Back
[03:16:14] <OndraSter> abc
[03:16:18] <OndraSter> hmm he is not around
[03:16:37] <CapnKernel> _abc_ or abcminiuser?
[03:16:42] <OndraSter> abcmu
[03:17:45] <OndraSter> so, the problem with dragon fw update: I was not the only one appearantly. He didn't manage to get the same problem, but I got email from atmel ATTicket saying "we have more people happening this"
[03:18:01] <OndraSter> CapnKernel, have you seen my nuclear battery? :D
[03:18:02] <OndraSter> http://clip2net.com/s/1VQv2
[03:19:31] <CapnKernel> LOL
[03:20:58] <Olli> Hi, I'm new to AVR and so on. I got an AVR Dragon and want to build up something with an ATmega1284P. I'll use an external crystal with 18.4320MHz. Right now I'm programming it with ISP and the internal clock source fused. I understand, when I now fuse an external full swing crystal (which is what I guess I have to fuse?), I can no longer use the ISP method when placing the ATmega on the Dragon as I have no clock
[03:20:59] <Olli> signal there - but - can I still use the HVPP mode of the Dragon? Is there any downside to use the HVPP mode?
[03:21:28] <OndraSter> you have the mega in DIP?
[03:21:33] <Olli> yes DIP40
[03:21:49] <OndraSter> yep, once you set it to use external clock and you do not feed one, it won't work
[03:22:00] <OndraSter> if you have another mega - you can create square wave from it and feed the xtal on that
[03:22:04] <Olli> but hvpp would work?
[03:22:09] <OndraSter> yes
[03:22:17] <OndraSter> I suggest breadboarding ISP connector as well
[03:22:20] <Olli> is it "bad" to use hvpp from than on?
[03:22:56] <OndraSter> well it shouldn't be.... but any play with HVPP is playing with danger, I've seen people fry their Dragons when they accidentaly shorted HVPP pins :)
[03:23:10] <Olli> i also have some clock generators (those 4 pin tings), would it be "better" to build a small circuit to supply lets say 16 MHz to XTAL1 of the ATmega?
[03:23:48] <OndraSter> you can use either actually I think - tbe difference between xtal and external square is just that the xtal circuitry is powered on
[03:24:04] <OndraSter> so when you set xtal, you can use both xtal or the oscillator
[03:24:14] <Olli> ok
[03:24:33] <OndraSter> I _think_ -- I heard it either here or on forums :)
[03:24:53] <Olli> by the way... is "full swing crystal" the correct fuse to set? atmel stuid 6 is a bit cryptic about the clock fuses
[03:25:25] <OndraSter> do you see anything else for high speed xtal? :)
[03:25:27] <Olli> I have EXTCLK, INTRCOSC, FSOSC, EXTOSC....
[03:25:36] <Olli> and EXTLOFXTAL
[03:25:47] <Olli> not sure about EXTOSC or FSOSC
[03:26:00] <Olli> i guess EXTCLK is this 4-pin clock generating circuit
[03:26:02] <desaster> http://www.engbedded.com/fusecalc
[03:26:03] <OndraSter> extosc should be the "4 legged chip"
[03:26:11] <OndraSter> desaster, that doesn't have the mega1284 :)
[03:26:21] <Olli> but what is then extclk ondraster?
[03:26:27] <desaster> oh
[03:26:33] <OndraSter> Olli, hmm
[03:26:33] <Olli> see? ;)
[03:26:45] <OndraSter> maybe extosc = external RC oscillator?
[03:26:49] <OndraSter> check datasheet
[03:27:33] <OndraSter> woah the datasheet is 24MB
[03:28:15] <Olli> datasheet says: low power crystal, full swing crystal, low freq. crystal, int. 128jhz rc osc, calibrated int. rc osc. external clock
[03:28:16] <CapnKernel> Or:
http://www.frank-zhao.com/cache/avrfusecalc.php
[03:29:43] <OndraSter> http://clip2net.com/s/1VTpG
[03:30:17] <Olli> yes this was what i meant with "datasheet says" ;)
[03:30:23] <OndraSter> ah
[03:30:26] <OndraSter> didn't notice that line :D
[03:30:35] <OndraSter> anyway, full swing xtal is what you need
[03:30:40] <Olli> but i'm working with atmel studio.... and I only have those crypted selections
[03:30:42] * OndraSter hides before it blows up
[03:30:43] <Olli> okay
[03:30:53] <OndraSter> since you have Dragon, you have HVPP
[03:30:57] <OndraSter> so if it screws up... :D
[03:30:59] <Olli> ;)
[03:31:00] <OndraSter> you can always recover it :D
[03:31:39] <Olli> so I set up full swing osc, then build a small circuit to supply 16 MHz from an external clock generator (4pin) to xtal1 and then I can still programm with ISP?
[03:31:54] <Olli> the "target circuit" will of course have a crystal - not a clock generator
[03:32:11] <OndraSter> well if you keep it in the breadboard you can put there the five wires for ISP anyway
[03:32:33] <Olli> the target circuit will be wired on a prototype board... so no breadboard
[03:32:38] <OndraSter> even better :)
[03:32:55] <Olli> and since i want to attach an sdcard to the SPI pins I do not want to place a isp header and so on on the board
[03:33:06] <Olli> and jtag on the board is also not possible as i need the port for other things
[03:33:12] <OndraSter> oh
[03:33:20] <OndraSter> I have similar problem :)
[03:33:24] <OndraSter> JTAG occupying 4 pins
[03:33:29] <OndraSter> but no debugging = pain
[03:33:52] <Olli> i'll have rs232 output... manual debugging then ;)
[03:33:57] <OndraSter> :D
[03:34:22] <Olli> doesn't supply the dragon on its XTAL1 pin of the HVPP connector a clock?
[03:34:34] <OndraSter> well on the HVPP it most likely does
[03:35:11] <Olli> wanted to measure it with my oscilloscope but have to install atmel stuido first on the laptop.... no way to bring the scope to the dragon - better the other way round ;)
[03:35:37] <OndraSter> no space left on the table? :)
[03:35:44] <Olli> just need to check if the clock is always generated there.... if yes i could just route this pin then to XTAL1
[03:35:52] <Olli> scope is much to heavy ;)
[03:36:01] <OndraSter> it can't be _that_ heavy
[03:36:06] <Olli> old 40MHz dualbeam scope...70s like
[03:36:14] <OndraSter> I have 80s scope
[03:36:41] <OndraSter> it was fairly heavy after bringing it from school to home
[03:36:46] <OndraSter> ~40 minutes ride :)
[03:37:04] <OndraSter> 30 minutes ride *
[03:37:08] <OndraSter> + 20 minutes walks :P
[03:37:29] <Olli> 37kg the datasheet says...
[03:37:33] <OndraSter> :o
[03:37:38] <OndraSter> okay, it _is_ _that_ heavy :D
[03:37:50] <Olli> don't want to bring it from the cellar to the 1st floor if I could do the other way round with the laptop ;)
[03:38:23] <Olli> http://files.pofo.de/P1070427.JPG
[03:38:31] <OndraSter> wow
[03:38:35] <OndraSter> that is oldschool =)
[03:38:42] <Olli> 70s - as i said ;)
[03:39:05] <OndraSter> the knobs look great though :D
[03:39:42] <OndraSter> I have got this
[03:39:42] <OndraSter> http://www.po-recycle.com/shop/p/po-recycle/img-lib/spd_20100529122027_b.jpg
[03:39:57] <Olli> looks more modern ;)
[03:40:12] <OndraSter> the volts/div knobs are a bit screwed already, so one has to gently tap it when it shows junk :D
[03:40:50] <OndraSter> plus function gen
http://regimuszereklhanzi.5mp.eu/honlapkepek/regimuszereklhanzi/rc4jvjQies/nagy/p0039009.jpg
[03:42:11] <Tom_itx> mmm 1284 comes in dip?
[03:42:30] <Tom_itx> i thought that was a 100 pin part
[03:42:38] <Tom_itx> but then i didn't look
[03:42:51] <Olli> the 1284p yes
[04:16:46] <OndraSter> hmm is it known that any _delay_us or _delay_ms that should be >2ms and <100ms to make problems?
[04:16:59] <OndraSter> when I put there either _delay_ms(1) or 2, it is immediate
[04:17:11] <OndraSter> but when I put there _delay_ms(5) or _delay_us(5000) it takes about 16 seconds ?!
[04:17:18] <OndraSter> but when I put there _delay_ms(5000) it does 5 seconds
[04:17:20] <OndraSter> hmh
[04:18:09] <OndraSter> trying in emu to be sure
[04:20:20] <OndraSter> oh I know where the problem lays
[04:20:23] <OndraSter> ignore me =)
[04:20:37] <OndraSter> I forgot that _delay uses just loops instead some timer
[04:20:58] <OndraSter> and well, when the loop gets interrupted it create delays
[04:22:36] <OndraSter> but how come that when using 5000ms delay, that it is approx 5 seconds..
[04:22:50] <OndraSter> the error should be the same all the time?!
[04:22:55] <OndraSter> I mean
[04:22:57] <OndraSter> linearily
[04:26:00] <OndraSter> ok, refreshing DRAM takes 106ms every second... about 10% CPU load "only"
[04:26:07] <OndraSter> hello sirdancealot2
[04:26:25] <OndraSter> that is on 8MHz clock
[04:27:27] <Olli> mh... looks like XTAL1 on the HV_PROG socket of the dragon is not supplied with a clock signal when programming in ISP mode
[04:33:49] <sirdancealot2> hello
[04:34:11] <OndraSter> oh I know where is my error, I am launching refresh 488x per second rather 64 times lol
[04:34:12] <OndraSter> oh well
[04:40:33] <OndraSter> hah, now it works!
[04:40:48] <OndraSter> 15ms between refreshes sounds about right, the required time is 16ms
[04:42:03] <OndraSter> okay, 11.1% load, now we're talking
[05:41:36] <Sordna> Tom_itx, the atmega1284p is an awesome chip in 40pin dip package...complete with 16K of sram....
[05:52:59] <OndraSter> yep
[05:53:01] <inflex> PDIP... *shudder*
[05:53:02] <OndraSter> but how much does it cost :P
[05:53:23] <OndraSter> mouser says $6.29 for 1pcs
[05:53:26] <OndraSter> @ DIP
[05:53:41] <OndraSter> I think that getting xmega256a3u with adapter board would be cheaper :)
[05:53:55] <OndraSter> I need to make one
[05:53:56] <OndraSter> ..
[06:06:35] <Sordna> some of us like pdip...they fit breadboards well after all.
[06:06:51] <Sordna> comes in other packages as well.
[06:12:13] <inflex> I'd rather they phase it out and just replace it with a SMD mounted PDIP breakout
[06:12:24] <inflex> makes the assembly line cheaper to avoid PDIP
[07:37:25] <OndraSter> so, should I decouple EACH vcc pin with 100nF + 10uF or just one or two 10uFs globally and 100nf at each pin?
[07:37:34] <OndraSter> it is quite hardcore having 14 caps around the MCU :)
[07:46:14] <Steffanx> " just one or two 10uFs globally and 100nf at each pin" should do it ..
[07:46:25] <OndraSter> fine
[07:46:29] <OndraSter> that's what I needed to hear :D
[07:47:12] <OndraSter> if we're talking breadboard adapter one hasn't got much space for caps if he wants to fit .8"
[07:55:21] <CapnKernel> The rule with 74xx and 40xx chips used to be one per chip. If you have a fancy chip, the datasheet should tell you the recommended decoupling.
[07:56:14] <mrfrenzy_> when we're going into MHz-range thumbrules don't really help. I would certainly read the datasheet
[07:56:40] <CapnKernel> OndraSter: My guess is that the 10uFs are for power input noise and prevention of sag when loads are switched in and out. The 100ns should be right next to the chips, and are to soak up the switching noise the chips induce on the power rail
[07:57:02] <OndraSter> yeah
[07:57:03] <OndraSter> this is xmega
[07:57:08] <OndraSter> 100nF per vcc is sure
[07:57:20] <OndraSter> 10uF is usually per chip
[08:00:12] <CapnKernel> Check out the decoupling caps:
http://www.obsolyte.com/sunPICS/3_280/CG3_damage.jpg
[08:00:48] <mrfrenzy_> sweet
[08:01:04] <OndraSter> :D
[08:01:10] <OndraSter> they are talking hundreds of MHz I bet :P
[08:01:15] <mrfrenzy_> for some old logics chips they had some kind of foil running along the lenght of chips providing vcc+gnd and acting as huge decoupling caps
[08:04:08] <mrfrenzy_> http://imgur.com/BvvtV
[08:06:29] <CapnKernel> Wow, what a picture. That's just odd.
[08:06:36] <karlp> OndraSter: not necessarily hundreds of megs, just you need more decoupling if you have more power switching,
[08:06:46] <karlp> bimos or actual ttl chips
[08:23:58] <Tom_itx> OndraSter, i've seen as many as 3 different values on each vcc pin
[08:24:25] <Steffanx> Someone who liked over engineering?
[08:24:45] <OndraSter> 100n 1u 10u .)
[08:24:46] <Tom_itx> i think it may have been called out
[08:25:06] <Tom_itx> i don't recall what it was on
[08:25:31] <CapnKernel> A crack pipe
[08:25:46] <OndraSter> :D
[08:26:08] <Steffanx> First digital crack pipe ever..
[08:26:28] <CapnKernel> I see a lot of really really lame PHP code. Seems to be that PHP implies lame.
[08:26:52] <CapnKernel> My point is that just because it exists in a product someone paid real money for, it doesn't mean it was actually a good design, or that the person knew what they were doing.
[08:26:53] <Steffanx> Or .. lots of beginners play with php
[08:26:55] <Steffanx> and write crap
[08:27:19] <CapnKernel> Yeah. No training in software engineering, no concept of architecture.
[08:27:56] <CapnKernel> That board with the 100n 1u 10u, we don't know that it wasn't designed by the intern.
[08:28:19] <OndraSter> OH GOD PHP
[08:28:27] * OndraSter runs away to hide from the crappiest language ever
[08:28:35] <Steffanx> There we go .. the rants have started
[08:28:55] <Steffanx> I wonder when Java comes in
[08:29:01] <CapnKernel> There's nothing terrible with the language. It's just the crowd it attracts.
[08:29:01] <Steffanx> And perhaps linux/windows/os x
[08:29:09] <OndraSter> Java is about the same crappiness as PHP :P
[08:29:23] <Tom_itx> now he's just crying
[08:29:31] <CapnKernel> Java in a nutshell:
http://ptrthomas.wordpress.com/2006/06/06/java-call-stack-from-http-upto-jdbc-as-a-picture/
[08:29:49] <OndraSter> I used to work as a PHP programmer
[08:30:00] <CapnKernel> PHP is generally under-architected. Java is generally over-architectured.
[08:30:00] <OndraSter> I was really desperate
[08:30:26] <CapnKernel> I could take that to mean that being a PHP programmer wasn't getting you laid.
[08:30:38] <OndraSter> :D
[08:31:00] <CapnKernel> It's the second most disturbing thought I can think of.
[08:31:56] <CapnKernel> The first being the idea that out there, there are programmers who because they're writing in PHP, are getting laid.
[08:33:14] <OndraSter> PHP is every child's wet dream, because of ignoring data types. It is also every programmer's horror, because of ignoring data types.
[08:33:15] <Tom_itx> must be a very very high ratio of females in that area
[08:34:49] <CapnKernel> OndraSter: In other words, PHP is a sheet stain.
[08:35:01] <OndraSter> sheet?
[08:35:08] <CapnKernel> bedsheet
[08:35:23] <OndraSter> more like shit stain :)
[08:35:38] <OndraSter> I am glad I moved away from PHP
[08:35:46] <CapnKernel> Python doesn't have strong typing either.
[08:35:47] <OndraSter> if I will ever do web stuff again - my heart will go to ASP.NET
[08:35:55] <OndraSter> yeah and Python doesn't have { } either
[08:36:00] <CapnKernel> The difference is that Python attracts a different kind of people
[08:36:05] <OndraSter> or at least not for the proper usage
[08:36:22] <CapnKernel> A lot of python code is graceful, and people try to make it graceful
[08:41:14] <karlp> asp.net? are you _trying_ to leave in the 90s?
[08:41:19] <karlp> s/leave/live/
[08:42:14] <OndraSter> no?
[08:42:18] <OndraSter> asp.net is new
[08:42:20] <OndraSter> asp is 90s
[08:42:46] <OndraSter> whole .NET, Silverlight etc are <3
[08:42:56] <karlp> right.....
[08:43:07] <CapnKernel> OndraSter: Whatever respect I had for you has just evaporated.
[08:43:12] <OndraSter> ...
[08:43:12] <OndraSter> lol
[08:43:21] <OndraSter> what's wrong on .NET?
[08:43:39] <OndraSter> hell, they machine rewrote Android to .NET and ran on Mono
[08:43:44] <OndraSter> and many stuff was many times faster
[08:43:58] <Steffanx> Mono ..
[08:44:13] <OndraSter> (yeah, because there is no official .net fx for linux)
[08:44:37] <CapnKernel> I'm getting a soft-on.
[08:45:22] <OndraSter> do you remember that I said that I am getting a Netduino? :P
[08:49:35] <OndraSter> xmega breakboard:
[08:49:35] <OndraSter> http://clip2net.com/s/1VXSl
[08:49:46] <OndraSter> http://clip2net.com/s/1VXSy
[08:50:08] <OndraSter> I always need to take a look at renders
[08:50:12] <OndraSter> so I can see the bottom layer
[08:54:44] <Steffanx> See the bottem layer?!
[08:55:01] <OndraSter> see it clearly
[08:55:35] <OndraSter> Eagle is missing out some things, showing just bottom layer takes IMHO too much steps
[08:55:53] <Steffanx> Two many steps?! 3 clicks?
[08:56:16] <Tom_itx> Steffanx, it should just be a thought process
[08:57:17] <OndraSter> Steffanx, how 3 clicks? I have to open layers and disable all except all bottom layers
[08:57:27] <Tom_itx> no jtag header?
[08:57:34] <OndraSter> PDI
[08:57:52] <Tom_itx> no rtc xtal?
[08:58:02] <OndraSter> nope
[08:58:13] <Steffanx> It's not clear enough when you only disable the top layer OndraSter ?
[08:58:52] <OndraSter> no, not just top layer, you have to hide all the top silkscreens etc
[08:59:10] <Steffanx> No, you don't have to :P
[08:59:29] <OndraSter> how come?
[09:00:29] <OndraSter> Tom_itx, plus there is builtin rtc xtal
[09:00:42] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/misc_stuff/mega128.jpg
[09:01:11] <OndraSter> <OndraSter> Tom_itx, plus there is builtin rtc xtal
[09:01:21] <Tom_itx> you just said that
[09:01:40] <OndraSter> just repeating to be sure :P
[09:01:53] <Tom_itx> to be sure that you said it?
[09:02:16] <OndraSter> that you noticed it
[09:02:33] <CapnKernel> Tom_itx: Interesting board
[09:02:35] <OndraSter> when you write something the same second I blink on you, I have already faced few times that they overlooked that lin
[09:02:35] <OndraSter> e
[09:02:35] <Steffanx> Why no usb mini B connector .. usb mini B is HOT OndraSter :P
[09:02:44] <Steffanx> And it should at least be SMD
[09:02:45] <Tom_itx> CapnKernel, pepsi made that one
[09:02:51] <Tom_itx> long time ago
[09:02:59] <OndraSter> Steffanx, I broke too much mini usb's
[09:03:03] <CapnKernel> It was coke's day off? (Not sure who pepsi is)
[09:03:13] <Tom_itx> he's a coke head
[09:03:32] <Steffanx> He wasn't back than ..
[09:03:45] <Tom_itx> yeah i bet he was
[09:04:00] <Tom_itx> i think he's cleaned up a bit recently
[09:04:30] <Tom_itx> moved back to Chicago and got away from all the California nuts
[09:04:52] <Tom_itx> anyway. gtg
[09:05:05] <karlp> don't use miniusb, use micro or fullsize
[09:05:20] <Steffanx> No, not fullsize
[09:05:40] <karlp> so, micro then :) but not mini.
[09:05:47] <OndraSter> micro is awful
[09:05:52] <Steffanx> So go mini b :P
[09:06:27] <Steffanx> A fatass connector on a small board is ugly
[09:06:36] <OndraSter> I broke already 3 miniUSBs in phones
[09:06:38] <OndraSter> I hate miniUSB
[09:06:47] <Steffanx> Maybe you should be more careful :)
[09:06:50] <CapnKernel> Interesting comparison: The design lifetime in terms of insertion, between mini-B and micro.
[09:07:19] <CapnKernel> mini-B is rated for a few thousand. micro is rated for 10,000.
[09:07:40] <karlp> this is why mini has been killed.
[09:07:46] <karlp> don't put mini on anything new
[09:07:57] <karlp> also, failure mode for mini was the socket, not the cable,
[09:08:06] <karlp> which is the wrong side to break first.
[09:08:36] <Steffanx> Never had any trouble with mini b
[09:09:43] <OndraSter> I don't trust micro USB either
[09:09:50] <OndraSter> I always feel like I am going to break it any second
[09:09:57] <OndraSter> what's wrong on USB B? :P
[09:10:00] <OndraSter> it is nice
[09:10:06] <Steffanx> [15:56:10] <Steffanx> A fatass connector on a small board is ugly
[09:10:09] <OndraSter> bla bla
[09:10:18] <OndraSter> any other reason? :P
[09:10:32] <OndraSter> more people has at homes USB A - USB B btw
[09:10:34] <OndraSter> than mini USB or micro USB
[09:10:35] <Steffanx> Hey, and you are the person with the design tiger on a board?
[09:10:40] <OndraSter> yes
[09:10:44] <OndraSter> that is different board
[09:10:53] <OndraSter> that is on hold before I get familiar with xmega completely
[09:10:57] <Steffanx> I think i have more usb min B than any other usb cable here ..
[09:10:58] <OndraSter> thus before I make this board :)
[09:11:02] <OndraSter> lucky you
[09:11:04] <Steffanx> Yeah
[09:11:47] <Steffanx> Who gave you that info btw OndraSter ? :)
[09:12:12] <Steffanx> "more people has at homes USB A - USB B btw" <= that
[09:12:31] <OndraSter> because everybody has printer
[09:12:35] <OndraSter> which usually has USB B
[09:12:38] <OndraSter> (or ethernet)
[09:13:29] <OndraSter> CapnKernel, can I put more designs on one board and have it somehow separatable? I have got 27x72 board and the minimum is usually 50x100 for this size :)
[09:13:41] <OndraSter> so I'd put there some SOP->DIP adapters for example
[09:14:36] <Steffanx> btw .. that GREAT mono even lacks serialport support OndraSter
[09:15:04] <Steffanx> At least, I can't find it in the most recent version
[09:15:10] <OndraSter> don't ask me
[09:15:14] <OndraSter> SerialPort is in original .net fx
[09:20:38] <CapnKernel> OndraSter: You can have whatever size board you want, from 1cm x 1cm to around 50cm x 110cm. You will pay for what you use, not some arbitrary {5,10} quantization.
[09:20:46] <OndraSter> oh
[09:21:01] <CapnKernel> (Exception: Special cheaper prices apply if your board fits within 5cm by 5cm
[09:21:15] <OndraSter> and how much would it be then? :) 72x25 is the size then
[09:22:04] <OndraSter> the area of 72x25 is smaller than 50x50 :D
[09:22:10] <OndraSter> mm
[09:22:45] <CapnKernel> The 5 x 5 deal is not about area. It's a special case if both dimensions are no bigger than that.
[09:22:50] <OndraSter> oh
[09:23:04] <CapnKernel> If it's not within that rectangle, you'll be charged according to the area.
[09:23:35] <CapnKernel> The default board type is 2 layer, 1.6mm thick FR-4, white silkscreen on green soldermask, 1oz copper and Pb solder finish - perfect for hobbyists. Any of those attributes can be changed, which lifts the price a little.
[09:24:02] <OndraSter> RoHS in price, ENIG is + how much?
[09:24:09] <OndraSter> (where is the online calculator you talked about? :)
[09:24:28] <CapnKernel> Not available yet
[09:24:31] <OndraSter> I don't need ENIG anyway
[09:24:32] <OndraSter> ah
[09:24:40] <CapnKernel> RoHS and ENIG are extra
[09:25:09] <CapnKernel> Do you have a name for this board?
[09:25:21] <OndraSter> err I ment not RoHS but ... the other thing
[09:25:24] <OndraSter> the other thing than ENIG
[09:25:32] <OndraSter> HASL
[09:25:46] <CapnKernel> Yep. w/ Pb solder.
[09:25:48] <OndraSter> name... "ATxmega256a3u breakboard"
[09:27:44] <CapnKernel> How many boards you looking at?
[09:27:48] <karlp> OndraSter: people right now might have usbA to usbB, from printers, but usb micro is the euro mandated phone charger connector, and that means you're about to have wall-wart->usbA->usb micro all over the place.
[09:28:57] <OndraSter> hmm
[09:29:07] <OndraSter> but I am worried that it is easily rippable
[09:29:12] <karlp> stop being worried
[09:29:33] <CapnKernel> OndraSter: How many boards? (Obviously the more boards, the cheaper the price per board)
[09:29:43] <OndraSter> well first test run so as least as possible
[09:29:47] <OndraSter> I think you mentioned 10 minimum?
[09:29:58] <CapnKernel> 5 min
[09:30:01] <karlp> choose a slightly more expensive receptacle, that has mounting lugs, not just soldered onto the board, and you're good.
[09:30:15] <CapnKernel> But he may tear the whole board out!
[09:30:34] <OndraSter> hmm karlp will take a look
[09:30:41] <CapnKernel> 5 or 10?
[09:30:48] <OndraSter> prices would be how much?
[09:30:54] <OndraSter> CapnKernel, have you checked out some microUSB connectors when you were @ Shenzen?
[09:30:58] <CapnKernel> You tell me what you want, I'll look it up.
[09:31:05] <OndraSter> well 10 is standard
[09:31:08] <OndraSter> so let's go with that
[09:31:17] <CapnKernel> OndraSter: I have several good suppliers for things like that.
[09:33:02] <CapnKernel> OndraSter: 10 boards is USD28, and that includes shipping to Europe.
[09:33:19] <OndraSter> for 25x72?
[09:33:23] <CapnKernel> Yes.
[09:33:41] <CapnKernel> +USD3 for registered, +USD25 for DHL. Since you'll be paying with Paypal, please add 5%.
[09:33:51] <OndraSter> okay, after my final exams I will most likely order it (if my brother pays me his debt before that)
[09:33:55] <OndraSter> 2 weeks :)
[09:34:06] <OndraSter> I _should_ start learning already
[09:34:18] <OndraSter> thanks
[09:34:24] <CapnKernel> You're welcome.
[09:35:27] <CapnKernel> Oh, and that's 100% e-test, and there's no restrictions on board shape.
[09:36:29] <CapnKernel> Snaplines and slots are no problem.
[09:37:24] <CapnKernel> I just checked on Seeed, they have restrictions on board shape and slots, and their price of USD35 doesn't include shipping.
[09:38:16] <OndraSter> well itead has $22 for 50x100 + $5 shipping so it is the same price as yours... although I could fit two of them on one board and split them at home somehow
[09:38:39] <CapnKernel> Seeed and Itead don't like you to put two boards in one.
[09:40:34] <OndraSter> hmm
[09:46:49] <OndraSter> "The small PCB had twice the same design. It seems to be OK, as long as you don’t put anything that eases the board cut (slots, holes line or V-groove):"
[09:54:49] <chupas> so is AVR studio 5 end of life now?
[09:54:57] <chupas> should i just migrate to studio 6?
[09:55:10] <OndraSter> I'd sait
[09:55:11] <OndraSter> wait
[09:55:13] <OndraSter> I found few bugs
[09:55:24] <chupas> I still dont really understand
[09:55:30] <chupas> why they did 6 so soon after 5
[09:55:36] <OndraSter> because avr studio is no more
[09:56:00] <OndraSter> it is atmel studio now :)
[09:56:00] <chupas> Oh, ok
[09:56:00] <OndraSter> added support for ARM
[09:57:16] <theBear> no shit ? that's the only thing i heard in probly 5years that makes me consider even LOOKING at avr studio again :)
[10:34:58] <landonf> I wish they'd gone in the other direction :/ avr32 studio -> atmel studio.
[10:35:10] <landonf> So much for getting rid of my windows dependencies.
[10:50:05] <specing> landonf: What do you need windoze for?
[10:52:41] <Steffanx> I wonder if specing has a highlight on Windows
[10:52:57] <landonf> specing: avr/atmel studio. and solidworks. and altium. sigh.
[10:55:40] <asteve> xilinx has a windows only client as well, you can't program the FPGAs in a VM because of the way the usb host interface works
[10:55:50] <asteve> windows is still a requirement :/
[11:09:35] <Tom_itx> just alot more overhead just to download a program
[11:12:37] <landonf> asteve: gah.
[11:15:09] <Tom_itx> having a Z moment?
[11:15:16] <OndraSter> ZL or ZH?
[11:23:53] <specing> Z+!
[11:24:58] <Tom_itx> Z#
[11:25:59] <OndraSter> R30:R31...
[13:06:53] <OndraSter> what voltage should be ESD suppressors on USB?
[13:07:20] <OndraSter> I see some USB rated that are 150V
[13:07:22] <OndraSter> clapming
[13:07:25] <OndraSter> clamping
[13:08:11] <Tom_itx> http://search.digikey.com/us/en/products/PESD0603-240/PESD0603-240CT-ND/1813513
[13:08:46] <OndraSter> thanks
[14:37:41] <OndraSter> hmm Release Notes for AS6 say that xmega is supported with Dragon - but JTAG or PDI or both? :)
[14:37:53] <OndraSter> I remember somebody mentioning that only PDI is supported for xmegas and not JTAG
[14:37:56] <OndraSter> or the opposite
[14:38:03] <OndraSter> did they "fix" that?
[14:50:19] <Tom_itx> OndraSter, look at the studio device help and see
[14:51:09] <OndraSter> it says supported devices, but not supported methods on each peripheral..
[14:51:18] <Tom_itx> it used to
[14:51:39] <Tom_itx> i suppose they fcked that up too
[14:53:09] <OndraSter> documents @ atmel.com @ dragon's page is empty
[14:53:27] <OndraSter> abcmu! where are you :P
[14:53:44] <OndraSter> It can perform a symbolic debug on all devices with OCD with SPI, JTAG, PDI (selected devices), high voltage serial programming, parallel programming, and aWire modes, and supports debugging using SPI, JTAG, PDI interfaces.
[14:53:59] <OndraSter> does the "selected devices" mean "those that support it" or "those that we decided that will support it on Dragon"?
[14:54:02] <OndraSter> :(
[14:55:11] <RikusW> OndraSter: read the jtagice mkii page instead
[14:55:20] <RikusW> and the stk500 page too
[14:55:24] <OndraSter> how will that help me? They do not share the firmware
[14:55:31] <OndraSter> and Dragon is cheap ass version of mkii :)
[14:55:37] <OndraSter> so who knows what they cut out from it
[14:55:43] <RikusW> jtag debugging is the same for both
[14:55:49] <OndraSter> but PDI?
[14:55:50] <RikusW> and the protocols too
[14:56:07] <RikusW> there is an issue with PDI and the dragon
[14:56:25] <RikusW> does your xmega only have PDI or JTAG too ?
[14:56:26] <OndraSter> which one, I've heard many
[14:56:27] <OndraSter> both
[14:56:36] <RikusW> use JTAG instead
[14:56:39] <OndraSter> hmm
[14:56:45] <OndraSter> I can take out both pins
[14:56:54] <RikusW> PDI don't appear to work when jtag is available
[14:57:13] <OndraSter> hmm weird
[14:57:14] <RikusW> which xmega is that ?
[14:57:16] <OndraSter> 256a3u
[14:57:20] <OndraSter> anyway, Dragon seems to support both
[14:57:33] <RikusW> dragon jtag works on 128a1 but pdi don't....
[14:57:34] <OndraSter> http://clip2net.com/s/1W3xz
[14:57:47] <RikusW> I mailed atmel about it....
[14:57:49] <OndraSter> hmm
[14:58:02] <RikusW> seems PDI only xmegas might work
[14:58:09] <da78> .6.32-5-686
[14:58:11] <OndraSter> yeah, that's what I heard long time ago
[14:58:23] <OndraSter> da78, 3.x kernels are out for some time now, upgrade :P
[14:59:00] <OndraSter> darn, it is on PortB
[14:59:02] <OndraSter> right in the middle...
[14:59:49] <Tom_itx> odd bunch they are
[15:00:06] <Tom_itx> it's the protocol a week company
[15:00:47] <OndraSter> http://clip2net.com/s/1W3Aw
[15:00:50] <OndraSter> where should I put it...
[15:01:55] <OndraSter> I suppose one could connect it through the broken out pins
[15:02:02] <OndraSter> if the PDI would really not work
[15:02:27] <RikusW> email atmel about your xmega pdi support ?
[15:03:22] <OndraSter> might do
[15:04:59] <OndraSter> I did not find anything in the "known problems" with AS6 release notes
[15:05:06] <OndraSter> about PDI/JTAG problems on Dragon
[15:06:31] <RikusW> its not a problem its a feature :-P
[15:06:35] <OndraSter> :D
[15:06:39] <RikusW> a marketing feature...
[15:06:50] <RikusW> buy a jtagice mkii instead ;)
[15:06:59] <OndraSter> I lied
[15:06:59] <OndraSter> • Issue #AVRSV-1512:
[15:06:59] <OndraSter> XMEGA PDI mode on AVR Dragon does not work for the following XMEGA devices: A3/D3 - revisions
[15:06:59] <OndraSter> B, C and E
[15:07:02] <OndraSter> heh
[15:07:05] <OndraSter> more like jtagice 3
[15:07:34] <OndraSter> any reasons why to get mkII over ice 3?
[15:07:54] <RikusW> no
[15:08:20] <RikusW> or ask abc, he probably have both
[15:08:33] <OndraSter> jtagice 2: $337, jtagice 3: $214
[15:08:36] <OndraSter> dragon: $52
[15:08:38] <OndraSter> :P
[15:08:49] <OndraSter> I am cheap student, you know that!
[15:08:55] * RikusW also have a dragon
[15:09:20] <RikusW> but my own programmer is actually a little faster for ISP programming
[15:09:35] <OndraSter> • Issue #AVRSV-1512:
[15:09:35] <OndraSter> XMEGA PDI mode on AVR Dragon does not work for the following XMEGA devices: A3/D3 - revisions
[15:09:35] <OndraSter> B, C and E
[15:09:36] <OndraSter> but
[15:09:49] <OndraSter> @ datasheet of xmega:
[15:09:50] <OndraSter> Rev. A-F
[15:09:50] <OndraSter> Not sampled.
[15:10:25] <OndraSter> most likely they had problems with "not sampled" chips?
[15:10:58] <RikusW> I think dragon PDI only works on PDI only xmegas
[15:11:15] <OndraSter> I lied again, the non-U versions have revE right now
[15:11:15] <RikusW> JTAG should work
[15:11:35] <OndraSter> I suppose I could e-mail to Atmel
[15:12:50] <RikusW> The AVR Dragon only supports PDI on XMEGA A4 and XMEGA D4 devices.
[15:12:55] <RikusW> from atmel support
[15:16:41] <OndraSter> oh
[15:16:41] <OndraSter> hmm
[15:16:42] <OndraSter> darn
[15:16:54] <OndraSter> so I can get rid of that PDI header and put there JTAG instead
[15:16:54] <OndraSter> darn
[15:17:01] <OndraSter> RikusW, that depends how new that page is
[15:17:06] <OndraSter> because *U chips are fairly new
[15:17:11] <Steffann> mail them OndraSter
[15:17:17] <Steffann> or perhaps ask abcminiuser
[15:17:17] <OndraSter> doing so right now
[15:17:32] <OndraSter> Your request has been registered blabla
[15:17:33] <OndraSter> done
[15:17:37] <OndraSter> thanks guys :D
[15:17:58] <OndraSter> I thought that PDI was faster than JTAG
[15:18:05] <Steffann> It is
[15:18:12] <Steffann> afaik
[15:18:14] <OndraSter> also, AVRISP mkII can work with PDI but not JTAG
[15:18:15] <Kevin`> it shouldn't be faster
[15:18:33] <OndraSter> well the PDI can go upto 32MHz itself
[15:18:34] <Kevin`> jtag is spi, it can run faster than the cpu pretty easily
[15:18:49] <Kevin`> (faster than the cpu can handle data, that is.. not limited by jtag)
[15:18:55] <OndraSter> PDI is half duplex SPI :P
[15:19:04] <Kevin`> spi is full duplex spi ;p
[15:19:16] <OndraSter> jtag is full duplex spi*
[15:19:33] <Steffann> Doesn'tt PDI skip the overhead of TAP etc. ?
[15:19:37] <Steffann> *the TAP
[15:19:40] <OndraSter> most likely
[15:19:54] <OndraSter> since it can't be chained
[15:20:15] <Steffann> For ARM you have this fancy SWD, I wonder how similar it is to PDI
[15:20:23] <Steffann> SWD is also clk+io
[15:24:53] <OndraSter> duhh the thing is, if I get rid of PDI and put there JTAG, it won't be programmable with AVRISP
[15:24:55] <OndraSter> arrgh
[15:24:59] <OndraSter> and I can't fit both on the PCB
[15:26:14] <OndraSter> we need new JTAG header... this 10 pins is too big, 7 pins would be enough...
[15:26:18] <OndraSter> custom header eh?
[15:26:21] <Steffann> Stay with PDI unless Atmel tells you it doesn't work..
[15:26:44] <Steffann> OndraSter .. 1.27mm spacing
[15:27:02] <OndraSter> well one would need to enclose also new cable with it
[15:27:42] <Steffann> It seems to be a "standard"
http://www.keil.com/coresight/connectors.asp
[15:28:15] <OndraSter> huh
[15:28:21] <Steffann> *there
[15:28:23] <OndraSter> it is also flipped there
[15:28:49] <OndraSter> well it is weird alltogether :D
[15:29:03] <Steffann> Just use PDI :P
[15:29:36] <OndraSter> :D
[15:29:39] <OndraSter> but if it won't work :(
[15:33:28] <RikusW> Steffann: the Programming part of PDI is documented, not the debug part
[15:33:34] <RikusW> its a byte based protocol
[15:33:53] <RikusW> and xmega JTAG is just PDI encapsulated inside JTAG
[15:34:31] <RikusW> abc used a synchronous uart for pdi
[16:17:58] <OndraSter> hah I set JTAG speed from 250kHz to 1.7MHz and it is waaay faster
[16:21:26] <prpplague> why such a low speed?
[16:21:35] <OndraSter> well 250kHz is maximum
[16:21:38] <OndraSter> and 2MHz is maximum
[16:21:59] <prpplague> OndraSter: max of the device or of the jtag dongle?
[16:22:12] <OndraSter> well AVR Studio allows me to set only upto 2MHz
[16:22:16] <OndraSter> the slider won't go after that
[16:22:34] <prpplague> ahh, probably software related then
[16:22:43] <OndraSter> because the internal oscillator is in default 1MHz
[16:22:45] <OndraSter> set by FUSEs
[16:22:53] <OndraSter> and JTAG is supposed to be <= 1/4 of master clock
[16:23:14] <OndraSter> so for builtin 8MHz oscillator you can use only upto 2MHz supposedly
[16:23:40] <OndraSter> funny thing - I was running my atmega128a on 1MHz internal oscillator yet I could do 2MHz JTAG... either it has fallen back to 250kHz on its own or.. magic!
[16:24:11] <prpplague> OndraSter: ahh right, i forget about the low clock freqs for avr, hehe
[16:24:23] <OndraSter> :D
[17:06:37] <Steffann> I knew that RikusW
[17:06:40] <Steffann> Oh, he left
