#avr | Logs for 2012-04-26

Back
[02:38:27] <inflex> not I
[02:39:39] <mitsakos> hello, does anybody use avrispmkII with avrdude in mac os?
[02:42:20] <ziph> Sane people use AVR Studio in Windows under Parallels. ;)
[04:20:05] <mitsakos> does anybody use avrispmk2 on mac os?
[04:45:04] <grummund> i think RikusW made a cross-platform GUI that will drive various programmers
[06:33:23] <Tom_itx> mitsakos i'm pretty sure mine works on osx
[06:35:30] <Steffanx> me too
[06:56:43] <inflex> hi folks
[06:57:00] <Steffanx> Lo Mr. inflex
[06:58:44] <ziph> inflex: lo
[07:04:20] * inflex needs to get super busy tomorrow and the w/e to finish this order of BAC units
[07:04:25] <inflex> soooooo not wanting to do it
[07:05:59] <Steffanx> Then, don't do it :)
[07:07:15] <inflex> Steffanx: a prepaid invoice of $800 says I should
[07:07:55] <Steffanx> Hmm, yeah you have no choice this time :)
[07:09:09] <inflex> I don't mind so much the SMD stuff.... but dang, the wirelooming, testing, buzzer & cap fitting... hotgluing, heatshrinking and finally packaging are a real pita
[07:11:12] <inflex> wonder if it'd just be easier to get all his wires precrimped with knurls
[07:29:17] <Roklobsta> hmmm, seen this? http://ethersex.de/index.php/Main_Page
[07:52:52] <mitsakos> Steffanx so why i have this huge delay?
[07:53:21] <mitsakos> i have also made a post here if you want to see it http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=120287
[07:53:29] <Steffanx> Sorry, i missed half of the conversation mitsakos :)
[07:58:56] <Steffanx> This is about a real avrispmkii mitsakos
[07:59:25] <Steffanx> I remember I had some 'trouble' with Tom_itx 's clone on os x, but that was more a delay of ~3-5 seconds
[07:59:37] <Steffanx> (haven't used it for a while, so i'm not 100% sure)
[08:02:06] <Steffanx> You could try to see if anything shows up in the USB prober, but you need the debug version of the usb kernel module for that (to get anything useful)
[08:02:20] <Steffanx> Or maybe something shows up in 'dmesg', mitsakos ?
[08:03:47] <Steffanx> You also tried some more v's in the -v parameter for avrdude mitsakos ?
[08:20:46] <mitsakos> more v's what does?
[08:24:00] <mitsakos> Steffanx here is what dmseg shows com_apple_driver_AppleUSBCardReaderUMC:: Stop::Controller Reset
[08:24:00] <mitsakos> 0 0 AppleUSBCDC: start - initDevice failed
[08:24:00] <mitsakos> USBMSC Identifier (non-unique): 000000009833 0x5ac 0x8403 0x9833
[08:24:00] <mitsakos> USBF: 59003.157 AppleUSBOHCI[0xffffff800b880000]::Found a transaction past the completion deadline on bus 0x6, timing out! (Addr: 5, EP: 2)
[08:24:00] <mitsakos> USBF: 59005.157 AppleUSBOHCI[0xffffff800b880000]::Found a transaction past the completion deadline on bus 0x6, timing out! (Addr: 5, EP: 2)
[08:31:23] <Steffanx> With more v's avrdude shows more debug info mitsakos
[08:37:08] <mitsakos> do you understand anything from dmseg?
[08:48:18] <Steffanx> I tried the clone from Tom_itx again, without a target. The delays are still there, but are around ~3-5 second. And i get the same error as you mitsakos :)
[08:48:21] <Steffanx> *:(
[08:48:26] <Steffanx> No idea why
[10:03:22] <Xata> sup
[10:03:48] <Steffanx> Tha sky bro
[11:38:03] <Xata> dear anyone. got schematics for low-pass filter with Q-factor? pleaaaase
[11:40:35] <alexh> huh?
[11:40:55] <alexh> q factor is a property of resonant, bandpass, etc circuits
[11:42:09] <alexh> not to mention that there are quite a few different low-pass filter "architectures"/circuits
[11:42:32] <alexh> and further, the actual component values, as well as the architecture choice itself depends on your actual requirements
[11:44:57] <karlp> R in series, C to ground will give you low pass right?
[11:45:09] <karlp> doesn't get much simpler.
[11:45:34] <alexh> will give you a first order passive lowpass filter, yes
[11:48:38] <Steffanx> A low pass filter is a low pass filter :)
[12:12:57] <ziph> This channel gets some weird questions.
[12:14:28] <Steffanx> You want more weird questions ziph ?
[12:14:58] <ziph> I want to build an AVR programmer using a toaster, and I only have $1.42 for my electronics budget for this year. How do it do that?
[12:15:48] <ziph> (Despite the fact that I clearly own a computer worth around $1000 at minimum and obviously pay $50-100 a month for Internet)
[12:17:46] <karlp> we're on teh internet! of course there's weird questions!
[12:17:58] <ziph> I don't mind weird.
[12:18:05] <ziph> Most of them are well into the retarded.
[12:18:08] <ziph> And they're all the same.
[12:18:15] <ziph> (See above)
[12:18:28] <Xata> ziph: grov programmer
[12:18:32] <Xata> *gromov
[12:18:51] <Xata> alexh: but i know there is low-pass with resonance
[12:18:59] <Xata> moog filter for example
[12:20:16] <ziph> You might want to find an audiophile forum, I do fairly advanced filter design and I still have no idea what you're talking about.
[12:23:19] <Xata> ziph: for example http://www.youtube.com/watch?v=bkHZjO2Eh-A
[12:24:32] <Steffanx> That's more than just a low pass filter ..
[12:24:44] <ziph> Yeah, that doesn't sound linear. ;)
[12:25:09] <Xata> damn. i will even fraw what i want. second, please
[12:28:28] <Xata> Steffanx: ziph take a look here. http://img232.imageshack.us/img232/3572/etgwgt.png
[12:28:30] <karlp> Xata: if you can draw what you want, clearly you don't need any more from us then :)
[12:28:58] <ziph> Xata: That's called gain peaking.
[12:29:19] <ziph> You can't do it with a single pole filter.
[12:29:35] <Steffanx> Go fancy go digital :P
[12:29:47] <ziph> Digital is terrible for audio.
[12:29:55] <ziph> Analog is the only way to go.
[12:30:00] <Xata> ziph: in pc music software it is called resonance <_< how do i do this in circuity??
[12:30:04] <Steffanx> He didn't say it was for audio, did he?
[12:30:16] <karlp> ziph: moah bits!
[12:30:25] <Xata> Steffanx: it is for audio. xata = guy who makes synth on atmega32
[12:30:30] <ziph> You can tell it is for audio because while they use real words it doesn't make sense. ;)
[12:30:39] * karlp laughs. nice one ziph
[12:31:23] <ziph> Xata: A second or higher order filter with one of the poles pushed around.
[12:32:09] <ziph> Xata: e.g. same as a normal low pass filter but with different component values.
[12:32:50] <Xata> ziph: what means pushed around? sorry, i have some huge problems with engrish sometimes
[12:35:10] <ziph> Xata: Have you got spice? You could do a Sallen-Key filter in something like LTSpice and play with the values.
[12:35:47] <Xata> ziph: proteus will do? or spice is something more advanced?
[12:36:06] <ziph> Any circuit simulator that can do AC sweeps.
[12:37:40] <Xata> ziph: look what i found. this will do?
[12:38:43] <Xata> ziph: http://sim.okawa-denshi.jp/en/OPseikiLowkeisan.htm
[12:39:10] <ziph> Yeah, that'll work.
[12:39:21] <ziph> That equation is all you really need.
[12:41:17] <Xata> ziph: thanks!
[12:41:43] <exDM69> abcminiuser: yo! I tried debugging my LUFA problem with svn trunk. The problem still exists and I'm getting dmesg messages about usb resets
[12:42:05] <exDM69> abcminiuser: I sent you a wireshark trace and some details by e-mail
[12:42:41] <abcminiuser> exDM69, excellent, I was trying to match up a name to an email :P
[12:43:02] <abcminiuser> I've been goddam tired all week after work, I now have a very full inbox and a very unhappy concience
[12:43:34] <vectory> :X
[12:43:57] <vectory> ment to ask about your bluetooth stack, i rather wont now
[12:44:05] <Steffanx> Time to work less hard abcminiuser
[12:44:25] <vectory> heard of people who just delete all mail after a long vacation :O
[12:44:40] <Steffanx> And old teacher of me did ..
[12:44:58] <abcminiuser> Steffanx, time to automate responses
[12:44:58] <Steffanx> "Oh, I've too many email" *click* "Problem solved"
[12:45:10] <abcminiuser> Steffanx, autoresponder "Could not reproduce."
[12:45:10] <Steffanx> *emails
[12:45:27] <abcminiuser> vectory, I always answer, even if I'm super late :(
[12:45:38] <abcminiuser> Which is more and more likely these days, like 2 weeks late :S
[12:45:52] <Steffanx> People should use the mailing list anyway :P
[12:47:02] <ziph> abcminiuser: Open up your e-mail so that anyone can reply. ;)
[12:47:38] <RikusW> or redirect it to #avr :-P
[12:47:48] <Steffanx> No thanks
[12:47:59] <Steffanx> Oh, yes .. that means more questions for ziph
[12:48:10] <ziph> I bags replying to his girlfriend's e-mails.
[12:48:31] <Steffanx> Doesn't she live next to him nowadays?
[12:48:50] <ziph> Good point.
[12:48:59] <ziph> I bags replying to he defacto wife's e-mails.
[12:49:45] <Steffanx> Nice marriage when your "defacto wife" sends you emails :)
[12:50:00] <ziph> My non-defacto wife sends me e-mails.
[12:50:13] <abcminiuser> ziph, heh
[12:50:28] <Steffanx> Hmm, the means you work to much/hard ziph
[12:50:31] <ziph> Which is even funnier given that I work 100% from home.
[12:50:39] <Steffanx> Hmpf
[12:50:43] <exDM69> abcminiuser: I'd need to know which svn revision match which LUFA releases, how can I do that? I havn't tagged the revisions
[12:51:56] <Steffanx> I understand why, it's because you work at night ziph
[12:52:23] <Steffanx> *know
[12:52:32] <abcminiuser> I'm on AVR support the next few weeks, so just email me there to guarantee a response within 72 hours
[12:52:32] <abcminiuser> Or your money back
[12:52:49] <ziph> Xata: I just found my worksheet with my Sallen-Key derivations, if you want to tune the peaking keep C1 = 1/C2 but increase or decrease C2.
[12:53:14] <abcminiuser> exDM69, if it's an old release grep the logs for "Commit for the" if it's new I've made tags
[12:56:50] <exDM69> abcminiuser: ok, thanks
[12:57:04] <exDM69> of course there's timestamps as a fallback
[12:58:31] <Xata> ziph: oh. thanks for advice.
[13:11:10] <exDM69> abcminiuser: git bisect started, roughly 6 steps to go
[13:24:07] <vectory> why is it, that people often type "and" instead of "an"? btw, found a typo on fourwalledcubicle.com :D
[13:25:40] <norbi> hi people, is there a guie of using pwm with avr, i really forgot how to do it
[13:25:44] <abcminiuser> vectory, where? and muscle memory
[13:25:47] <norbi> im using an atmega64
[13:25:49] <abcminiuser> exDM69, let me know if you find anything :P
[13:26:12] <vectory> yeah, i think i know what you mean
[13:26:18] <vectory> abcminiuser: http://fourwalledcubicle.com/ExplorerBot.php
[13:26:29] <norbi> i know that i set the timer counter to count, then compare the ocr0
[13:27:45] <norbi> but there is just only one duty cycle value register that is used to compare? we dont need more?
[13:28:07] <norbi> if i want to use 4 motors for pwm, ocr0 register is enough to use it for compare?
[13:30:22] <vectory> norbi: the datasheet should tell you. then there is google
[13:30:53] <abcminiuser> vectory, now you see it...
[13:30:58] <abcminiuser> ...now you don't
[13:31:05] <vectory> hehe
[13:31:16] <vectory> are you a wizard?
[13:34:32] <vectory> norbi: if your german i would recommend the mikrocontroller.net tuts
[13:36:45] <norbi> vectory: no im not german, im reading the datasheet, it says there is 2*8 bit pwm channels and six 1-16 bit pwm channels
[13:37:01] <vectory> ok
[13:37:01] <norbi> ocr0 is the only register that holds the compare value?
[13:37:26] <norbi> or each pwm channel has its own ocr?
[13:37:30] <vectory> im not familiar with that chip, none to be exact ;)
[13:38:20] <vectory> look for ocr in the register summary
[13:38:36] <vectory> if you cant find more than one, thats it
[13:39:17] <vectory> as in ocr0, ocr1 etc
[13:39:46] <norbi> ah
[13:41:30] <norbi> do i need to install the avr toolchain first or avr studio?
[13:41:37] <norbi> i completely forgot these things
[13:42:31] <norbi> winavr is in the toolchain?
[13:42:47] <exDM69> abcminiuser_: damnit, I chose the wrong starting points
[13:43:31] <exDM69> gotta do it again
[13:46:48] <asteve> bayern!!!
[13:46:55] <asteve> take that rinaldo!
[13:49:45] <Steffanx> ronaldo :)
[13:50:22] <asteve> that's right, the fliers have rinaldo
[13:50:51] <RikusW> norbi: yes
[13:54:05] <ferdnaO> i have a 74hc595 connected to a resistor to a 2n2222 base
[13:54:14] <ferdnaO> what would my resistor had to be?
[13:55:23] <RikusW> 4700 Ohms should work
[13:55:43] <ferdnaO> i have a 1K
[13:55:52] <RikusW> should be ok
[13:56:30] <ferdnaO> still my leds wont turn on =(
[13:56:32] <RikusW> at 5V it will put 4.3mA through the base
[13:56:55] <ferdnaO> hmmm no, its getting power from external supply
[13:57:04] <RikusW> what resistor did you use with the led ?
[13:57:11] <RikusW> and what voltage ?
[13:59:19] <ferdnaO> red,black, brown,gold = 200 ohms
[13:59:23] <RikusW> you do know its pinout is EBC right ?
[13:59:40] <Xata> fock. i just don't get this stuff about filters.
[13:59:46] <ferdnaO> 4 x AA batteries
[14:00:08] <RikusW> ferdnaO: so 15mA at 5V through the led
[14:00:18] <RikusW> ferdnaO: thats 6V....
[14:00:27] <RikusW> maybe a little too high
[14:01:36] <Xata> got that filterpro thing from TI. Q not less than 11 thay said. ok. but how do i move it? and if it is already 11 in output, why i can not see it on output diagram?
[14:02:15] <Xata> chebyshev, 4 op-amps, and still linear boring lpf
[14:04:01] <ferdnaO> RikusW, thanks for the help
[14:04:04] <ferdnaO> computer crashed
[14:04:22] <ferdnaO> the 4 pack double AA battery is not enough to power it
[14:04:43] <Xata> you mean pc?
[14:04:56] <ferdnaO> ya
[14:05:03] <RikusW> 4 is too much, it is 6V....
[14:05:22] <RikusW> 3 would be better, 4.5v
[14:05:39] <Xata> RikusW: oh you :3
[14:05:41] <ferdnaO> RikusW, getting 5v from the arduino
[14:06:07] <Xata> oh, arduino
[14:06:22] <Xata> yes. 5v. to power the pc. yeah.
[14:06:45] <RikusW> whats the batteries for then ?
[14:07:02] <Xata> or i got something misunderstood?
[14:11:35] <ferdnaO> RikusW, battery pack is for the leds
[14:12:05] <RikusW> why not use usb power ?
[14:12:13] <RikusW> how many leds btw ?
[14:12:50] <ferdnaO> hmmm leds wont turn with battery pack =/
[14:12:55] <ferdnaO> RikusW, 8x8
[14:13:21] <RikusW> thats a lot...
[14:13:25] <ferdnaO> i know
[14:13:26] <ferdnaO> =)
[14:14:23] <RikusW> if you only turn on one column at a time it will take 10mAx8 =80 which can be taken from usb
[14:16:06] <ferdnaO> RikusW, http://arduino.cc/en/Tutorial/ShftOut12
[14:17:15] <ferdnaO> that is supposedly doing one row at a time...
[14:22:33] <RikusW> I'd suggest checking your led circuit seperately from the avr first...
[14:23:38] <Kevin`> you can normally drive those shift registers quite fast btw. for example the full speed of the hardware spi interface (clk/2)
[14:23:55] <Kevin`> and using that lets you do the entire thing with one or two cpu instructions
[14:25:30] <Essobi> Soooo, I've got a motherboard with a TTL level GPIO that has an i2c driver in kernel land for it. :D
[14:26:12] <Essobi> wonder how hard it'd be to slave some 328ps on there
[14:28:35] <Kevin`> Essobi: the linux i2c stuff is fairly easy to use. i2c in slave mode is a probably the harder part
[14:28:46] <Essobi> Kevin`: It's a netbsd box.
[14:29:52] <Essobi> Kevin`: But I was thinking the same thing... I think the 328p has a hardware implementation.
[14:31:39] <Kevin`> Essobi: oh, I think it does too. you need real-time code for it though. not that realtime is exactly uncommon on a microcontroller, but it's a definitely harder than the host side where you can stop whenever you want
[14:32:47] <Essobi> Good point..
[14:33:36] <Essobi> I don't really what to do too much with the uCs thou.. sensor interfaces more then anything.
[15:24:46] <rue_house> clever, there?
[15:27:08] <OndraSter> I have just noticed that avr studio 5.1 now also gives you C++ and not just raw C
[15:27:18] <OndraSter> how much more space does it use in the flash? :)
[15:36:48] <abcminiuser> exDM69, you're pretty GIT knowledgable, right?
[15:42:42] <abcminiuser> Or anyone else here a git-master?
[15:42:57] <Steffanx> you
[15:43:05] <abcminiuser> Damnit
[16:00:41] <OndraSter> screw it, I am going back to regular C... :P
[16:02:53] <Casper> C++ on uC is usually a bad idea
[16:03:37] <norbi> is there any adc lib?
[16:03:49] <norbi> adc looks a bit complicated to me
[16:03:59] <norbi> never tried fully
[16:04:57] <Casper> ADC is easy
[16:04:58] <OndraSter> norbi, it is as easy as eating a .. pie
[16:05:22] <Casper> set the ADC, start the conversion, wait and get the data
[16:05:24] <Casper> btw
[16:06:13] <Casper> you can use ADC instead of ADCL and ADCH... ADC is a 16 bits containing the 10 bits of the ADC, so no need to read both register and manually combine them
[16:06:43] <OndraSter> Casper, the thing is, now even AVR Studio themselves "promote" C++ :P
[16:06:51] <OndraSter> so I expected it to be quite optimized
[16:06:52] <OndraSter> but oh well
[16:06:58] <OndraSter> 4kB RAM and 32kB flash "only"
[16:07:29] <OndraSter> "640kB ought to be enough for anybody"
[16:08:01] <Casper> they should have added an onchip divider
[16:08:17] <Casper> would have helped so much with math
[16:08:46] <OndraSter> yep, few more 16bit instructions would have been k3wl
[16:08:57] <OndraSter> mostly for my asm masochism
[16:09:32] <Casper> I started with asm, and gave up when I started to need 16 bits math :D
[16:10:00] <OndraSter> hehe
[16:10:11] <OndraSter> I have done some 16bit mul/add/sub/.. stuff
[16:10:18] <OndraSter> I wrote macros to ease up the pain
[16:10:27] <OndraSter> C wouldn't be fast enough most likely on that AVR
[16:10:54] <OndraSter> in asm I get about 60% CPU load just for the main routine (that runs actually in the interrupt... small code but it runs many times a second)
[16:11:29] <Casper> it's sad that it's so hard to figure out the cpu usage
[16:11:33] <Casper> and even the ram usage
[16:11:42] <OndraSter> I figured it quite simply actually
[16:12:13] <OndraSter> I always dumped the counter value when entering the interrupt (actually it was 00 because it was interrupt for overflow) and then grabbed the timer value when it was RETI-ing
[16:12:24] <OndraSter> did a bit of math and boom, I had the result
[16:12:53] <OndraSter> (how long does it take for the counter to count to the ending value * how many times per second it runs) :)
[16:13:32] <Casper> but ram is a pita
[16:13:37] <Casper> specially in C
[16:14:58] <OndraSter> yep
[16:18:24] <norbi> ok i did this until now: ADMUX |= (1 << REFS1) | (1 << ADLAR) ;
[16:18:24] <norbi> ADCSRA |= (1 << ADEN) | (1 << ADSC) | (1 << ADATE);
[16:19:00] <norbi> after this how do i read the adch?
[16:21:37] <norbi> anybody?
[16:21:49] <OndraSter> you should wait for the flag
[16:21:53] <OndraSter> and then you can read it :)
[16:24:11] <Essobi> oh ffs... shadio rack doesn't sell crystals? :|
[16:30:32] <norbi> im getting 0 from the adc, what can it be wrong
[16:30:33] <norbi> ?
[16:30:52] <Essobi> umm..
[16:31:08] <Essobi> You have no vcc?
[16:31:55] <norbi> hear is the code: http://pastebin.com/3i3LZYLk
[16:33:06] <norbi> the sensor is working
[16:33:18] <norbi> but the sensor max voltage is on aprox 2V
[16:33:23] <norbi> as it looks
[16:33:32] <norbi> and the internal voltage is 4.7V
[16:35:56] <norbi> i forgot to init the adc
[16:36:03] <norbi> but now im having constantly 255
[16:36:27] <norbi> and the sensor voltage is changing from 0.4 to 2.3
[16:39:39] <Xata> norbi: aref?
[16:40:43] <norbi> Xata: ?
[16:41:50] <Xata> norbi: what is your reference vo;ltage? also why not use interrupts?
[16:42:22] <norbi> Xata: i will use interrupts, but not yet, at least i want to see some values first, then i will optimize to interrupts
[16:43:37] <norbi> Xata: what ref voltage should it be?
[16:43:51] <norbi> the mcu is on aprox 5 volts
[16:44:04] <norbi> the sensor returns me values between 0 and 2... volts
[16:44:26] <Essobi> norbi: verified with a volt meter?
[16:44:30] <norbi> yea
[16:45:35] <Xata> norbi: but you have to choose the reference. try internal 2.75v (read about adc registers in datasheet). also you know that you redeclare bits in line 29 after line 27? it is 0x80, not 0x87 no more.
[16:47:40] <norbi> still 255
[16:47:49] <Xata> norbi: also use method from line 43 for setting bits. this is more text, but i, for example, too lazy to calculate and see what bits are set on 0x80
[16:47:51] <norbi> ADMUX |= (1 << REFS1) | (1 << REFS0) | (1 << ADLAR); ADCSRA |= (1 << ADSC) | (1 << ADEN);
[16:49:04] <Xata> norbi: cap. connected between AREF pin and gnd?
[16:49:28] <norbi> Xata: dunno, this is a prefabricated board
[16:49:40] <norbi> i should check the schematics for it
[16:49:57] <norbi> but im convinced it has
[16:50:46] <norbi> i see it has cap
[16:50:49] <Xata> hm. i am eating right now. when i'll be done i'll recheck code (mb curent version to pastebin?) and try to halp you
[16:51:47] <norbi> Xata: good appetite, and thx : http://pastebin.com/68NpfuCQ
[17:02:06] <Xata> norbi: first - i do't know if they go pre-set (but usually you have to set bits manually), but you have no prescaler setting in ADCSRA
[17:03:39] <Xata> uh, you are not using interrupts. well, try this, but i don't know if they are mandatory for adc, or just new conversion bits are enough
[17:07:03] <norbi> http://pastebin.com/iNse3hj5
[17:07:42] <norbi> xata now im using prescale of 32
[17:07:54] <norbi> i dont know what prescale would be better for 8mhz
[17:08:55] <norbi> ive set it now for 64
[17:09:05] <norbi> 64 should be good for 8mhz
[17:09:29] <Xata> 32 is far enough, i think. well, also i don't actually know - is it needed without ADC_vect
[17:10:17] <norbi> how would this code look like with interrupts?
[17:21:31] <norbi> http://pastebin.com/bFrriBmw
[17:21:47] <norbi> now i transformed it into interrupts, not it gives me a constant values of 127
[17:22:54] <norbi> any ideas?
[17:24:31] <Xata> norbi: try this http://pastebin.com/XLrG3BXp
[17:25:07] <Xata> norbi: no adc pin selector at a moment, but seems you don't really need it yet
[17:25:52] <norbi> 255
[17:26:05] <norbi> constant 255
[17:26:11] <Xata> hmmmm
[17:27:07] <norbi> it is needed to set it as input?
[17:27:15] <norbi> adc can be used as output too right>
[17:27:15] <norbi> ?
[17:27:41] <Xata> no
[17:27:54] <Xata> adc is only analog to digital, not digital to analog
[17:28:55] <Xata> but you can use it as digital output, as all the other ports. and now, you dont have to set DDRx, PINxy, etc digital stuff to make adc work
[17:30:00] <Xata> norbi: i had similar problem when on synthesizer there was a problem cpu stoping to respond after it had to work too hard
[17:31:00] <norbi> Xata: what can it be?
[17:34:19] <Xata> norbi: well, i don't get all the other stuff you do except adc - this is not even know registers for me. xmega or something? and second - this can bee just too hard work for mcu. try making a simple programm just lighting on led if you have high on adc1 and turning it off if you have low.
[17:34:30] <Xata> *known
[17:34:47] <sabesto> norbi: first off, ADMUX |= input is a bad idea
[17:35:03] <sabesto> at least do ADMUX = input
[17:35:19] <Xata> so you will find out where is the problem - in hardware or software.
[17:35:21] <sabesto> you will end up with all bits set after the first round
[17:35:39] <Xata> sabesto: check this http://pastebin.com/XLrG3BXp
[17:35:44] <Xata> he still has 255
[17:37:44] <norbi> Xata: it is an atmega64
[17:37:49] <norbi> it has lot of features
[17:38:03] <Xata> norbi: also i think if this is fabric board - maybe they have 5v on AREF? so you can try not setting REFS1 and REFS0 bits.
[17:38:23] <Xata> or not 5v, but exact voltage you need
[17:38:36] <norbi> how can i check what voltage is on aref
[17:38:37] <norbi> ?
[17:38:53] <Xata> norbi: voltmeter?
[17:38:57] <sabesto> its a physical pin
[17:39:11] <norbi> aref is 4.7
[17:39:42] <sabesto> what is connected to the adc inputs?
[17:39:59] <Xata> so, you have aref connected. try not setting refs bits
[17:40:25] <Xata> norbi: "fabric board" is arduino? :3
[17:40:27] <norbi> not setting the ref did not solved the problem
[17:40:35] <norbi> not arduino, cerebot 2
[17:41:27] <norbi> this is the schematic: http://digilentinc.com/Data/Products/CEREBOT-II/Cerebot%20II_sch.pdf
[17:42:10] <norbi> im using the servo controllers for the sensor, they are saying that pwm is on them, but i dont get how adc0...adc7 has any pwm
[17:42:13] <sabesto> is jtag enabled in the fuses?
[17:42:50] <sabesto> because that would fuck up ADC4-7 at least
[17:42:52] <norbi> no JTAGEN is not "checked"
[17:43:37] <norbi> Reading fuses address 0 to 2.. 0xEE, 0xC9, 0xFF .. OK!
[17:45:06] <Xata> sabesto: but he is working on adc0, at least with my code i send you
[17:45:24] <Xata> and there is nothing that could not working in that program
[17:45:41] <Xata> *could not be working
[17:45:45] <sabesto> i dont see any errors
[17:46:03] <Xata> so conclusion?
[17:46:08] <norbi> PF0 is the ADC0 right?
[17:46:10] <Xata> something with hardware
[17:46:13] <norbi> for the atmega64
[17:46:30] <sabesto> measure the pin with a multimeter and check the voltage
[17:46:36] <Xata> norbi: i use a32, because it is top for dip-models <_<
[17:47:08] <norbi> xata can you please check out this atmega64 for me?
[17:47:15] <norbi> am i using the fkin right pin/
[17:47:15] <norbi> ?
[17:47:15] <sabesto> PF0 is ADC0 yes
[17:47:29] <norbi> PORTF pin0
[17:47:32] <norbi> adc0
[17:47:39] <sabesto> yes
[17:48:43] <Xata> sabesto: but as far as i get it, you can not read info from that pin if adc is on, or you'l fuck up the conversion. y?
[17:49:22] <sabesto> i'm not shure
[17:49:46] <sabesto> you can disconnect the digital input buffer in DIRD0 register to make shure
[17:49:59] <sabesto> dont quote me on that
[17:50:05] <Xata> adc and pins are different stuff from cpu point of view. but if that is similar port - u'r screwed
[17:51:50] <Xata> sabesto: there is even no such stuff for a32 T_T. one day i'll buy a breakout
[17:52:21] <sabesto> hehe
[17:53:00] <sabesto> i just held a 3 day crash course in AVR
[17:53:11] <sabesto> finished today
[17:53:19] <sabesto> last task, ADC :P
[17:53:34] <sabesto> so i should know this :P
[17:54:16] <Xata> avr courses... oh civilized world...
[17:55:38] <sabesto> funny that a 1.st grader is teaching 2.nd->5.th graders how to use an AVR
[17:58:36] <Xata> this is university course? well, as i am 7 years in uni now, and still on 4th grade - i can say the question is in will
[17:59:57] <sabesto> university college
[18:00:39] <sabesto> yes of course, if you are interested you are golden, if not you are fucked
[18:00:57] <Xata> if you wish you can know almost everything about some subject in year, but you can fuck that stuff and just be there doing almost nothing for your diploma
[18:01:33] <sabesto> Xata: yeah, but your not getting a job
[18:01:35] <Xata> problem when you understand why first is better than second when you are 23 <_<
[18:03:38] <norbi> guys, fk this, still cant make it work
[18:03:44] <norbi> searched on net a lot
[18:04:03] <norbi> prescale set to 64, avcc is the ref
[18:04:13] <norbi> interrupt driven
[18:04:16] <Kevin`> I assure that it IS possible to use the adc on avr chips. it really exists.
[18:04:38] <norbi> i cant figure out just why the conversion its not happening
[18:04:51] <norbi> if i disconect the sensor the same value is given by mcu
[18:04:57] <norbi> Kevin`: save me:P
[18:05:11] <Casper> norbi: still have adc issue?
[18:05:15] <Kevin`> you started the first conversion, right? even if you tell it to run continuously you have to start it once
[18:05:17] <Xata> sabesto: this is not such maximalistic here. people have rich parents to place them somewhere just for everybody know their lovely kid has a job, people try to live in other stuff as me (i am metallurgist myself, and i hate metallurgy), people just go and work office lemmings. life will never leave you starving, if you don't really want this.
[18:05:17] <norbi> Casper: yea
[18:05:58] <norbi> guys, here is my code: #include <stdio.h>
[18:05:58] <norbi> #include <stdint.h>
[18:05:58] <norbi> #include <ctype.h>
[18:05:58] <norbi> #include <avr/io.h> // include I/O definitions (port names, pin names, etc)
[18:05:58] <norbi> #include <avr/interrupt.h> // include interrupt support
[18:06:09] <Xata> lol
[18:06:43] <norbi> omg
[18:07:07] <Xata> norbi: pastebin.
[18:07:16] <norbi> that was bad copy, it seemed that i didnt copied well the pastebin link
[18:07:34] <norbi> http://pastebin.com/wnVJWGFH
[18:07:40] <norbi> sry for that paste before
[18:07:45] <norbi> it was an accident
[18:08:21] <Xata> norbi: lol. about code i sent you. Kevin` is right - you better start a first conversion there, i forgot :D
[18:08:37] <norbi> what first conversion?
[18:08:42] <Xata> and maybe after that it will work
[18:08:52] <norbi> i've seen ppl talking about it on forums but i dont get it
[18:08:54] <Kevin`> it does look like he's setting ASDC, although i'm not sure about the ordering
[18:09:01] <Kevin`> I really don't want to open the datasheet :D
[18:09:11] <norbi> Kevin`: i understand you:d
[18:09:37] <norbi> Kevin`: ordering mathers?
[18:09:46] <Kevin`> norbi: oh, look at the part of the datasheet about the adc register locking
[18:09:48] <Xata> norbi: add ADCSRA|=1<<ADSC just after sei()
[18:09:51] <Casper> norbi: I see some code error
[18:10:05] <Casper> in the commented out section
[18:10:20] <Kevin`> norbi: if you read ADCH and ADCL in the wrong order, or read one but not the other (I forget which), it will never update the register
[18:10:23] <Casper> and you might want to avoid interrupt stuff at the begening
[18:10:42] <sabesto> Kevin`: its ok to read ADCH only
[18:10:45] <Casper> and don't read ADCH ADCL, just read ADC
[18:10:45] <Xata> Casper: http://pastebin.com/XLrG3BXp is mine conversion to interrupts, and i really don't understand that commented stuff
[18:11:05] <Kevin`> sabesto: are you 100% sure? you wouldn't be the first person to mess it up
[18:11:06] <norbi> Kevin`: yea, thats true but it is in 8 bit mode, not 10 bit, so i can ignore the ADCL
[18:11:21] <sabesto> ADCH must be read last
[18:11:43] <norbi> yea in case it is 10 bit mode
[18:11:49] <norbi> but it is in 8 bit and datasheet says in 8 bit mode adcl can be ignored
[18:12:28] <Casper> norbi: why don't you use the example in the datasheet?
[18:12:29] <Casper> also
[18:12:33] <sabesto> but yeah, i would not set ADSC high in the init
[18:12:35] <Casper> beware with ADMUX
[18:12:58] <Casper> it's where you select the channel, so you can't just ADMUX |= channel
[18:13:02] <Kevin`> norbi: that seems correct. here's an example from me anyway: http://kwzs.be/~kevin/logger.txt
[18:13:06] <norbi> now im getting constantly 192
[18:13:28] <Kevin`> norbi: I mean the ADCH seems correct, not the fiddly register bits I don't want to look up :)
[18:13:46] <norbi> there is no such file: #include "uart/USI_UART.h"
[18:13:57] <Casper> norbi: question: have you selected the proper AREF? is AVCC connected?
[18:14:03] <Kevin`> norbi: you wouldn't need that anyway, you already have uart code
[18:14:14] <Kevin`> norbi: it's for a different chip (attiny45, fwiw)
[18:14:37] <sabesto> aww, the darn USI module
[18:15:09] <Xata> guys, this all is a lot of fun, but also it is 2 am here, so i got to go and get some sleep. goodnight/bye
[18:19:50] <Kevin`> hmm, why do I have ADCH sampled in the main loop there
[18:19:52] <norbi> hm: http://pastebin.com/uLqg5dwW
[18:19:57] <Kevin`> oh well, I suppose it illustrates that it works that way too
[18:20:02] <norbi> now some strange value im getting
[18:20:03] <Kevin`> probably just random code scraps
[18:20:25] <norbi> Casper i've selected AVCC
[18:20:37] <norbi> but what does that propper avcc connected mean?
[18:20:38] <Kevin`> norbi: don't do this: a|=( ((uint16_t)ADCH)<<8 );
[18:21:02] <Kevin`> or at least I forget what it does, if it serves any purpose
[18:21:12] <Kevin`> norbi: also don't assume the registers are completely the same
[18:21:16] <Casper> norbi: I see a problem right away in your main
[18:21:31] <Casper> why are you only reading ADCH ?
[18:22:08] <norbi> Casper: because i had it set on 8 bits
[18:22:44] <Casper> why are you making it more complicated than the basic setup?
[18:23:44] <Casper> sample[] is misdeclared
[18:23:57] <inflex> hehehe
[18:23:59] <inflex> morning all
[18:24:07] <Casper> hi inflex
[18:24:13] <Kevin`> Casper: sample isn't an array, it's a number 0-63
[18:24:33] <Casper> samples[]
[18:24:35] <Casper> and sample
[18:24:44] <Kevin`> explain
[18:24:52] <Casper> they ain't volatile
[18:25:04] <norbi> ok lets start from this code: http://pastebin.com/9yiQw2Zx because im mixing it now, and started to understand nothing
[18:25:09] <Kevin`> ah indeed they should be
[18:25:09] <Casper> the content is lost once you leave the function
[18:25:09] <norbi> now im readinx ADC
[18:25:18] <Kevin`> ahtough fwiw it DOES work
[18:25:24] <Kevin`> although*
[18:25:40] <Kevin`> probably because samples is a pointer and sample isn't used outside the isr
[18:26:27] <norbi> o this adc is not that easy at it looks
[18:26:28] <norbi> :(
[18:27:12] <Casper> norbi: it is easy, you screw it up by trying to make it more complex than it is
[18:27:44] <Kevin`> norbi: do you have it working without the interrupts?
[18:28:09] <norbi> Kevin`: no
[18:28:35] <Kevin`> do that first, since it's a bit simpler. although it would work either way if the datasheet is followed to the letter the first time
[18:28:45] <norbi> Casper: i wont complicate it, this is the most simple code i can write: http://pastebin.com/9yiQw2Zx
[18:29:47] <inflex> norbi: as a general bit of guide... don't put too much in your ISRs
[18:30:05] <inflex> eg, the print() call would probably be a bit long under normal circumstances.
[18:30:29] <inflex> better to set a flag, then do the print in the while(1){} loop in the main body
[18:31:05] <inflex> but perhaps you already know that and was just being succint in this source example.
[18:32:03] <Casper> norbi: why are you using interrupt right at the start? this is a bad way to learn
[18:32:09] <norbi> inflex: thanks, yea you are right, im just trying to "scatching" this out
[18:32:28] <norbi> Casper: didnt used interrupts for it
[18:33:03] <norbi> i switched
[18:34:00] <Casper> I see interrupt enabled for it
[18:37:43] <norbi> http://pastebin.com/sVf77aBU
[18:37:56] <norbi> now is without interrupt
[18:37:59] <norbi> still same problem
[18:38:27] <Kevin`> norbi: you don't have free-running enabled and are only starting the conversion once
[18:38:39] <Kevin`> add the code to wait for the conversion to be complete, print, start new conversion
[18:39:00] <norbi> ive read this tutorial: http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=56429
[18:39:05] <norbi> but it doesnt seems correct
[18:39:16] <norbi> because that for will never end and just printing out the same value
[18:39:46] <norbi> what is the exact part of waiting for the code?
[18:39:59] <norbi> free running mode exists for this mcu?
[18:40:04] <norbi> i cant see in datasheet
[18:40:20] <Kevin`> norbi: it exist for almost all of them. check carefully. you have a line commented out for it
[18:40:46] <Kevin`> in any case your current code needs one or the other to be corrected
[18:41:10] <Kevin`> you shouldn't continuously poll adch if you want a stream of numbers, although it's sane in the example code for controlling leds or whatever it does
[18:44:16] <Kevin`> norbi: anyway, if you aren't using interrupts and aren't using free-running mode, there's a status flag you can poll for when a conversion is complete
[18:45:14] <Kevin`> eg while(ADCFOO&(1<<CONVERSIONINPROGRESS))
[18:45:17] <norbi> ok tried: http://pastebin.com/SWjJxcy1
[18:45:29] <norbi> still same problem
[18:46:09] <Kevin`> is ADSC what you poll? again I don't want to open the datasheet :D
[18:49:02] <norbi> yea
[18:49:04] <norbi> i think
[18:49:13] <norbi> still cant figure out what is wrong
[18:56:22] <norbi> i think im going to bed, its almost 3 am here
[18:56:35] <norbi> i cant figure out why this wont work, i think im doing all right
[18:56:39] <norbi> at least almost
[19:13:38] <Kevin`> go to bed, it'll work when you fix the obvious problems in the morning
[19:15:45] <Roklobsta> yeah sleep will help
[19:23:13] <Essobi> usually does... unless the problem is narcolepsy.
[19:23:56] * rue_house falls asleep on his keyboard while typing
[19:25:09] <Casper> do you imagine the problem that someone that suffer from it get?
[19:25:12] <Casper> no driver license
[19:25:37] <Casper> high risk of taking a train/subway from possibly falling on the tracks
[19:25:41] <Casper> and so on
[19:25:47] <Essobi> __ ___ _ _______ _ _ ____ ____ _ _ _____ _ ____
[19:25:47] <Essobi> \ \ / / \ | |/ / ____| | | | | _ \ | _ \| | | | ____| | _| _ \
[19:25:47] <Essobi> \ \ /\ / / _ \ | ' /| _| | | | | |_) | | |_) | | | | _| | | (_) | | |
[19:25:47] <Essobi> \ V V / ___ \| . \| |___ | |_| | __/ | _ <| |_| | |___|_| _| |_| |
[19:25:47] <Essobi> \_/\_/_/ \_\_|\_\_____| \___/|_| ( ) |_| \_\\___/|_____(_) (_)____/
[19:25:49] <Essobi> |/
[19:26:03] <Essobi> Casper: mind the gap
[19:26:24] <Roklobsta> well essobi, that's for the ascii mess. what does it say?
[19:26:31] <Roklobsta> thanks
[19:26:37] <Essobi> Roklobsta: Wake up rue. :D
[19:26:55] <Essobi> Get a nice monospaced font man. :D
[19:26:59] <Casper> Essobi: sadly, many modern client fuck up the "irc must use fixed font" rules...
[19:27:11] <Essobi> Yea... sheeeeitzweak.
[19:27:22] <Roklobsta> yeah quassel does that. damned open sores.
[19:27:29] <Essobi> I still IRC from UNIX. UTF-8 now, mind you.
[19:27:37] <impulse32> rue?
[19:27:45] <Roklobsta> whatever happened to dcc?
[19:27:49] * Essobi points up. He passed out on the keyboard.
[19:28:08] <Essobi> Roklobsta: It's still around... still sucks with NAT thou since it wasn't designed for that,
[19:29:25] <Essobi> Crazy to think I've been on IRC for... fuck I'm old.
[19:29:55] <Essobi> 24 years this May... Christ on a crutch.
[19:31:37] <Casper> the main problem with DCC is that the clients ain't upnp
[19:32:29] <Essobi> well.. upnp is just a work around for non-nat compat protocols..
[19:33:04] <Casper> yes, but a good workaround
[19:33:35] <Essobi> I guess... manufacturers and client programmers alike can't agree on the f'ing RFC thou.
[19:34:09] <Essobi> Did I mention I was a network engineer by day? :D
[19:34:26] * Essobi kicks his firewall.
[19:35:04] <Casper> do you have experience with VoIP ata?
[19:35:08] <Essobi> hahaha
[19:35:36] <Casper> I guess bad one?
[19:35:43] <Essobi> I was a VoIP R&D engineer for 8 years before parlaying back into security.
[19:36:19] <Essobi> SIP/SCCP/H323
[19:36:26] <Casper> I'm considering to get VoIP, but... man it's a mess of lack of info
[19:36:56] <Essobi> SIP RFC is yet another unfollowed IEEE spec.
[19:37:18] <Casper> the problem I get is that all the reviews are unreliable
[19:37:21] <Essobi> Too many should's and coulds, and not enough musts.
[19:37:22] <Roklobsta> IAX2 is the go.
[19:37:56] <Casper> "oh this review is positive, but look like it was written by an advertisement compagny"
[19:38:02] <Essobi> Meh.. IAX isn't much better considering you have to use asterisk. :D
[19:38:14] <Roklobsta> is asterisk crap?
[19:38:15] <Casper> "this review... 1 star... look like a user error"
[19:38:21] <Essobi> Roklobsta: lulz.. Yea.
[19:38:32] <Roklobsta> i am thinking of setting up voip at home - what is a good linux voip server?
[19:38:37] <Casper> asterix
[19:38:40] <Essobi> Roklobsta: FreeSwitch.
[19:38:54] <Roklobsta> essobi: does freeswitch have different guts?
[19:39:07] <Essobi> Casper: what kind of ATA you looking for? price wise, etc..
[19:39:08] <Casper> thing is... everyone have their POV...
[19:39:28] <Essobi> Casper: :D Asterisk is a heaping pile of shit if you ever read their code.
[19:39:40] <Casper> Essobi: http://www.microbytes.com/advanced_search_result.php?manufacturers_id=&keywords=voip&osCsid=b3e28553ff8364efcada61f67f02833f ← one of those 3
[19:39:41] <Roklobsta> i did set up asterisk for an intercom system for a security gatehouse. i hated the syntax.
[19:40:37] <Essobi> Roklobsta: ehh, yea, the call processing language kind of sucks.
[19:40:48] <Casper> Essobi: the first 2, I read that the firmware is full of bug and cisco ain't fixing them (took over 1 year to fix the caller id bug, 2 major bug is still there)
[19:40:52] <Essobi> Roklobsta: FS isn't much better.. it's XML but atleast it's extendable.
[19:40:58] <Casper> didn'T checked the more expensive one yet
[19:41:49] <Essobi> SPA3102 is one of the better ones..
[19:41:51] <Roklobsta> i rellie just got an iphone so i was thikning of installing a voip server on their mythtv box so they could use it as an extension at home and when they are out.
[19:41:56] <Essobi> I've got a few SPA's sitting around here.
[19:42:44] <Essobi> Casper: most of the linksys ATAs are pretty good.
[19:42:47] <Casper> Essobi: how's the 112?
[19:43:20] <Essobi> ah, that's a new one.. not much experience with that one.
[19:43:40] <Essobi> Usually you can ask around in #freeswitch ... someone usually has played with the name brand gear in there.
[19:43:47] <Casper> thing is, I don't really need the landline option
[19:43:55] <Casper> oh will check
[19:44:06] <Essobi> I was building primarily large enterprise custom call switch routers.
[19:44:31] <Essobi> We had about 20K of those early linksys SPAs dangling off that system.
[19:45:10] <Casper> the 2 lines option is interresting, but the landline option is also interesting
[19:45:46] <Essobi> I've got a 3 port FXO/FXS PCI card sitting here...
[19:46:20] <Casper> in the short/medium term, it will be used a few minutes per month
[19:46:30] <Casper> in the longer term it will be the main phone line
[19:46:40] <Essobi> They ain't bad either if you just want to dump a dialtone on existing wiring. Disco at the gray box demark on your house.. pop it into a server, and plug it into the nearest jack and your lines are hot.
[19:46:44] <Casper> so the 2 lines could get interresting
[19:47:25] <Casper> like voip.ms look good, but there is basically no review
[19:47:27] <Essobi> If I was going to do that... I've just get a SIP client for my cell phones and term it locally.
[19:48:04] <Essobi> yea, most of those guys in #fs have experience with the smaller voip vendors... hell some of the vendors hang out in there on occasion.
[19:48:21] <Essobi> MMmmmkay. Time to wire up a DAPA and a 328p and break some shit.
[19:48:55] <Casper> basically 1$/month for DID, then 1 cents/min incomming, 0.51 cents outgoing... or 5$/month DID 3500 mins incomming, 0.51cents outgoing
[19:50:08] <Casper> but some say that the sound quality isn't great
[19:50:41] <Essobi> !!>
[19:50:51] <Essobi> Get someone who supports G711.
[19:51:27] <Essobi> It's a 64K stream, in both directions.
[19:53:01] <Essobi> It's about the best you can do over PSTN copper, so not much sense in exceeding that unless you go wide-band for point to point.
[19:54:40] <Casper> 1 G.711u (μ-law/pcmu)
[19:54:50] <Casper> that one?
[19:56:45] <Essobi> yessir
[19:57:00] <Essobi> That's got the best MOS score for a PSTN line too.
[19:57:10] <Casper> what is better? sip or iax2?
[19:57:43] <Essobi> SIP is supported by more ATAs, and other endpoints.
[19:58:02] <Casper> voip.ms support both
[19:58:05] <Essobi> Personally I hate both of them. I just hate SIP a little less.
[19:58:17] <Casper> http://voip.ms/ check the feature list
[19:59:25] <Essobi> ehh, doesn't look like a complete hackjob.. so that's good thing.
[20:01:03] <Casper> they are a bit too flexible so the config is kinda confusing, but it's not too bad
[20:31:37] <Essobi> What's the minimal circuit I need for a 328p at 5V?
[20:34:43] <mog> Essobi, a 5v power source and a 328p
[20:35:08] <mog> if your smart youd put in a decoupling capacitor to have cleaner power draw
[20:42:28] <Essobi> mog: what size you think?
[20:42:42] <Essobi> mog: I'm not doing anything with the ADC on here..
[20:47:58] <mog> 1uF would be a lot
[20:48:10] <mog> it depends how bad the 5v source is
[20:48:29] <mog> i have used avrs powered just from a battery and never had a problem
[20:48:42] <mog> decoupling caps are just a best practice thing
[20:49:40] <Essobi> arn't you supposed to put it between the 5v source and the VCC?
[20:52:41] <mog> decoupling caps go from the power source and are bridged to ground
[20:53:11] <mog> that way if the voltage source has a bad signal or gets cut off badly it can clean the signal
[20:54:18] <clever> its not just to get rid of existing noise, but to prevent creating more noise
[20:54:30] <mog> clever, is also right
[20:54:48] <clever> the wires/traces from the power to the avr act like very low value inductors, and resist changes in current
[20:55:04] <clever> plus a linier regular may have a sudden voltage drop if current spikes, before it can react
[20:55:24] <clever> the caps at the chip are much closer, so you have less voltage drop on the wires to it
[20:59:32] <mog> but given what you said earlier you dont need anything but the source and avr
[21:11:02] <spjt> I just run my AVR's off a USB cable I cut open, a breadboard and an old iPod charger.
[21:12:20] <Tom_itx> good for 400ma iirc
[21:12:31] <Tom_itx> or 100 unless you request more
[21:54:33] <gkwhc> hey guys, when designing PCBs, can silkscreen width generally be smaller than the min trace width?
[21:55:58] <Tom_itx> there's a point it won't be very legible
[21:56:23] <ziph> And it depends on how they're doing the silkscreen.
[21:56:38] <Tom_itx> yeah, some mfg are sloppier than others
[21:57:08] <gkwhc> hm most mfgs dont specify the min width though
[22:10:10] <Casper> Essobi: decoupling cap are 10-100nF ceramic
[22:10:41] <Casper> Essobi: it go between ground and the power pin, as close physically as possible to those 2