#avr Logs

Dec 11 2017

#avr Calendar

01:57 AM Casper: Cheaterman: if you screw up the clock source, it is not that bad
01:59 AM Casper: all you need is a way to send a clock signal, it can be another AVR that you program with an insane led flasher (aka while(1) PINA = 0xFF; or the like), some have the option to have a clock out
02:17 AM Cheaterman: Casper: yeah but this one "can't be attached to crystals", and also I wasn't too keen on soldering wires to my TQFP UC
02:17 AM Cheaterman: but I realize another AVR is a clock source, not a crystal
02:17 AM Cheaterman: But yeah, it's complicated :3
02:18 AM Casper: still need the crystal pin unfortunatelly to feed the signal in
02:19 AM Casper: if it is a board that you made, next time consider to add provision on the board for a crystal, or atleast a test point for the external clock in
02:20 AM Cheaterman: I currently use a 10MHz xtal as mentioned
02:20 AM Cheaterman: but as you said, a test point for the xtal1 pin could help
02:20 AM Cheaterman: Thanks for the hint :)
02:21 AM Cheaterman: Also I'm trying to setup a UART-programmable AVR with a FT232R, it has DTR wired to the ATmega168 RESET through a 100nF capacitor, and it doesn't work
02:21 AM Cheaterman: I have a 10k resistor pulling up the ATmega side
02:21 AM Cheaterman: if you have any hints on how to troubleshoot, I'd love that :3
02:33 AM Casper: with serial bootloader?
02:33 AM Casper: look up how arduino do it
02:40 AM raek: Cheaterman: have you verified that the bootloader works when resetting the atmega manually?
02:41 AM raek: 100n cap and 10k resistor sounds right (assuming you have the pull up on the atmega side of the cap)
02:42 AM raek: I have done that myself on one of my boards I use the most: https://pbs.twimg.com/media/DQI4z18W4AAZ1CC.jpg:large
02:43 AM raek: also, I think I connected the RTS pin to reset, not the DTR one
02:44 AM raek: but I vaguely recall that the Arduino IDE pulls both RTS and DTR low for compatibility or something
02:44 AM raek: I don't know if "avrdude -c arduino" also uses both or just RTS
02:46 AM raek: before I wired up the reset pin to the FTDI header, I had to push the reset button manually, but that way I could tell that the bootload (optiboot in my case) worked.
02:49 AM raek: *bootloader
03:05 AM Cheaterman: raek: I got it, it was the bootloader indeed
03:05 AM Cheaterman: raek: I forgot mentioning I was using a bootloader made for 16MHz
03:05 AM Cheaterman: and I'm using my MCU at 10MHz
03:05 AM Cheaterman: I compiled a new bootloader, and then I set the proper baudrate on avrdude, and everything uploaded nicely
03:05 AM Cheaterman: I'm happy now ^__^
03:11 AM raek: great!
04:52 AM polprog: morning
05:01 AM nohitzwork: morning
08:15 AM amigojapan_bnc is now known as amigojapan_festi
08:15 AM amigojapan_festi is now known as amigoj_festivus
08:31 AM Jartza: https://twitter.com/_youhadonejob1/status/939972059609927681
08:47 AM thardin: is there a way to get avr-gcc to not choke on arrays 32k and larger?
08:48 AM thardin: would be nice not to have to limit by dubble buffered sample collection to 2x16383
08:48 AM cehteh: which AVR chip has so much flash or ram?
08:49 AM thardin: ATmega128 with 64k XMEM
08:50 AM cehteh: never used that
08:51 AM LeoNerd: Yah I don't think any AVRs have that much internally, but some have external RAM buses which might support it
10:23 AM nohitzzzzzz is now known as nohitzzzz
10:36 AM nohitzzzz is now known as nohit_
11:38 AM Emil: Anyone know of a ram chip with two read/write busses?
11:42 AM Thrashbarg: look for dual port RAM
11:52 AM Emil: Ah that's the word
11:56 AM Emil: Thanks
11:56 AM Emil: Expensive as heck though :D
12:09 PM cehteh: possibly easier to workaround than use such special shit?
12:34 PM LeoNerd: Dualport with two *writable* buses sounds terrible
12:34 PM LeoNerd: I could imagine situations needing dualport where one only reads
12:40 PM cehteh: shared memory in hardware
01:10 PM Emil: LeoNerd: well just reading would be enough
01:30 PM cehteh: cant you just duplicate or snoop on the memory, send updates
01:33 PM Emil: cehteh: issue is that sending shit out actually takes time :D
01:33 PM Emil: I wish we had dmas
01:35 PM cehteh: yes, but maybe you just use the wrong hardware then, avr+expensive dual port ram is prolly more expensive and certainly much more complex to program than starting with a stm32 or?
01:36 PM cehteh: dunno what you plan to do, but somtimes, if you are not Jartza, then AVR's are just the wrong tool :D
01:39 PM Emil: cehteh: gib https://emil.fi/avr but for stm32s (or any arm pretty much) and I'll jump ship instantly
03:50 PM Cheaterm1n is now known as Cheaterman
03:56 PM Santh_ is now known as Santh
03:58 PM specing_ is now known as specing
08:46 PM _ami_: hey guys,
09:16 PM rue_mohr: why the F*** can I open that link with lynx, but not with iceweasel
10:36 PM _ami_: rue_mohr, using internal proxy?
11:13 PM day__ is now known as daey
11:44 PM rue_mohr: no
11:44 PM rue_mohr: it turns out its because of a spat between deian and mozilla
11:44 PM rue_mohr: firefox and debain policies didn't work togethor,
11:44 PM rue_mohr: so debain created icewesasel
11:45 PM rue_mohr: and like all things debain, is always 5 years out of date
11:45 PM rue_mohr: which mattered when someone cracked ssl3 and everyone started not using it
11:45 PM rue_mohr: which is the only encryption iceweasel supported
11:46 PM rue_mohr: the policies are essentially the same, but people dont care about the legal stuff as much now, leading to firefox-esr
11:46 PM rue_mohr: aka "the intentionally out of date version of firefox for debian users"
11:47 PM rue_mohr: by 2019 it should all settle out, that, or the incorperation of closed source binaries to firefox will have it outsted from debain again
11:49 PM rue_mohr: it seems that debian is not winning on the "operating system that caused an open-source landslide" front
11:51 PM rue_mohr: I wanted to add a command line calculator package of mine to the debain repo, when I was done reading up howto, it was obvious that it was impossable for a jeo-blow to add software to debian, and made me wonder how ANY package ends up in debian
11:51 PM rue_mohr: joe...
11:52 PM rue_mohr: (it would take me 4 years and hundreds of dollars in travel to become someone who could add a package, there is NOBODY available to help (sponsor) any new package)
11:53 PM rue_mohr: so, if you want to make something into a debian package, make the package and put it on your own repository server