[16:49:53] <karlp> not because you're worried about anyone touching it or anything?
[16:50:19] <amee2k> four heatsinks at different electrical potential are pretty difficult to fit, compared to one electrically neutral heatsink
[16:50:59] <amee2k> 30V should be mostly harmless to touch, only shorting would result in sparks and a blown fuse
[16:51:11] <amee2k> plus downtime to fix it, of course
[16:51:58] <amee2k> so yeah, thats probably a factor too besides the mounting issues
[16:52:44] <karlp> it sounds like you're trying to fit a square peg in a round hole. in my completely unprofessional opinion.
[16:52:51] <karlp> if things I touch need a heatsink, I've done something wrong.
[16:53:01] <karlp> so not really my field :)
[16:53:28] <amee2k> i never said i like the fact that the board has thermal issues
[16:54:20] <amee2k> in fact, someone deserves to lose his fertility over this, if it were my decision
[16:56:33] <karlp> fair enough :)
[16:57:11] <amee2k> also, the touching the heatsinks part, if it were to happen, would not be by design
[16:57:23] <amee2k> so that much is still sane, really
[16:57:31] <karlp> talk to these guys:
http://www.chomerics.com/products/therm_chotherm.htm
[16:58:32] <karlp> says they make diecut to any shape too..
[16:58:47] <amee2k> in low quantities too?
[16:59:12] <amee2k> or do they have a reseller for that, preferably in europe
[16:59:27] <karlp> ... ask them, not me :)
[16:59:39] <karlp> like I said, I don't do much thermal stuff
[16:59:43] <amee2k> you recommended them, chances were you could know :)
[17:00:43] <karlp> I found them here:
[17:00:52] <karlp> http://lmgtfy.com/?q=thermally+conductive+insulators
[17:01:51] <karlp> sil-pads were number two, but it sounds like you already know them :)
[17:02:32] <amee2k> i'm only excluding them due to the price, really. otherwise they would be fine
[17:02:39] <karlp> but again, 5x20cm is a pretty enormous piece
[17:02:45] <amee2k> exactly
[17:02:54] <karlp> do you really need to cover teh whole thing?
[17:03:16] <karlp> how much power are you trying to dissipate here?
[17:03:23] <amee2k> i could do it with four pieces of 5x5cm each, but smaller than that and alignment is going to get tricky
[17:04:34] <karlp> are you trying to bolt this directly to four plane parts of the pcb?
[17:04:35] <amee2k> only about 50W, but the problem is that the design is kinda shit thermally so i need to get rid of it effectively and constantly
[17:04:38] <karlp> not onto the back of an IC?
[17:05:03] <karlp> "only about 50W"
[17:05:06] <amee2k> i can't compensate for shitty thermal resistance with a larger heatsink because it'll just heat soak and overheat
[17:05:48] <amee2k> 50W isn't exactly much considering the area i have available to get rid of it
[17:07:02] <amee2k> thats half a watt per square centimeter
[17:08:14] <amee2k> thats why thermal resistance isn't a top priority as long as its reasonable
[17:09:03] <karlp> so, instead of 4x5x5, do 4x4x4,
[17:09:07] <karlp> no alignment issues,
[17:09:16] <karlp> no insulation issues
[17:09:22] <amee2k> ?
[17:09:22] <karlp> probably enough transfer?
[17:09:28] <amee2k> how does making it smaller help?
[17:10:09] <karlp> we're still only talking about 1.4W/cm2
[17:10:14] <karlp> is that completely unreasonable?
[17:10:40] <amee2k> no, but i don't get how making it smaller helps when i allready said 4x 5x5 would be acceptable
[17:11:02] <karlp> I thought you said 4x5x5 was _not_ acceptable, as it would be too difficult to align them all without touching
[17:11:08] <amee2k> its just that i can't put in 50x TO220 sized thermal pads because that would be bit of a bish to assemble
[17:11:53] <amee2k> karlp: note the "but smaller than that" in my line ;)
[17:24:22] <amee2k> i'm tempted to just try OHP transparencies >_<
[17:25:44] <karlp> I seen them melt in printers.
[17:27:50] <amee2k> the fuser unit in a laser printer easily reaches 200+C
[17:28:45] <amee2k> i've successfully used OHP transparencies on transistors and voltage regulators before
[17:29:31] <amee2k> their thermal resistance isn't too hot, but they're dead cheap and apparently good dielectric strength
[17:31:18] <amee2k> specs on celulose acetate are like 10kV/mm, and i've successfully checked them with an insulation tester i borrowed at 6kV
[18:16:05] <LikWidChz> Dense!
[18:20:29] <mark4> also flamable?
[18:31:57] <amee2k> dense?
[19:48:54] <mark4> why the hell did atmel make them mounting hoiles in the avr dragon SOOOOOOOOO freeking small
[19:49:00] <mark4> there isnt a screw in existence that will go through that fucking hole
[19:49:09] <mark4> how the hell m i supposed to mount standoffs?
[19:51:11] <Casper> how small are they?
[19:51:44] <mark4> they are like 2mm diameter or maybe VERY slightly larger
[19:56:40] <Casper> yeah that's small
[19:56:48] <Casper> I wonder if #4 screw would fit... think not
[19:58:30] <Casper> yeah #1 screw would fit... that's not an hardware store size
[19:58:38] <Casper> smallest I gound was #4
[19:59:37] <Tom_itx> 2-56
[19:59:38] <Tom_itx> is
[20:00:58] <Tom_itx> feel special if yours has mounting holes
[20:01:01] <Tom_itx> mine has none
[20:02:48] <mark4> lol
[20:13:55] -wolfe.freenode.net:#avr- [freenode-info] channel trolls and no channel staff around to help? please check with freenode support:
http://freenode.net/faq.shtml#gettinghelp
[20:16:16] <mark4> http://www.pololu.com/picture/view/0J1767 those would work
[20:16:38] <mark4> "this product has been discontinued"
[20:16:53] <Tom_itx> you can get those numerous places
[20:17:00] <mark4> yup
[20:17:01] <Tom_itx> try jameco for one
[21:38:59] <k-man> anyone seen any kind of keyless entry projects around the interwebs that are actuall neat and tidy and mission critical ready?
[21:39:49] <CMiYC> See, I taught you how to ask a better question.
[21:39:53] <CMiYC> I added to the conversation last time
[21:41:03] <Casper> define mission critical
[21:50:56] <mark4> i would say mission critical would mean not letting unauthoried people in and allowing authorized people entry lol
[21:51:19] <mark4> and not allowing people at a distance to read any NFC signals
[21:51:55] <CMiYC> I never considered getting into my house mission critical
[21:52:14] <CMiYC> Well, except this one time I ate at a shady Mexican restaurant.
[21:52:57] <Casper> I consider mission critical to always work no matter what
[21:53:15] <Casper> safety is not the primary function fo mission critical
[21:54:53] <CMiYC> based on that, mission critical means every read works 100% of the time?
[21:55:01] <mark4> well gaining access when authorized and not gaining it when not authorized sounds kind of critical to me :)
[21:55:30] <keenerd> Casper: Dunno about you, but there are redundant power supplies on *my* garage door opener.
[21:55:31] <mark4> if security is not an issue then i have a FREE keyless entry system for you
[21:55:35] <mark4> leave the door unlocked
[21:56:08] <rue_house> FREE?
[21:56:13] <rue_house> is it avr based?
[21:56:30] <mark4> its NOTHING based. just leave door unlocked, no key needed
[21:56:44] <mark4> but if security IS an issue then ... ..
[21:57:10] <keenerd> I like mark4's idea. It is mission-critial reliable too, unless your spouse locks the door by accident.
[21:57:26] <keenerd> Better take off the lock to make it more reliable.
[21:57:36] <mark4> :)
[21:57:41] <mark4> that wouldnt fly at the whitehouse
[21:57:57] <mark4> and i used to work for the company that did the access control and alarm monitoring for the whitehouse :)
[21:58:18] <mark4> and the smithsonian institute, and the hoover building and every US Marine base in the country
[21:58:26] <mark4> oh and bill gates front door lol
[21:58:54] <Casper> k-man: I didn't checked for those, but I know the london hack space have one of such system
[21:59:07] <Casper> which even spam who enter and leave the building in their irc channel
[21:59:56] <k-man> CMiYC, piss off
[22:00:07] <mark4> ?
[22:00:22] <CMiYC> He's mad because he asked the question before, I told him to use google
[22:01:12] <CMiYC> But in the 2nd round, we learned he wanted "mission critical" and "tidy" (which were not used in the first question.)
[22:02:12] <Casper> keenerd: I wish I had a garage door opener in service
[22:02:23] <Casper> that would mean I would have a garage, which I do not
[22:03:41] <keenerd> Well I made it up too. Though it would be fun to add a second set of wires to everything in the building, for Apollo level of redundency.
[22:05:02] <keenerd> (Did they do that on the Shuttle, too? Not that it matters now.)
[22:19:41] <inflex> just make sure you route them differently
[22:20:01] <inflex> besides, modern planes have triple redundancy now afaik