Back
[00:03:56] <geckosenator> so everything breaks horribly with the new gcc
[02:14:28] <ferdna> guys can the goverment... track gps devices?
[02:15:12] <ferdna> like for example i create an arduino with gps capabilities will they be able to log its gps?
[02:16:41] <geckosenator> yes if you can see them they can see you
[02:19:04] <ferdna> geckosenator, oh
[02:20:23] <inflex> wow, it actually worked
[02:20:29] <inflex> my negative regulator system actually didn't blow up
[02:21:24] <ferdna> cool
[02:21:36] <inflex> yeah, I'm shocked
[02:22:10] <inflex> basically floating my digital bits 3.3V relative to the high-side (so +ve = Vcc, GND = Vcc -3.3
[02:22:15] <geckosenator> it shocked you instead?
[02:23:12] <geckosenator> you touched the wires?
[02:25:05] <ferdna> lol
[02:47:04] <inflex> humour aside, i was surprised because it was a bit of an odd configuration
[02:47:11] <inflex> at least I knew everything else should work
[02:50:58] <kd0mto> /quit sleep
[03:01:52] <soul-d> oh we humans are so smart
[03:02:19] <soul-d> lets take the h5n1 virus and if it doesn't spread from human to human why not geneticly alter it that it does
[03:04:08] <sh00p> I have this atmega on a hole-card, it doesnt act deterministic, if i power it on one time it might do something, if i turn it off and then on the next it does another
[03:04:14] <sh00p> even though all the other conditions are the same
[03:04:18] <sh00p> anybody seen this before?
[03:05:13] <geckosenator> yeah
[03:05:28] <geckosenator> you are sampling from the adc pins that are floating
[03:06:06] <sh00p> hmm, am I ? :o
[03:06:10] <sh00p> i do use the adc
[03:06:28] <sh00p> should I pull the adc pins that aren't used high or something?
[03:06:48] <sh00p> or can I circumvent this in code somehow?
[03:10:51] <sh00p> geckosenator, an you help me out?
[03:11:12] <geckosenator> probably not
[03:12:37] <sh00p> but by "sampling from floating adc pins" you're not talking about pins that I aren't using right
[03:12:50] <sh00p> it's the pins i'm actually using that are floating
[04:06:41] <amee2k> ziph: i just found a full datasheet for them and it says the heatsink pad is "electrically neutral", which i'd interpret connecting it to cathode or anode potential, or any potential inbetween is acceptable
[04:06:59] <amee2k> the online shop was linking to some outdated or castrated version of it
[04:08:04] <ziph> I'd ground it, heatsinks aren't half bad antennas some times. ;)
[04:08:14] <inflex> heh
[04:08:28] * inflex goes off to try make a nice 0.002R shunt
[04:09:46] <inflex> have to say, when you have a LOM, making precise shunts is easy :D
[04:11:00] <amee2k> the led is on the low side anyway so i can just expand the ground plane to cover the heatsink pad too
[04:12:08] <amee2k> inflex: i'd give my left nut for one of these cool precision impedance bridges and a place to store it
[04:14:40] <inflex> amee2k: well, store it in the place where your nut used to be ;)
[04:14:59] <amee2k> ...
[04:15:32] <inflex> Still having my own LOM has been bloody expensive
[04:15:37] <inflex> since I've taken ages to make it
[04:15:40] <amee2k> i said nut, these aren't exactly the same size. i'm not giving away my ... nevermind
[04:15:49] <inflex> *laugh*
[04:16:04] <amee2k> you made your own lom?
[04:16:38] <inflex> yes
[04:16:44] <amee2k> nice :)
[04:16:56] <inflex> 15th revision now... 10 micro-ohm resolution
[04:17:05] <amee2k> 0.0
[04:17:16] <amee2k> how does it work?
[04:17:25] <inflex> http://nqrc.com/?vp=PLD-LOM7
[04:18:24] <inflex> uses a 24-bit ADC to measure the voltage across the unknown and also a known internal reference
[04:18:30] <inflex> then just uses a bit of ratiometric math to do therest
[04:18:32] <inflex> the rest
[04:20:11] <amee2k> i see
[04:22:21] <inflex> mmm... just realised, I need a few 50V rated caps
[04:25:19] <amee2k> nice one :)
[04:49:25] <earthshine> o/
[04:58:09] <errordeveloper> hey ..
[04:58:36] <errordeveloper> I have got a Bus Blaster from Dangerous Prototypes
[04:59:00] <errordeveloper> but I wonder if I could use it as AVR ISP?
[05:00:50] <errordeveloper> I suppose with the CPLD on board, I can achive that quite esily
[05:01:04] <soul-d> stupid question how do you read ->
http://tinyurl.com/cfymlqf only mine says 1J or 22J
[05:01:10] <soul-d> in place of 04
[05:03:14] <errordeveloper> I know there is a plenty of DIY ISP code, but a pointer to a decent implementation would be appreciated :)
[05:04:12] <inflex> ?
[05:04:16] <inflex> what about the USBtiny?
[05:04:45] <RikusW> errordeveloper: iirc avrdude supports it already
[05:05:33] <RikusW> err the bus pirate anyways
[05:11:48] <errordeveloper> RikusW: ok, i'll check :)
[05:13:37] <errordeveloper> pwd
[05:13:39] <errordeveloper> oops
[05:13:41] <errordeveloper> :)
[05:13:57] <RikusW> /freenode/avr ;)
[05:18:06] <w|zzy> For those following raspberry pi
http://wiki.qt-project.org/Qt_RaspberryPi/Device_program
[05:19:18] <errordeveloper> w|zzy: kewl :)
[05:19:43] <w|zzy> It is indeed. Nokia is paying ;)
[05:20:49] <errordeveloper> lol., it's just a few quid really
[05:21:37] <errordeveloper> dunno, I thought Nokia has given up on embedded linux
[05:21:56] <errordeveloper> however, may be they just wanna sell qt now :)
[05:22:10] <inflex> ooookay, let's see if fail2ban does its job nicely
[05:22:37] <w|zzy> qt is very capable... Ive been following the raspberry pi more than qt though
[05:23:20] <errordeveloper> sure .. I just wonder what about the new deal - Nokia -> win8
[05:23:51] <w|zzy> I ban 2-3 ip addresses a day.
[05:24:10] <w|zzy> Its a pity that... nokia had some potential there.
[05:25:45] <RikusW> w|zzy: ban ip's ? what for ?
[05:25:56] <w|zzy> inflex was talking about fail2ban
[05:26:06] <w|zzy> Sorry. Should have put context there.
[05:26:26] <inflex> w|zzy: should be interesting - I get a lot of scripted scans
[05:27:01] <w|zzy> I used to get more... First month of running fail2ban my logs were massive
[05:28:45] <w|zzy> Not sure why.. It would be naive of me to believe it was fail2bans doing.
[05:31:42] <inflex> w|zzy: I've noticed the scans / attempts seem to go in waves here
[05:31:45] <inflex> but not sure that's the same
[05:38:40] <errordeveloper> w|zzy: cfs is the quite good
[05:39:00] <errordeveloper> you just get the tarball from here -
http://configserver.com/cp/csf.html
[05:39:35] <errordeveloper> then run install script - it does self-update also, if you ask it to
[05:39:50] <errordeveloper> there is csf and lfd
[05:40:05] <errordeveloper> lfd track suspecious stuff, i.e. ssh login attempts
[05:40:19] <errordeveloper> it's very easy to use, yet robust
[05:42:09] <sh00p> I have this atmega on a hole-card, it doesnt act deterministic, if i power it on one time it might do something, if i turn it off and then on the next it does another
[05:42:13] <sh00p> even though all the other conditions are the same
[05:42:16] <sh00p> anybody seen this before?
[05:42:27] <sh00p> some guy told me it could be due to floating adc pin
[05:42:37] <sh00p> I am using the adc, but I don't see how this behaviour comes from that
[05:42:54] <RikusW> or uninitialized variable ;)
[05:43:02] <RikusW> that one bit me before...
[05:43:04] <sh00p> i didn't get this behaviour when the mcu was on the sdk500
[05:43:15] <sh00p> hmm
[05:44:21] <sh00p> I've put delays on the top of main()
[05:44:26] <sh00p> but it doesn't seem to fix it
[05:46:18] <sh00p> "In order to set up internal register properly to their initial state after power is applied, there is reset on power-on needed. Atmega8 have several options"
[05:46:21] <sh00p> could it be something like that?
[05:47:05] <RikusW> rather som variable...
[05:47:37] <sh00p> they don't get initialized to zero by default?
[05:47:46] <sh00p> although i have manually initialized most of them to zero afaik
[05:48:09] <karlp> global statics do, nothing else does.
[06:11:55] <scuzzy> hey RikusW
[06:15:26] <RikusW> hi scuzzy
[06:26:25] <sh00p> is there any sense in putting an LC-filter on ADC inputs?
[06:27:27] <RikusW> avcc can use LC 10uH + 100nF
[06:28:10] <sh00p> yeah, avcc has that
[06:28:13] <sh00p> as well as vcc
[06:28:18] <sh00p> they're actually on the same filter
[06:28:21] <sh00p> is that weird?
[06:28:31] <sh00p> and aref too
[06:28:50] <RikusW> avcc should be on a different filter than vcc
[06:29:34] <sh00p> same value for LC's though?
[06:29:44] <RikusW> yes
[06:30:25] <sh00p> but avcc and vcc can ultimately be powered from same source?
[06:39:19] <inflex> tyes
[06:39:39] <errordeveloper> ah, it should be quite easy in fact
[06:39:42] <sh00p> http://sh00p.mine.nu/~fredrik/schematic2.png
[06:39:47] <sh00p> that's how I've done it now
[06:40:03] <sh00p> does that look weird?
[06:40:17] <errordeveloper> cause the Bus Blaster has ft2232 which can do SPI
[06:40:31] <errordeveloper> and is actually supported by avrftdi driver
[06:42:41] <sh00p> RikusW, could you have a look at my schematic?
[06:48:20] <RikusW> sh00p: rahter use 18pF caps for the crystal
[06:49:05] <sh00p> ok
[06:49:24] <sh00p> do you see any critical errors on there? cause my software seems to behave differently every time i turn it on
[06:50:34] <sh00p> and i'm sure there are no uninitialized variables
[06:59:13] <RikusW> sh00p: are you using avrstudio ?
[07:00:36] <sh00p> avrdude
[07:00:41] <sh00p> with eclipse and avr-gcc
[07:01:00] <sh00p> but I have avrstudio in vmware, have set some fuses from there
[07:01:27] <RikusW> I have a jtagice clone it you want ?
[07:01:29] <RikusW> for free
[07:02:50] <sh00p> whats a jtagice clone?
[07:03:00] <Tom_itx> debugger
[07:03:24] <sh00p> ah, it's a piece of hardware?
[07:04:01] <RikusW> yes
[07:04:13] <RikusW> do you have a m32u2 or m168 or m328 around ?
[07:04:21] <RikusW> you can easily build it
[07:04:26] <Tom_itx> yes
[07:04:28] <Tom_itx> i have a few
[07:04:47] <RikusW> Tom_itx: a few ?! more like a million ;)
[07:05:01] <sh00p> I have a couple of atmega32-pu
[07:06:32] <RikusW> you could compile it to run on that too
[07:09:20] <sh00p> hmm, I wish I had time for it
[07:09:25] <sh00p> going on vacation tomorrow
[07:09:33] <RikusW> sh00p:
http://sites.google.com/site/megau2s/home/jtag
[07:09:44] <sh00p> this thing i'm building is supposed to be sold and installed while I'm away
[07:09:49] <sh00p> I'm sort of in the shit :)
[07:11:31] <carp3> is it possible to communicate with uC using programmer ? avrisp mkii for example
[07:15:39] <Tom_itx> to program it yes
[07:16:24] <Tom_itx> sh00p, you might try setting all the unused adc pins as output
[07:16:33] <carp3> no i mean sending something using SPI . (by connecting Reset pin to CS )
[07:17:26] <Tom_itx> i doubt it since the programmer has a certain protocol it uses
[07:18:24] <carp3> Protocol is hardcoded in Hardware or it's software that do the job (PC Side)? like STK200/300
[07:18:27] <RikusW> Tom_itx: it should have been possible, I added some custom commands to my programmer
[07:18:43] <RikusW> like turn on/off SPI and then used the standard SPI_MULTI command
[07:18:54] <Tom_itx> ok but he said avrisp mkii
[07:19:16] <RikusW> yeah
[07:19:36] <RikusW> carp3: its in the mkii firmware...
[07:19:44] <RikusW> no changing that
[07:20:21] <carp3> and there is no openFirmware for mkii ?
[07:20:29] <RikusW> too bad atmel didn't think of it
[07:20:33] <RikusW> carp3: no
[07:20:40] <RikusW> its encrypted anyways
[07:20:47] <RikusW> the bootloader decrypts it
[07:28:13] <carp3> what's MKII uC? mega128 ?
[07:34:00] <RikusW> won't help you
[07:34:22] <RikusW> its bootloader expects encrypted fw
[07:35:28] <carp3> but it's possible to write firmware from scratch?
[07:35:45] <RikusW> no
[07:36:06] <RikusW> you don't really want to mess with the mkii's fw
[07:36:14] <RikusW> you'll just break it...
[07:36:31] <RikusW> don't know if it has a jtag/isp header inside
[07:36:45] <carp3> i have a bricked mkii anyway. so it's OK
[07:37:15] <RikusW> how did you brick it ?
[07:37:39] <RikusW> fw update failed ?
[07:38:58] <carp3> i don't know. LED is always red ( and wire is OK )
[07:42:30] <RikusW> carp3: even with target power connected ?
[07:42:42] <carp3> yes
[07:44:00] <RikusW> have you tried manual fw upgrade ?
[07:44:18] <carp3> by shorting pins inside mkii ? yes
[07:44:24] <RikusW> inside there is a holes for a pinheader
[07:44:43] <RikusW> pin 1-3...
[07:46:04] <RikusW> that might actually be a jtag header....
[07:46:26] <carp3> i think something fried inside.. doesnt matter i have 2 more.
[08:01:14] <RikusW> Red Idle - No target power
[08:01:21] <RikusW> so not sensing VTG
[09:01:19] <RikusW> hi vectory, got that 64 bit inf ?
[10:13:33] <sh00p> thanks for all the help guys, now I'm off to thailand
[10:48:37] <ben1066> Hey
[10:48:46] <ben1066> So I decided to switch to a 168, more speed, more space
[10:48:52] <ben1066> Still got UART issues
[10:48:59] <ben1066> I added an LED on the uart tx pin
[10:49:03] <ben1066> and it stays on
[10:49:10] <ben1066> However my multimeter tells me its 2.5V
[10:49:19] <ben1066> So it appears to be sending something
[10:52:17] <ben1066> for FOSC 16mhz is 16000000
[10:52:18] <ben1066> yes?
[10:54:00] <ben1066> also
[10:54:13] <ben1066> is it normal that touching the xtal pins without a crystal makes it reallly fast :P
[10:54:31] <ben1066> or slower
[10:55:11] <carp3> do LED test again to makesure Clock is OK.
[10:59:27] <ben1066> also
[10:59:31] <ben1066> I know why atmega wastn working
[10:59:39] <ben1066> reset was being held low with avrisp
[11:00:14] <carp3> heh
[11:00:32] <carp3> avrisp mkii ?
[11:00:36] <ben1066> ja
[11:00:46] <ben1066> also
[11:00:49] <ben1066> Im using referance code
[11:00:51] <ben1066> #define MYUBRR FOSC/16/BAUD-1
[11:00:54] <ben1066> so it should work
[11:00:58] <ben1066> FOSC is 16000000
[11:01:28] <carp3> define FOSC as 16000000UL
[11:01:50] <ben1066> should baud be 57600UL
[11:02:07] <carp3> yes
[11:03:30] <carp3> also 57600 is not good at 16MHz
[11:03:57] <carp3> use 56000
[11:04:00] <ben1066> It seemed to work well enough with the arduino
[11:04:16] <ben1066> or Ill run at 20mhz
[11:04:16] <ben1066> :p
[11:06:25] <ben1066> I wish I had a logic analyser
[11:06:27] <ben1066> Really badly :p
[11:06:45] <ben1066> anyone here got one of those salea logic thingys, are they any good?
[11:10:30] <ben1066> hm this is very wierd
[11:24:48] <Casper> you can also use wormfood avr baud calculator
[11:25:00] <Casper> to see what speed work at what clock
[11:25:02] <ben1066> this is referance
[11:25:05] <ben1066> I dont understand
[11:25:09] <Casper> and what UBR to use
[11:25:40] <Casper> http://www.wormfood.net/avrbaudcalc.php
[11:25:47] <ben1066> i get 20.7 for the 20mhz clock
[11:26:07] <Casper> use 21
[11:26:18] <ben1066> Casper: thing is
[11:26:21] <Casper> you'll be 1.4% off, which is tolerable
[11:26:23] <ben1066> Im not even getting garbage
[11:26:25] <ben1066> Im getting nothing
[11:26:35] <ben1066> im send U and I get nothing :S
[11:26:42] <ben1066> while(1){uart_putc('U');}
[11:27:10] <Casper> have you actually enable the outut?
[11:27:10] <carp3> upload your hex file
[11:27:23] <Casper> and try to set the pin to output too
[11:27:32] <ben1066> Its outputting
[11:27:33] <ben1066> I can see that
[11:27:43] <ben1066> the voltage is 2.11 about half my supply
[11:27:56] <ben1066> with a DMM anyway
[11:28:24] <ben1066> http://dl.dropbox.com/u/11197643/168fail.hex
[11:28:30] <ben1066> remeber its 168 now not 8
[11:28:47] <ben1066> Im getting an output, my other avr just isnt seeing it :S
[11:29:07] <carp3> 16MHZ and 57600 ?
[11:29:21] <Casper> ben1066: be sure that your terminal is set to 56700 8N1 no flow control
[11:29:38] <ben1066> 20mhz
[11:29:41] <ben1066> 57600
[11:29:48] <Casper> err ya
[11:29:50] <ben1066> Casper: Im not going to a terminal, im going to another AVR
[11:30:05] <ben1066> That avr is expecting 57600, 8N1
[11:30:11] <ben1066> Just to check
[11:30:17] <Casper> same xtal speed?
[11:30:31] <ben1066> http://pastebin.com/Ev2Pn6pM
[11:30:42] <ben1066> its running on a 16mhz xtal
[11:30:47] <ben1066> but I dont see the issue with that
[11:31:22] <Casper> math time
[11:31:35] <ben1066> my arduino works fine with it
[11:31:38] <ben1066> it can send data
[11:31:52] <ben1066> Serial.Enable(57600); Serial.Send("Hello World");
[11:31:54] <ben1066> or whatever
[11:31:56] <carp3> ben1066: should output f ? i geting lots of f fffffffffffff
[11:31:59] <ben1066> is that pastebin corrent?
[11:32:06] <ben1066> carp3: should be UUUUUUUUUUU
[11:33:26] <Casper> 16MHz = 58823
[11:33:48] <ben1066> surely slight error shouldnt make a difference?
[11:34:24] <ben1066> it didnt work at 16mhz either
[11:34:25] <ben1066> :p
[11:34:36] <Casper> 20MHz = 58818
[11:34:58] <Casper> closer than I tought, it's almost a perfect match
[11:35:29] <ben1066> still
[11:35:34] <ben1066> Im not even getting garbage
[11:35:40] <Casper> so be sure that you enabled the rx and tx, that you have common ground, that the rx is set for input and the tx is set for output
[11:35:42] <ben1066> if the baud was wrong id get garbasge
[11:35:48] <Casper> not always
[11:36:03] <ben1066> http://pastebin.com/Ev2Pn6pM
[11:36:07] <ben1066> is my init code
[11:36:12] <Casper> can't che ck code, I'm at work
[11:36:23] <ben1066> Casper: you cant open links? :p
[11:36:41] <ben1066> UCSR0C = ((1<<UCSZ01)|(1<<UCSZ00));
[11:36:50] <ben1066> is that right for 8 bit, 1 stop, no parity
[11:37:04] <Casper> can't check datasheet
[11:37:15] <Casper> and the boss is about to "break in"
[11:37:29] <Steffanx> Haha
[11:37:33] <ben1066> Lol
[11:37:44] <digi001> Hello. I am looking to send serial data from my Micro to a computer for plotting. So far I setup all the control registers and have established communications, but what is the best way now to transmit the data?
[11:37:45] <Casper> he's on the road
[11:37:46] <ben1066> Why wouldnt you get garbage?
[11:37:51] <digi001> Im sending floating point numbers
[11:37:53] <digi001> 32 bit
[11:38:01] <ben1066> digi001: send them as hex?
[11:38:03] <Casper> he was going to do grocery shopping, then come to make a phone call
[11:38:03] <digi001> but they dont need to be recieved as floating point
[11:38:12] <ben1066> like, raw hex
[11:38:21] <Casper> ben1066: wait... do you use a serial library?
[11:38:25] <ben1066> Nope
[11:38:26] <carp3> ben1066: hmmm you are outputing a square wave @2.4khz. nothing more.
[11:38:30] <ben1066> I use my own functions
[11:38:32] <Casper> ahAH!
[11:38:41] <ben1066> carp3: Ja, im looping UUUUUUUUUU
[11:38:53] <digi001> ok but if i send as raw hex. ill need to send 4 transfers of 8 bits each?
[11:39:03] <digi001> how do i reassemble that on the other end, and keep track of it?
[11:39:13] <ben1066> Have a buffer
[11:39:14] <Casper> you know that printf do not connect to the uart directly?
[11:39:18] <ben1066> and a start and stop bit
[11:39:21] <ben1066> I dont use printf
[11:39:28] <Kevin`> digi001: how will you tell on the recieving end which byte you are at
[11:39:30] <ben1066> I use uart_putc uart_puts
[11:39:36] <digi001> ok for start and stop bit, do you mean just like a special charachter?
[11:39:45] <digi001> each transfer has a start and stop bit already
[11:39:46] <ben1066> digi001: and a start bit? you dont need a stop bit
[11:39:50] <ben1066> just wait for a start bit
[11:39:55] <ben1066> then count off 4
[11:40:03] <Casper> digi001: that's the fun thing.... serial is 8 bits (plus a start and stop bit)
[11:40:07] <ben1066> if you get a start bit first, you know you have to start again :p
[11:40:18] <Casper> so you need to keep track of which part you are receiving
[11:40:19] <ben1066> start magic number more correct I geuss
[11:40:48] <digi001> but if im sending raw data then there could be a chance my data has the same magic number as the start byte
[11:40:49] <Kevin`> digi001: do you have any reserved data? otherwise it might be easiest to use breaks between numbers
[11:40:52] <ben1066> anyone any idea with my problem?
[11:41:26] <ben1066> digi001: have your datalogging program request data then
[11:42:04] <digi001> i dont need the precision really, i just want a basic graph. The output is a number oscillating between -3 and 3. WOuld it make sense to convert the number to a string with 2 or 3 decimal places and send that?
[11:42:15] <digi001> then i can have start and stop characters
[11:42:20] <ben1066> No
[11:42:27] <ben1066> Thats much less inefficient :p
[11:42:32] <Kevin`> digi001: the last time I did this I sent one nybble per byte to have extra data for framing, but that was because of a deep error somewhere that was causing problems
[11:42:33] <ben1066> effecient*
[11:42:50] <digi001> ah ok kevin. thats an idea
[11:43:23] <ben1066> Casper: carp3 any ideas?
[11:43:29] <digi001> ben1066: what do u mean datalogging program request data?
[11:43:34] <digi001> like with the RTS CTS?
[11:43:41] <ben1066> how are you capturing data
[11:43:48] <Kevin`> digi001: why do you need floating point? floating point math on avr8 is really slow because there's no instructions for it
[11:43:51] <ben1066> cause you could send 1 set of 8 bits as a request
[11:43:54] <digi001> well, to start i was just using hyperterminal
[11:44:06] <ben1066> Kevin`: Add FPU :p
[11:44:07] <digi001> but then i was gonna try and use matlab captureing tools, or maybe labview
[11:44:54] <digi001> Kevin: well this is actually a bigger AVR
[11:45:11] <Kevin`> digi001: alternately, send the data like as ascii with linefeeds. it's inneficient and takes a lot of flash space for printf, but you can read it, and include debug output in the same data
[11:45:23] <Kevin`> digi001: avr32?
[11:45:31] <digi001> TI TMS320F28335
[11:45:46] <Kevin`> that's not an avr. whatever ;p
[11:46:06] <digi001> sorry ha. but i alsio use avr. :)
[11:46:18] <digi001> just not for this problem
[11:46:26] <carp3> ben1066: your output ... pure square wave
http://1.aseo.hostei.com/images/f3aebc649f9e.jpg
[11:46:36] <digi001> Kevin': yes thats what i was thinking to send it as ascii
[11:47:15] <digi001> how do i convert a floating point to a ascii?
[11:47:29] <digi001> with like 2 or 3 decimal places
[11:47:40] <ben1066> sprintf
[11:48:15] <Kevin`> no need to use sprintf most likely, use normal printf with the serial as stdout
[11:48:39] <Kevin`> that IS the way to do it thoughi
[11:49:29] <digi001> thanks
[11:51:12] <Kevin`> gnuplot and whatever expensive stuff you said before will also directly graph the ascii output, which is nice sometimes
[11:53:28] <digi001> yes TI Code Composer, the API, has a graph feature. but it sucks
[11:54:12] <ben1066> carp3: thats wrong isnt it :p
[11:54:25] <ben1066> even for a loop of U?
[11:55:00] <Casper> there is another way for those big number
[11:55:06] <Casper> use 9 bits transfert
[11:55:17] <Casper> use the 9th bit as a "packet start"
[11:57:18] <digi001> Ah yes. the last bit?
[11:57:43] <digi001> parity bit?
[11:57:44] <ben1066> first...
[11:57:52] <ben1066> there is a 9 bit mode
[11:58:19] <ben1066> carp3: shouldnt a looped U look like that?
[11:58:24] <ben1066> or will anything just ignore that
[11:59:33] <carp3> donno
[12:02:33] <ben1066> helpful :p
[12:04:51] <carp3> ben1066: yes it should be square wave
[12:05:13] <carp3> but your clock or baudrate calculations is wrong
[12:07:39] <carp3> ben1066:
http://1.aseo.hostei.com/images/aa98789cfb69.jpg
[12:08:01] <carp3> ben1066: first row : your hex file /NOT working
[12:08:20] <carp3> ben1066: second row: mine, working
[12:08:34] <carp3> both @20mhz/56700 8n1.
[12:11:51] <ben1066> how are you initing the usart
[12:13:06] <carp3> UCSR0B=0x18;
[12:13:07] <carp3> UCSR0C=0x06;
[12:13:09] <carp3> UBRR0L=0x15;
[12:13:44] <ben1066> ew magic numbers :p
[12:23:25] <ben1066> CARP
[12:23:32] <ben1066> YOU ARE BRILLIANT
[12:23:44] <carp3> ;)
[12:23:48] <ben1066> It works :D
[12:23:57] <ben1066> Now to calculate mine and why they were wrong
[12:23:58] <ben1066> >)>
[12:24:22] <ben1066> also, how come the 168 was released as a 24mhz but now the datasheet says 20mhz?
[12:25:07] <ben1066> Hmm
[12:25:16] <ben1066> Aparently this would fit on an atmega48 :D
[12:57:54] <grummund> Interesting. i didn't know atmel had previously released atmega at >20MHz, but looking at the datasheet revision history it clearly shows they reverted at around 04/04.
[13:00:06] <grummund> Not sure when mega168 actually hit commercial production though.
[13:02:27] <ben1066> also, whats the difference between 168 and 168P? is P the picopower version?
[13:03:10] <grummund> yes. but there maybe other subtle differences
[13:03:33] <grummund> also not there is the later A and PA
[13:04:02] <ben1066> woops I have the A
[13:06:51] <ben1066> Ooh, thats an interesting feature
[13:06:52] <ben1066> a RTC
[13:14:03] <ben1066> it suprises me how the avrs are limited by the flash, youd think atmel could use better flash for higher rated ics :S
[13:15:02] <Casper> ben1066: I find the main restriction is to not be able to execute code in ram
[13:15:18] <ben1066> Nono I meant in speed
[13:15:37] <ben1066> In functionality Id want a more common xmem and be able to run code in ram
[13:15:42] <Casper> flash speed is plentluy fast
[13:15:43] <RikusW> Casper: agreed
[13:15:46] <ben1066> although the latter is an architectual decision
[13:16:05] <ben1066> Casper: Its what stops avrs from being 40mhz instead of 16 aparently
[13:16:14] <RikusW> ben1066: xmega can run at 33MHz and AVR32 at 66MHz
[13:16:22] <RikusW> from flash afaik
[13:16:39] <Casper> if we were able to run code in ram, it would be really usefull, stick ab sd card and bang! "infinite" flash!
[13:16:59] <ben1066> RikusW: But who uses xmegas :p AVR32 is interesting
[13:17:15] <RikusW> AVR32 is more complicated too
[13:17:19] <ben1066> although the smd only parts are hardly hobbyist friendly, i hardly want to have to buy a PCB for everything
[13:17:28] <RikusW> it makes AVR8 look really easy
[13:17:58] <RikusW> AVR32 is tqfp 0.5mm or even worse bga
[13:18:23] <RikusW> ben1066: who said soldering smd is that hard ?
[13:18:39] <RikusW> once you have the pcb its not too difficult to solder tqfp
[13:19:14] <amee2k> pin pitch less than 1mm is doable, less than 0.5mm is a bitch
[13:19:18] <RikusW> Casper: there is an AVR + FPGA chip that executes from RAM
[13:19:29] <ben1066> I didnt, the thing is the fact you have to have a pcb, when prototyping its much easier to be able to just put an IC in a breadboard, that said, any tips for getting them out without bending the pins :p
[13:19:34] <RikusW> most avr's is 0.8
[13:19:58] <RikusW> ben1066: use a flat screwdriver
[13:20:11] <amee2k> soldering SMT works fine as long as the packages have their legs exposed. shit like BGA is awkward because its hard to align and impossible to inspect at home
[13:20:40] <amee2k> making SMT boards at home isn't too easy too because of the fine structures
[13:20:55] <ben1066> I dont really plan on using BGA, if I ever do decide to do SMD ill probly use a hotplate for reflow
[13:21:23] <RikusW> I soldered a lot of tqfp's using a Iron
[13:21:33] <RikusW> with Plato tip nr EW-103
[13:21:43] <RikusW> works very well
[13:21:44] <amee2k> reworked SMT boards with hot air. the problem is getting the dead part off, soldering the replacement on is pretty easy
[13:22:20] <RikusW> amee2k: yeah removing is more difficult
[13:22:39] <amee2k> especially if the parts were glued on during manufacturing
[13:22:55] <ben1066> Desoldering SMD is near impossible without good equipment, i tried to remove a 7805 SMD the other day, it didnt work :p
[13:23:00] <RikusW> maybe bending a piece thickish of copper wire to fit around the tqfp will help ?
[13:23:20] <RikusW> how about using a heatgun ? ;)
[13:23:23] * RikusW have one
[13:23:25] <ben1066> dont have one :p
[13:23:30] <ben1066> also, its has!
[13:23:37] <amee2k> before i had hot air i used to thread a piece of magnet wire under the legs. then heat the pins in sequence with the iron and pull the wire out to the side. that would break the solder connection and bend the pin upwards
[13:23:43] <RikusW> I stripped a few old motherboards entirely :-P
[13:23:51] <amee2k> only works for SOIC and *QFP though
[13:24:10] <ben1066> what are these AVR+FPGAs you speak of?
[13:24:32] <RikusW> its called avr slic or something
[13:24:40] <RikusW> but its dev tools are expensive
[13:24:46] <amee2k> i've got a hot air gun and it works pretty good. problem is lack of temperature control... stay anywhere for a few seconds too long and the board is toast
[13:24:51] <RikusW> not free like AS4 / AS5
[13:24:57] <amee2k> so its only really useful for a salvage run
[13:25:08] <RikusW> amee2k: mine does have power control
[13:25:49] <RikusW> i did manage to remove a dead m128 and replace it, with no flux.... didn't have any at the time
[13:25:49] <amee2k> mine only has two settings, 350°C at so-much airflow and 550°C with more airflow (i don't recall the airflow figures)
[13:26:00] <ben1066> Lol
[13:26:09] <ben1066> Rip it apart and add airflow control?
[13:26:10] <amee2k> 10EUR at the hardware store
[13:26:13] <RikusW> mine have that and another dial for heat
[13:26:23] <RikusW> 35E for mine
[13:26:45] <RikusW> RYOBI CPS-2000K
[13:26:53] <amee2k> its okay for mass salvage operations, and if i need fine control i can use the rework station now
[13:27:05] <ben1066> I wish there Von Nueman architecture AVR series
[13:27:12] <amee2k> i also figured out that the rework station is awesome for applying hot shrink the other day
[13:27:30] <RikusW> of course ;)
[13:27:49] <amee2k> no shit. made me facepalm why i didn't realize it sooner :P
[13:29:30] <ben1066> Would it be worthwhile getting a JTAGICE mkIII (they are cheaper than mkII)
[13:30:25] <RikusW> only works with AS5
[13:30:33] <ben1066> I use AS5, thats no big deal
[13:30:33] <RikusW> ben1066: why not get a dragon ?
[13:30:36] <RikusW> $50
[13:30:44] <ben1066> Dont dragons like to break?
[13:30:55] <RikusW> I'm carefull with mine
[13:30:58] * amee2k <3 avrisp mk2
[13:31:00] <RikusW> still working
[13:31:10] <ben1066> I have an avrisp mk2 amee2k :p
[13:31:17] <amee2k> hehe
[13:32:13] <ben1066> I dont know how I lived without it, I had a hacky arduino programmer, only worked half the time :p
[13:33:16] * amee2k idly wanders back to white LED shopping
[13:33:23] <ben1066> I think ill be getting a saleae logic analyser, since I really could use one :p
[13:33:45] <amee2k> i'm using my dso as ghetto 2ch la
[13:33:50] <ben1066> its that or a scope, and I dont do analogue stuff much, so a logic analyser seems more usefull :P
[13:34:00] <RikusW> since building my own isp programmer I don't use the dragon for isp anymore ->
http://sites.google.com/site/megau2s/home
[13:34:06] <ben1066> amee2k: where are you? US?
[13:34:27] <amee2k> my scope with the 16ch LA add-on now costs as much as i paid for without two years ago
[13:34:36] <amee2k> europe/germany
[13:35:09] <ben1066> ah right, for leds, if you were US there are loads of surplus places :p
[13:35:41] <amee2k> i don't want a whole load so i'll probably get some overpriced crees off ebay
[13:36:28] <ben1066> if you want stuff quick I use farnell, next day delivery :P
[13:36:33] <amee2k> 6.50 for a pack of 10 is cheaper than 5EUR... because shipping is 3EUR cheaper >_>
[13:37:00] <ben1066> lol, I got 25 transistors cheaper than 20 cause of the price boundries :p
[13:37:27] <amee2k> i get lots of stuff from farnell, they're awesome. only 6EUR shipping for everything is bit of a bummer sometimes
[13:37:32] * RikusW got 200 usb connectors for the same price as 120 ;)
[13:37:53] <amee2k> i tend to wait until at least 50EUR or so worth of stuff come together before i order from them
[13:38:09] <ben1066> I can spend £20 and get free shipping in the UK
[13:38:32] <amee2k> for fast small time shit ebay really is an option because shipping is either free or really cheap
[13:38:39] <ben1066> http://uk.farnell.com/tecstar/fge1202pc/software-function-generator/dp/5019564 HELL YEA
[13:38:50] <amee2k> nope, no free shipping limit i've seen from farnell over here
[13:39:15] <amee2k> wtf is that shit??
[13:39:25] <ben1066> No idea
[13:39:28] <ben1066> But its a floppy!
[13:39:40] <amee2k> pic says "function generator software or windows version 1.0"
[13:40:02] <ben1066> for*
[13:40:06] <ben1066> its for windows 1.0
[13:40:07] <amee2k> yeah, for
[13:40:20] <ben1066> they have 3 left
[13:40:22] <ben1066> :p
[13:40:43] <amee2k> if they make it two bucks instead of 20, i'll buy one :P
[13:41:00] <amee2k> i bet it sold for like 200 when it came out two decades ago
[13:41:47] <ElMonkey> hi, anyone have a butterfly the program via ISP?
[13:42:09] <ElMonkey> I'm not succeeding in attaching either a dragon or avrispmkii
[13:42:39] <ElMonkey> when the plug is supposedly the right way around, Vtarget is reported as 0.1V by avrdude
[13:43:09] <ben1066> There is 1 thing I envy pics for, the fact they have DIL versions of newer ICs
[13:43:19] <ElMonkey> and my voltmeter confirms that the VCC pin is indeed pulled low by the programmer
[13:43:24] <ElMonkey> but why?
[13:44:04] <RikusW> ElMonkey: did you apply power to the butterfly ?
[13:44:37] <amee2k> i seriously need to grow a pair of balls and try phototransfer for PCBs... SMT footprints with a sharpie suck. and my printer's toner stubbornly refuses to transfer
[13:44:39] <ben1066> oO freescale do micros?
[13:44:47] <ElMonkey> RikusW, the battery is in, and the lcd display what it should
[13:45:36] <ElMonkey> and when the ISP isnt plugged in, i get the expected just short of 3V on VCC/GND of the ISP pins on the board
[13:45:46] <RikusW> without the programmer connected what does vcc measure ?
[13:45:52] <amee2k> nice... 500mA warm white cree XRCWHT-L1, 10pcs for 6.56EUR
[13:45:53] <RikusW> ah
[13:46:43] <RikusW> ElMonkey: and the dragon is plugged into usb when you connected the butterfly ?
[13:46:49] <ElMonkey> RikusW, yes
[13:48:07] <RikusW> when the dragons usb connected is facing away from you pin 1 is on the lower left, ISP use the 2x3 header
[13:48:39] <ElMonkey> rikkus, yea
[13:48:47] <ElMonkey> same results with the avrisp as the dragon
[13:49:05] <ElMonkey> though when i turn around the plug on the avrisp, at least it reports the proper voltage
[13:49:13] <ElMonkey> though it also complains that the connection is reversed
[13:49:53] <RikusW> pin 2 == vcc and 6 == gnd
[13:51:04] <RikusW> the butterfly also have a jtag port.
[13:51:27] <ElMonkey> yea, jtag is harder to wire up though, i dont have a proper cable
[13:51:50] <RikusW> you don't have any ribbon cables ?
[13:52:06] <RikusW> I use that with a connector at both ends
[13:52:25] <RikusW> easy to press the connector onto the ribbon cable
[13:52:32] <ElMonkey> not here, unfortunately
[13:53:23] <ElMonkey> maybe this bloody cable is broken
[13:54:01] <ElMonkey> though i also tried plugging the butterfly onto the dragon directly
[13:54:04] <ElMonkey> without any success
[13:54:27] <RikusW> you're using the 2x3 connector ?
[13:54:57] <ElMonkey> yea
[13:55:25] <RikusW> and pin1 on the lcd side ?
[13:56:03] <RikusW> you really should get some ribbon cables...
[13:56:06] <RikusW> so much easier
[13:56:26] <ElMonkey> i was doing fine with stuff on the breadboard
[13:56:34] <ElMonkey> never needed a damn ribbon cable ;)
[13:56:45] <ElMonkey> but the butterfly isnt cooperating
[13:57:22] <ElMonkey> hmm, cable seems fine
[13:57:37] <ElMonkey> maybe you're not supposed to put the header on the top of th ebutterfly after all?
[13:57:42] <ElMonkey> doesnt make any sense
[13:57:49] <digi001> what is the order in a protacol for a external parallel bus?....1. Chip Select 2. address 3.read or write?
[13:58:35] <RikusW> CS first
[13:58:45] <ElMonkey> RikusW, anyway, thanks for your time, i unfortunately got to go
[13:58:49] <RikusW> apply address, then read
[13:58:57] <digi001> ok thanks
[13:59:01] <ElMonkey> i'll procure some further supplies in the morning, maybe something will help
[13:59:18] <RikusW> address+data then write
[14:00:09] <ben1066> I need to get more jumper cables
[14:00:19] <ben1066> Anyone got any suggestions for cables that dont fall apart :p
[14:00:57] <RikusW> jumper cables for connecting pinheaders ?
[14:01:10] <ben1066> male to male
[14:01:11] <RikusW> like YY09 ?
[14:01:15] <ben1066> for breadboarding
[14:01:28] <ben1066> The ones I have atm are gradually falling apart
[14:01:39] <ben1066> http://www.oomlout.co.uk/breadboard-jumper-wires-70-pieces-p-181.html?zenid=b6f52f61740922ddf3ef8e4c6ea35529
[14:01:42] <ben1066> These
[14:01:45] <RikusW> use single core 0.5mm wire
[14:01:56] <ben1066> I used to, but it kept breaking in the holes :S
[14:02:53] <RikusW> I've gotten some thick comms cable offcut containing a lot of colored wires
[14:02:59] <RikusW> works quite well
[14:03:29] <RikusW> I just strip of 5mm insulation on each ent
[14:03:31] <RikusW> end
[14:03:37] <RikusW> and use it like that
[14:03:57] <RikusW> but it must be singe cored, like say ethernet cable
[14:04:03] <ben1066> Do you know the gauge?
[14:04:06] <ben1066> is it single core?
[14:04:10] <RikusW> yes
[14:04:13] <RikusW> 0.5mm
[14:04:48] <ben1066> I meant AWG :p
[14:05:00] <soul-d> how long did you work with above cables before they fell apart ?
[14:05:15] <soul-d> i orderd thise a few cause im to lazy to cut and strip :P
[14:05:42] <RikusW> ben1066: #24
[14:05:55] <RikusW> or AWG # 25
[14:06:09] <RikusW> or maybe #23 too
[14:08:16] <ben1066> Yea, thats kinda what I used to use
[14:08:19] <ben1066> and kept breaking :p
[14:34:52] <ben1066> is a jtagice worthwhile?
[14:36:41] <Kevin`> ben1066: if you have to ask (if you are paying for it yourself, that is), probably not. you might want a dragon
[14:37:36] <ben1066> yea I would be
[14:37:56] <ben1066> I use the simulator when I need to check registers normally, I just wondered if Id ever find a time when I REALLY need one
[14:40:35] <Mithrandir> hiya, I'm having a bit of trouble getting avrdude to talk to my stk500 with a attiny85 on it. It complains either about a device signature mismatch or it fails to read the same values for lfuse/hfuse. I've tried with various -B values, but it doesn't seem to help at all.
[14:41:14] <Mithrandir> also, I've got an mega32l which I can program fine using uisp, but where avrdude complains when trying to verify, so I wonder if avrdude is doing something odd.
[14:41:26] <Mithrandir> (it seems uisp doesn't support the attiny85)
[14:42:07] <Kevin`> does uisp do a verify step?
[14:42:30] <ben1066> are you sure yoru voltage is rihgt? I got a wrong sig when I accidently connected the unregulated power (7v)
[14:42:38] <Kevin`> ooh nasty
[14:43:36] <Mithrandir> I'm using an stk500, so not sure how I'd manage to connect the unregulated power?
[14:43:51] <Mithrandir> and yes, uisp verifies when I pass --verify
[14:44:07] <ben1066> No idea, just checking :p also Kevin` my avr was fine, nothing burnt, no smoke released, nothing
[14:44:32] <Mithrandir> I've tried with another attiny85 with the same problems, so I don't think the chip is broken.
[14:45:07] <Mithrandir> (both are fresh out of the tube, so I _assume_ they should be ok)
[14:46:09] <Kevin`> it's strange, I would expect an stk500 to have pretty reliable wiring, and avrdude rarely has problems
[14:47:20] <Mithrandir> yes, that's basically my take as well. It's through a USB-to-serial adapter, but I doubt that should cause problems.
[14:47:23] <Kevin`> you didn't do anything different between the runs when it read the signature correctly and when it didn't, did you?
[14:47:29] <Kevin`> like say, ignore the signature
[14:47:51] <Mithrandir> I've tried ignoring the signature too, but it then fails reading the fuses.
[14:48:06] <Mithrandir> from running with -vvv it seems like it reads them three times and compares.
[14:48:24] <Kevin`> ah
[14:48:34] <Kevin`> don't ignore the signature
[14:48:58] <Kevin`> how do you have it connected to the board? are you sure the isp pins are connected to the right pins on the stk500?
[14:50:04] <Mithrandir> same as
http://blog.flipwork.nl/fp-content/images/.thumbs/isp_stk500.jpg
[14:50:42] <Mithrandir> hmm, I found a tip somewhere to run avrdude with -t and then sck 3 to adjust the timing delay.
[14:50:45] <Mithrandir> and now it works. :-)
[14:50:54] * Mithrandir has.. a blinking led!
[14:51:06] <amee2k> me too \o/
[14:51:19] <amee2k> i got it from conrad a loooong time ago
[14:51:41] <Mithrandir> and it seems to have verified fine and all this time around, so I suspect it just needed a bit of tickling (and/or kicking)
[15:09:03] <ben1066> I need a PS/2 keyboard
[15:09:07] <ben1066> or a serial keyboard
[15:45:36] <ben1066> I need to buy some ram Hmm
[15:54:58] <ben1066> are small ram chiops even made these days!
[15:56:01] <karlp> get a chip with mor eonboard...
[16:00:01] <amee2k> mmh, maybe this one is somewhat trivial, but are LEDs sensitive to average or RMS forward current?
[16:05:21] <karlp> sensitive in what way?
[16:05:22] <karlp> breaking?
[16:05:42] <amee2k> i.e. if it says nominal forward current is 400mA, would i need to size my driver for 400mArms? or 400mA average?
[16:06:34] <amee2k> karlp: well, specification wise i suppose. both figures are amperages, but how do i have to interpret the specs to use them correctly
[16:07:23] <pupy> hi!
[16:08:13] <amee2k> because the datasheets i've seen only give maximum pulsed current, and typical and maximum DC currents
[16:08:46] <amee2k> where pulsed is usually specified as square pulses only
[16:09:04] <amee2k> but what do i use if RMS and average aren't equivalent for my drive waveform
[16:11:34] <pupy> peaks?
[16:12:04] <pupy> :D don't even know what you're talking about, but sounds to me about some power metering or the sort
[16:12:22] <amee2k> i'm working on a power LED driver
[16:12:35] <amee2k> well, sort of. still getting into some of the theory details
[16:13:21] <pupy> ahm, I've been with something related lately, though it's mostly dealing with AC
[16:14:37] <amee2k> my first ones will run straight off mains. i've got the cases of some CFL bulbs to repurpose
[16:15:04] <amee2k> in the end i'm wondering about some self-oscillating SMP driver though
[16:15:47] <amee2k> either way would drive the LEDs with something that looks more like a sine than a square, so confusing RMS and average will result in a significant error
[16:16:18] <amee2k> for a clean sine its ~11% off
[16:16:52] <pupy> smp as in switched mode power supply? that's a bit too much for me
[16:17:47] <amee2k> i'm not too hot about RF and switched stuff, thats why i'm using mains and DC for practice
[16:18:27] <pupy> ah ok, with a voltage divider and an opamp you can easily grab wave's peaks if you feed ouput to the adc, or even
[16:18:55] <amee2k> mmh adc lol?
[16:19:17] <amee2k> my ac driver is just a rectifier and capacitive ballast
[16:20:33] <pupy> bias it to, say 0.5 - 4.5V, such that you can also grab zero-cross
[16:20:53] <amee2k> http://ompldr.org/vYmdkeg/led-ballast.png
[16:21:27] <karlp> if the leds say 400ma,
[16:21:37] <karlp> you want to design the power supply for substantially more anyway
[16:21:43] <karlp> regardless of rms or average
[16:22:52] <amee2k> but i need to set the driver itself for a specific one. the 11% error is quite significant compared to my driver tolerances
[16:44:35] <grummund> amee2k: you can do away with the bridge rectifier there if 50% drive onto the LEDs is acceptable... just connect the LEDs in reverse parallel.
[16:45:57] <amee2k> well, technically yes but then they're at the verge of flickering
[16:46:32] <amee2k> i did that on a meter for scale illumination and i could see it flicker out of the corner of my eyes
[16:46:37] <amee2k> very annoying
[16:47:52] <grummund> fair enough
[16:47:58] <amee2k> this way they're flickering at 100Hz same as a fluo tube
[16:48:40] <geckosenator> can someone tell me what --relax passed to the linker does?
[16:54:17] <grummund> it results in a smaller code image
[16:55:23] <geckosenator> apparently
[16:55:24] <geckosenator> but, what does it do?
[16:55:24] <geckosenator> because it results in a size of 0 since ld has a segmentation fault heh
[16:55:32] <grummund> http://sourceware.org/binutils/docs/ld/Options.html#Options
[16:55:34] <grummund> http://www.tty1.net/blog/2008-04-29-avr-gcc-optimisations_en.html
[16:56:18] <geckosenator> when I remove -Wl,--relax it links without crashing
[16:57:00] <Steffann> He, geckosenator is back alive?!
[16:57:09] <grummund> hmm, well not that it helps but... "it works for me".
[16:57:42] <geckosenator> avr-ld -v?
[16:57:54] <geckosenator> I am using 2.22
[16:58:08] <grummund> GNU ld (WinAVR 20100110) 2.19
[16:58:35] <geckosenator> oh thats over a year old mine is 6 days
[16:58:49] <geckosenator> maybe I'll try with 2.19 and see what happens then
[16:59:28] <grummund> that's what you get for being latest and greatest
[16:59:43] <geckosenator> haha
[16:59:50] <geckosenator> how much difference in code size do you get?
[17:00:03] <grummund> it made a big difference as i recall
[17:00:47] <geckosenator> like 5%?
[17:01:10] <grummund> i can't remember tbh
[17:01:41] <grummund> but it was one of (if not the) biggest contributor to code size reduction
[17:03:15] <geckosenator> ok, well its currently the biggest contributor to segmentation faults
[17:03:19] <geckosenator> and I have an empty main
[17:03:35] <grummund> hehe
[17:03:58] <geckosenator> with no libraries or includes or anything
[17:04:03] <geckosenator> just an empty program
[17:05:38] <grummund> are you also using --gc-sections ?
[17:05:42] <geckosenator> yes
[17:05:55] <geckosenator> but evene without it doesn't matter
[17:06:27] <geckosenator> compiler flags do though..
[17:06:57] <geckosenator> I always use function-sections and data-sections too
[17:09:12] <geckosenator> actually no scratch that.. "avr-gcc -mmcu=at90usb1287 test.c -o test -Wl,--relax"
[17:09:20] <geckosenator> with an empty main in test.c causes crash
[17:09:45] * geckosenator starts single stepping ld
[17:11:21] <grummund> what about same with atmega32 ?
[17:12:08] <karlp> I don't knwo what else it does, but relax is essential for devices with more than 128k flash
[17:12:17] <karlp> if you want to jump over the boundary at least
[17:14:00] <geckosenator> why?
[17:14:32] <geckosenator> I thought they use trampolines
[17:15:02] <geckosenator> I have used atmega 2560 because I needed about 180k of flash
[17:15:11] <geckosenator> but back then --relax worked heh
[17:20:20] <grummund> geckosenator:
http://pastebin.com/9U0szKSu
[17:21:08] <geckosenator> 4.3.3? I am using 4.7.0
[17:21:27] <grummund> well... that's what you get for being latest and greatest
[17:21:38] <geckosenator> anyway, try it with the latest binutils
[17:21:42] <geckosenator> I don't think this is a bug in gcc
[17:21:57] * Mithrandir sighs at his attiny suddenly just wanting to sulk in a corner.
[17:22:19] <geckosenator> well I _need_ the latest version of everything since I am working on the compiler
[17:22:38] <grummund> ah, then it is to be expected :P
[17:23:14] <geckosenator> I found a bug in avr-gcc yesterday.. posted to mailing list, it's already fixed :-P
[17:24:50] <grummund> http://sourceware.org/bugzilla/show_bug.cgi?id=13402
[17:27:22] <geckosenator> wow, it just crashes for me
[17:28:28] <geckosenator> of course we have no idea what version of binutils that guy used
[17:30:11] <geckosenator> I'm recompiliing binutils with optimization off so I can debug it heh
[17:30:17] <geckosenator> my computer is slow
[18:48:51] <geckosenator> I fixed the relax thing
[19:18:42] <inflex> lo folks
[19:18:58] <Tom_itx> hi inflex
[19:19:24] <inflex> do anything for Turkey day?
[19:19:38] <Tom_itx> ate turkey
[19:27:49] <inflex> good man :D
[19:28:21] <Tom_itx> just fallin in line like the rest
[19:28:40] <geckosenator> I dont live in america anymore
[19:28:52] <Tom_itx> some ways i envy you
[19:29:04] <geckosenator> no turkey for me haha
[20:22:13] <Valen> we have a frozen turkey in the freezer
[20:22:19] <Valen> its been there for like 2 years now
[20:22:28] <Valen> going to try to defrost it for this xmas
[20:32:41] <geckosenator> 2 years?
[20:33:42] <Tom_itx> time makes wine better, not sure about turkey
[20:34:29] <Landon> Valen: well, now is about the time to start defrosting it if it's been in there for that long :P
[20:42:06] <inflex> yeah, that'd be a bit dubious Valen
[20:45:43] <Valen> eh its been frozen
[20:45:47] <Valen> i mean in stasis
[20:46:09] <Valen> heat shrunk sealed package as well so there shouldn't be any biological activity
[20:46:19] <Valen> if i smoke it it should be good ;->
