#avr | Logs for 2015-05-04

Back
[06:08:35] <Lambda_Aurigae> kids with internet in schools...sheesh.
[06:08:45] <Lambda_Aurigae> we got our first Apple-II machines when I was in 9th grade.
[06:09:01] <Lambda_Aurigae> and had our first actual computer class when I was in 11th grade.
[06:09:14] <Lambda_Aurigae> didn't even have modems on them, much less interwebs.
[06:10:00] <twnqx> heh
[06:46:43] <Lambda_Aurigae> May the 4th be with you.
[06:47:29] <oal> Anyone here using platformio?
[06:48:15] <Lambda_Aurigae> never heard of it.
[06:49:06] <oal> It's great. Makes flashing lots of microcontrollers very easy. "platformio run" and it's on my board :)
[06:49:28] <Lambda_Aurigae> all the microcontrollers I need I can do.
[06:49:31] <oal> But... I'd like to know if I can flash assembly programs instead of c
[06:49:41] <Lambda_Aurigae> I don't see 8052 or pic32 listed there.
[06:52:02] <Lambda_Aurigae> looks like another do-it-all tool that's missing a lot of things that it needs to do it all.
[06:53:26] <oal> Good for newbies like me at least
[06:53:49] <Lambda_Aurigae> yes, like arduino...teaches bad habits and doesn't teach you how the core of things work.
[06:55:05] <Lambda_Aurigae> but, if it works for you, great.
[06:55:08] <oal> I started with arduino. Now I program the arduino pro mini without the arduino framework with platform.io (avr c, setting up ports, interrupts etc)
[06:55:39] <oal> So when I outgrow that, I'll probably look into using avrdude etc without something like platform.io
[06:55:58] <Lambda_Aurigae> no clue what platform.io is either.
[06:56:26] <Lambda_Aurigae> hmm..nevermind.
[06:56:29] <Lambda_Aurigae> too early this morning.
[06:56:31] <Lambda_Aurigae> same thing.
[06:56:57] <oal> A project structure and command line tool that contains configurations for many microcontrollers to ease development and flashing. Also it has a package manager I think
[07:02:33] <Lambda_Aurigae> heck, it doesn't support any pic chips, not just pic32...and still no 8051/8052 series.
[07:05:42] <Lambda_Aurigae> oh, I love the claim...PlatformIO supports over 100 Embedded Boards and all major Development Platforms.
[07:06:11] <Lambda_Aurigae> hell, pic and 8052 are still two of the most used microcontrollers on the planet...AVR might be better than either in many respects but they are still major.
[07:09:16] <Lambda_Aurigae> heck, I bet there is more development on Cypress stuff than STM32 even and Cypress is not listed either.
[07:10:38] <Lambda_Aurigae> so, to answer your first question, oal, it seems nobody here,,who is awake at the moment anyhow,,uses it.
[09:29:50] <FatalNIX> I actually have a super fondness for 65xx
[09:30:58] <FatalNIX> However, AVR is nice because of a built in clock, and theres IO ports, which I rather would have a full 65xx like memory mapped IO setup if it were a large project, but it's really nice for quickly doing some microcontroller stuff, while still capablre of doing the complicated projects if you really wanted.
[09:34:57] <specing> built in clock? stm32 cortex-mX chips usually have 3+ clock sources built-in
[09:35:14] <FatalNIX> I have never wired up an ARM
[09:35:35] <FatalNIX> mostly because I would rather build the PCB myself, and high frequency PCB design is not my thing.. yet.
[09:35:47] <malinus> Not much to wire up, as little as a avr8
[09:36:14] <malinus> high frequency PCB design? They go up to like 150MHz max.
[09:36:18] <specing> FatalNIX: everything up to and including a cortex-m4f is "connect Vcc, connect GND"
[09:36:28] <malinus> the cheaper ones are like in the 40-70MHz range
[09:36:29] <FatalNIX> oh, I was thinking of like Cortex A8's and stuff.
[09:36:35] <specing> malinus: Cortex-m4 goes to over 200 MHz
[09:36:37] <FatalNIX> or A15s
[09:36:43] <specing> cortex-m7 goes over 300 MHz
[09:37:09] <malinus> I thought we were talking about specific stm32. And sure they have the f7 line, but who the fuck uses that :P
[09:37:30] <FatalNIX> lol
[09:37:31] <specing> malinus: No idea because it is unobtainium
[09:37:32] <malinus> they are like $30/piece or so.
[09:37:44] <malinus> oh, or maybe even that
[09:37:55] <FatalNIX> apparently the cmos WDC 6502s work up to 200 Mhz.
[09:37:59] <FatalNIX> which is incredible
[09:38:02] <FatalNIX> (for a 6502)
[09:38:14] <FatalNIX> they're a modern redesign
[09:38:27] <malinus> But yeah, f4, which is the fastest normal stm32, goes up to like ~200MHz. That's what I was refering to.
[09:38:28] <FatalNIX> mine runs at 2 :D
[09:38:58] <specing> AT91SAM9260 is $10
[09:39:08] <FatalNIX> problem is then you gotta buy static ram that's runs there, and if you need a VIA, a VIA that runs there, MMU that runs that speed if you go that route..
[09:39:09] <specing> it can run Linux (mine runs Gentoo)
[09:39:15] <specing> It is in LQFP-208
[09:39:15] <malinus> I don't think you have to do much, pcb-design-wise, for that to work (200MHz)
[09:39:21] <FatalNIX> not hard but most of my parts are nmos @ 1 - 2 Mhz
[09:39:21] <specing> you can basically solder it at home
[09:39:35] <malinus> specing: LQFP-208? What's the pitch?
[09:39:39] <specing> malinus: 0.5
[09:39:46] <specing> malinus: ITS HUUUUGEEEEE (like 4x4 cm)
[09:39:54] <malinus> haha wow wtf
[09:40:01] <malinus> it's like those FPGA's
[09:40:09] <malinus> but they have 99999 pinouts
[09:40:37] <malinus> becuase BGA
[09:40:58] <malinus> specing: did you do it by hand?
[09:42:24] <specing> no I got it on a board
[09:42:32] <specing> Im just saying it is possible
[09:42:48] <specing> DRAM has 0.8 - 0.85 pitch in TSSOP
[09:43:10] <specing> some flash chip is 8-pin pdip
[09:43:16] <specing> and a 3-pin ldo
[09:43:20] <specing> basically all you need
[09:43:40] <specing> actually you could leave the RAM unconnected and use it like a microcontroller (cache as ram)
[11:04:01] <Lambda_Aurigae> FatalNIX, this is why I like 8052. Not quite 6502 that I grew up with but I can do memory mapped i/o,,,along with doing external parallel sram or flash or eeprom for program space and external sram for data storage.
[11:17:25] <twnqx> hm, seeed's pcb service is weird. same price for 5 to 300 pcbs?
[11:18:17] <twnqx> guess i broke something
[11:18:38] <twnqx> ah. website down, cloudflare page only left
[13:47:51] <hypermagic> hello my friends :)
[13:48:11] <hypermagic> what d you use as a random function in your mcu?
[13:49:00] <LeoNerd> Depends how good "random" needs to be
[13:49:21] <LeoNerd> E.g. to make pretty LEDs flash and blink in ways that aren't obviously neat patterns, I'd just use a linear-cong. generator
[13:49:39] <hypermagic> let's say it be unpredictable, or at least look good if doing something random?
[13:49:51] <LeoNerd> To make an IV to encrypt radio packets, I'd use something much more secure like measuring tiny timing drift between CPU and radio or CPU and watchdog
[13:49:55] <LeoNerd> ((yes I have done all of these things ;) ))
[13:50:22] <hypermagic> :)
[13:50:40] <hypermagic> i did something similar in concept on PC too
[13:50:58] <hypermagic> how did you encode that entropy into random data?
[13:51:25] <LeoNerd> Use the bottom two bits of the timer sampled on the WDT interrupt, to mix up the entropy pool some more
[13:51:31] <LeoNerd> And just generally keep a pool
[13:51:59] <LeoNerd> For me, it's rare I need an IV from the pool, so I can just keep stirring it constantly and just take a bit when I need it
[13:52:40] <antto> has anyone done UART over RF?
[13:52:46] <hypermagic> by mix you mean... ? :)
[13:53:22] <hypermagic> antto, im not sure i want uart :/ i rather use a clock and have any speed
[13:53:40] <antto> is it a bad idea?
[13:53:50] <hypermagic> it was probably done before...
[13:54:02] <hypermagic> i saw radio com with pc too on the net
[13:54:12] <hypermagic> they use soundcard
[13:54:28] <hypermagic> and they even do morse codes lol
[13:54:46] <hypermagic> that is a huge bandwidth...
[13:57:20] <hypermagic> LeoNerd, i find it more difficult to extract some entropy within the mcu than the pc, because it is much more deterministic, and has small, mostly fixed latencies :/
[13:58:15] <hypermagic> i think it can be possible on the simple avr to trigger an interrupt within +-1 clock ticks, what is your opinion?
[13:59:43] <tpw_rules> i've seen mcus which can use internal PLLs as truly random sources
[14:03:20] <hypermagic> and the internal rc oscillator is too stable i think :)
[14:04:42] <tpw_rules> i'm not sure an avr could do it
[14:04:42] <hypermagic> another entropy source could be the vcc voltage relative to the internal voltage reference, and the internal temperature sensor
[14:04:58] * tpw_rules was thinking propeller
[14:05:00] <hypermagic> the xmega has pll
[14:05:07] <tpw_rules> i've never used it
[14:05:17] <tpw_rules> it's not for clock. the propeller uses a pll for timers
[14:05:18] <hypermagic> me neither, :/
[14:05:58] <hypermagic> another source for some entropy might be a capacitive touchbutton, but that would require change too
[14:07:11] <tpw_rules> attach a geiger counter and radioactive source
[14:07:19] <hypermagic> hehehe
[14:07:27] <hypermagic> radioactive source is everywhere
[14:07:39] <tpw_rules> but it's slow
[14:07:43] <hypermagic> or pack your stuff with a fish from the ocean
[14:07:51] <tpw_rules> ambient radiation is low count
[14:08:11] <hypermagic> fish from japan ?
[14:08:18] <tpw_rules> fun fact: a becquerel (Bq) is one count per second
[14:08:34] <hypermagic> yes but you can measure time between 2 hits
[14:08:47] <hypermagic> it is analog
[14:09:02] <tpw_rules> you can but you'll get like 5 bits per minute maybe
[14:09:09] <hypermagic> also the tick may have an amplitude
[14:09:13] <tpw_rules> it doesn't
[14:09:17] <tpw_rules> that's not how decay works
[14:09:22] <hypermagic> it does, if you sense it analog
[14:09:40] <tpw_rules> but that doesn't give you randomness
[14:09:47] <hypermagic> the amplitude depends on angle, sensor hit position, and particly energy
[14:09:56] <tpw_rules> the time between decay events is truly random. that stuff isn't
[14:10:28] <hypermagic> well there are cosmic rays too