Back
[02:02:15] <carp3> is it better to use the GND pin on the right side of atmega8 beside AREF pin as return path for external reference or it doesn't matter ?
[02:07:05] <Casper> that ground on some avr is called "AGND"
[07:47:07] <evil_dan2wik> Anyone know how to convert a USBASP into a PDI programmer?
[07:47:27] <evil_dan2wik> I tried but I just end up with avrdude: initialization failed, rc=-1
[07:47:27] <evil_dan2wik> Double check connections and try again, or use -F to override
[07:47:27] <evil_dan2wik> this check.
[08:21:02] <hetii> evil_dan2wik: look at here :
http://www.elektroda.pl/rtvforum/topic2359783.html
[08:22:18] <hetii> evil_dan2wik its usbasp with PDI support
[08:22:37] <hetii> evil_dan2wik: if you will had trouble with Polish language let me know.
[08:28:36] <evil_dan2wik> hetii, still getting the same error.
[08:29:20] <hetii> didi you use code from the site that i give you ?
[08:29:49] <hetii> BTW test your hardware as regular usbasp programmer and then as pdi
[08:29:54] <hetii> if success
[08:30:22] <evil_dan2wik> hetii, yes.
[08:30:24] <evil_dan2wik> I used code from the site, the ASP was working fine before this.
[08:30:29] <evil_dan2wik> The only thing I am not sure works is the xmega.
[08:31:36] <hetii> well check connection then, maybe play with option -B (1,20,100,200...) for avrdude that will slow sck signal
[08:31:57] <evil_dan2wik> hetii, sck isn't the signal.
[08:32:46] <hetii> here is used for some purposes
http://obrazki.elektroda.pl/8962526200_1345075815.png
[08:33:54] <hetii> Also keep in mind that maybe you need voltage translator like here if your target mcu works with 3v3 volts:
http://obrazki.elektroda.pl/2407137700_1345075790.png
[08:34:28] <evil_dan2wik> do I *need* it?
[08:35:40] <hetii> No clue, ATxmega works at 3v3 and usbasp at 5v
[08:35:55] <hetii> don`t know what taget you use
[08:36:00] <hetii> *target
[09:38:37] <Jartza> ps. thanks for everybody who helped me with comments on my first pcb :)
[09:38:47] <Jartza> going to place an order of it
[09:38:52] <N1njaneer> Awesome!
[09:42:39] * Thrashbarg is working on a vacuum tube pong video game
[09:42:43] <Thrashbarg> I must be insane ._.
[09:42:54] <Jartza> sounds cool :)
[09:43:06] <Thrashbarg> yeah... just tricky
[09:43:12] <Jartza> http://gerblook.org/pcb/gZEZTT4v4w8vX9fWtsYPak
[09:43:20] <Jartza> that's the completed boad
[09:43:23] <Jartza> board, even
[09:43:27] <Thrashbarg> nice
[09:44:29] <Jartza> I think I would've had much harder time without helpful people here and on #hackvana :)
[09:44:53] <Thrashbarg> why spend ages figuring things out yourself when you can just ask? :P
[09:57:38] <Jartza> Thrashbarg: that's true, but on some channels people aren't that nice towards noobs like me :)
[10:02:41] <Thrashbarg> Jartza: indeed
[10:12:24] <__heisenbug> is it possible to force a very long instruction word manually?
[10:14:31] <__heisenbug> and if it does, how would it look like in c or asm?
[10:21:43] <twnqx> what
[10:22:00] <twnqx> vliw is a processor architecture, to my understanding
[10:22:12] <twnqx> so... youÄd have to design a processor in that way
[10:23:48] <__heisenbug> vliw is getting generated by a compiler, it sets some small instructions together and builds one big instruction, where all small instructions are parallel executed
[10:24:12] <twnqx> that requires a supersclaar processor
[10:24:19] <__heisenbug> so i dont think it is a processor architecture
[10:24:19] <twnqx> which the AVR is not.
[10:24:25] <twnqx> yes it is
[10:24:54] <twnqx> in order to decode such an instruction, the cpu fetch/decode unit must be able to handle it
[10:25:06] <twnqx> normal CPUs aren't
[10:25:19] <twnqx> iirc Itanium is an exmaple architecture for VLIW
[10:25:35] <twnqx> can't recall any other from top of my head
[10:25:47] <__heisenbug> no a superscalar processor needs some specials in the mikroarchitecture its not a vliw
[10:25:55] <__heisenbug> am i really that wrong?
[10:26:18] <twnqx> a superscalar processor just means that it is able to execute more than one command in parallel
[10:26:41] <twnqx> by virtue of multiple pipelines/pipeline parts
[10:26:46] <__heisenbug> yes, right, but it needs seperated memory and busses etc
[10:26:50] <twnqx> s/command/instruction/
[10:26:53] <twnqx> nah
[10:27:00] <twnqx> x86 is superscalar
[10:27:16] <twnqx> even the old single core ones are.
[10:27:41] <__heisenbug> hm so my prof tells me bullshit? xD
[10:27:47] <twnqx> possible
[10:28:15] <twnqx> the other possibility is that you took some parts out of context
[10:28:34] <twnqx> oviously only one unit can access a bus at one time
[10:28:49] <twnqx> but register->register operations are independent of example memory bus
[10:29:38] <twnqx> that's where first superscalar designs came from, separating load/store and integer units, for example
[10:29:46] <__heisenbug> he says: instead of a superscalar architecture, which is getting new arranged through hardware, the compiler can do this too, to get more exectionspeed
[10:30:09] <twnqx> that is the theory behind vliw designs,y es
[10:30:37] <twnqx> but that does not mean that any given architecture can suddenly use VLIW.
[10:31:20] <twnqx> it means you can design a CPU, leaving the parallizing to the compiler which might have a better overview than statistics and guesswork a superscalar cpu does
[10:31:30] <__heisenbug> hm yeah sounds logical
[10:32:06] <__heisenbug> okay thanks dude!
[10:32:21] <twnqx> you're welcome
[10:36:34] <mr_boo> is it equally simple to use a hardware timer to just read the contents to measure time rather than to trigger an interrupt from the register?
[15:08:47] <rewolff1> Re: "VLIW": twnqx: can't recall any other from top of my head
[15:08:47] <rewolff1> I studied at a group at the university that designed a VLIW processor in the 1990's...
[18:20:57] <hetii> Q: Is it possible to have canbus and dmx on the same physical bus ?
[18:23:26] <Lambda_Aurigae> likely not.
[18:23:34] <Lambda_Aurigae> dmx is rs485 as I recall.
[18:24:11] <hetii> yes but boath are a differential line
[18:24:31] <Lambda_Aurigae> yes...and what would stop one from interfering with the other?
[18:25:23] <hetii> dmx will suffer a bit but suppose canbus should works fine.
[18:27:03] <Lambda_Aurigae> I suspect they will both interfere with each other.
[18:27:12] <hetii> will need do some real test and check if thats make any sense :)
[18:29:03] <Lambda_Aurigae> they terminate differently too.
[18:31:07] <Lambda_Aurigae> can appears to be all positive voltage where rs485 is +12/-7
[18:31:27] <Lambda_Aurigae> that -7V might damage can devices.
[18:31:51] <hetii> not really, as I meet its one single 120 ohm resistor or in "T" mode like 2x60ohm and capacitor.
[18:32:18] <Lambda_Aurigae> http://en.wikipedia.org/wiki/File:Rs485-bias-termination.svg
[18:32:40] <Lambda_Aurigae> it's biased and terminated on rs485.
[18:33:02] <Lambda_Aurigae> I don't think you bias canbus
[18:39:45] <hetii> hmm with termination I saw different ways. But interesting point is with this voltage range
[18:40:03] <hetii> in my LPC11C24 the can indeed works at 5v range
[18:41:05] <hetii> but in datasheept i found parameter that is called: IO(rec) recessive output current and its -5ma till +5mA
[18:41:24] <hetii> but in condition V CANH = V CANL = 27 V to +32 V
[18:41:33] <hetii> *-27v
[20:40:17] <evil_dan2wik> Do atmel chips come with anything programmed on them when sent as a sample?
[20:42:31] <learath> I'd guess no
[20:47:05] <Casper> evil_dan2wik: nope
[20:48:01] <Casper> bbl
[21:31:00] <Tom_itx> evil_dan2wik only the USB ones with a bootloader
[22:44:42] <evil_dan2wik> Tom_itx, this is a USB one, it won't connect to USB
[22:44:45] <evil_dan2wik> not even any dmesg
[23:00:18] <evil_dan2wik> What was the requirements for using an xmega?
[23:00:19] <evil_dan2wik> It looks like it just needs power.
[23:02:39] <Casper> evil_dan2wik: if you didn't flash anything, then it's normal that it won't connect to usb
[23:02:51] <evil_dan2wik> Casper, but it isn't connecting to anything
[23:03:03] <evil_dan2wik> I don't know if the chip is busted or my programmer
[23:03:22] <evil_dan2wik> my programmer works fine in other modes but this is the first time I have needed to use PDI
[23:04:23] <Casper> I wonder...
[23:04:29] <Casper> do you need a crystal on xmega?
[23:04:37] <evil_dan2wik> I do not know.
[23:04:59] <Casper> might want to investigate
[23:05:03] <evil_dan2wik> It says it starts up with the 2MHz crystal and then switches to whatever other crystal is selected.
[23:05:33] <evil_dan2wik> I have seen other people with other xMegas who just have power and the same programmer and configuration.
