#avr Logs

Oct 22 2018

12:16 AM hjf: hi all. i'm trying to figure out something with atmel (microchip?) stuff. i read about ATSAML10 and L11. i want to try them. are the compilers for those micros free? or free-but-limited? or pay only?
12:20 AM rue_mohr: ?
12:20 AM rue_mohr: whats saml10?
12:20 AM rue_mohr: if its a new microchip product, I'd not be surprised if its pay-to-develop-with
12:20 AM rue_mohr: thats why i'm (slowly) bailing over to stm32
12:20 AM hjf: no it's atmel pedigree
12:20 AM hjf: works with atmel studio
12:22 AM hjf: "for now" ? or is it pretty much too big for microchip to XC?
12:22 AM rue_mohr: atmel studio isn't by atmel
12:23 AM rue_mohr: well
12:23 AM rue_mohr: its core is gcc
12:23 AM rue_mohr: and otherwise its just a programming ide
12:23 AM rue_mohr: you dont need it
12:23 AM hjf: i'm familiar with PICs, but pretty much only the 8 bit line. i'm used to their licensing bullshit
12:23 AM rue_mohr: yea, thats why I abandoned pics
12:24 AM hjf: so far i haven't needed more than the free version though
12:24 AM rue_mohr: that and the whole crap they did with the 16F877 vs the 16F877A
12:24 AM hjf: i do small shit anyways
12:24 AM rue_mohr: avrs: full clock speed, no pages memory, free C compiler
12:24 AM rue_mohr: and public domain programmers
12:25 AM hjf: 877 eh
12:25 AM hjf: good old times
12:25 AM rue_mohr: vs the 877A
12:25 AM rue_mohr: NOT THE SAME CHIP
12:25 AM rue_mohr: }:(
12:25 AM hjf: i remember when i got my first 877 (must have been an 877A tho). i was like whoa you can DEBUG THIS LIVE??
12:26 AM hjf: and built a ICD2 clone hehe
12:26 AM hjf: speaking of tools, do i need any fancy tools to program AVRs or SAM?
12:27 AM hjf: i mean i use arduino a lot
12:27 AM hjf: but never really looked into what's going on. is it JTAG or a proprietary serial programming thing?
12:28 AM rue_mohr: fro avrs you need a $2 programmer from china and free software
12:28 AM rue_mohr: or if you want, you can build a parallel port one
12:28 AM rue_mohr: or make one from an arduino
12:29 AM hjf: can you debug with that
12:30 AM rue_mohr: I do it differently
12:30 AM hjf: i was pissed the other day when i tried to debug a ESP32. the easy option is PlatformIO but it's $12 a month to debug so screw it
12:30 AM rue_mohr: I dont need to single step code and watch variable values
12:31 AM rue_mohr: I wont get into how people shoudl not write a huge amount of code and then try to run it
12:31 AM rue_mohr: cause otherwise I'm gonna start a rant about tested modules
12:33 AM hjf: hah at work i debug step by step all the time
12:33 AM hjf: but that's only because i get a giant project developed by a bunch of juniors and then i'm asked to "improve it"
12:34 AM rue_mohr: want to play with avrs?
12:35 AM vmt: get a cortex-m.
12:35 AM rue_mohr: oops, I was supposed to start some 3d prints for ma
12:35 AM hjf: so you have to step debug to even find out where this dumbass decided to start the program (which is of course deep in a User Control buried 4 User Controls down a Form that's called programmatically) that's .NET Winforms 2.0 for ya
12:35 AM vmt: -m0/1/3 is reasonable, 4/7 for somewhat unreasonable applications
12:36 AM rue_mohr: hmm, is it acceptable to put led pinstripes on a battle mech thats supposed to be moderatly cammo?
12:36 AM hjf: vmt: how about that one i mentioned? L10/L11
12:36 AM hjf: cortex M23, but like 100nA sleep current
12:37 AM hjf: 500nA with register retention
12:37 AM hjf: i like the idea of a 32-bit processor running off a CR2032 for 10 years
12:37 AM vmt: haven't actually tried m23s. they're quite new
12:38 AM vmt: -m0+ is very low-power too
12:38 AM rue_mohr: sleep isn't running...
12:39 AM hjf: ah but they have SleepWalking(R) peripherals
12:39 AM hjf: LOL
12:39 AM hjf: whatever the hell that is
12:39 AM hjf: i guess peripherals running off a secondary clock while the core sleeps
12:41 AM hjf: do i need any special tools for cortex?
12:41 AM hjf: i have a stm32f4 discovery board which also acts as a "ST LINK" jtag thing
12:41 AM hjf: and a BusBlaster 2.5
08:11 AM rue_bed: --
10:45 AM twnqx: any suggestion for µCs (any arch) that has parameters of 8 pins with separate uart and i²c, and something like 16+kB ram and 16+kb flash?
10:46 AM twnqx: or at least 4kB ram...
10:49 AM Emil: Consider joining #avrs
11:05 AM thardin: consider considering
11:09 AM LeoNerd: I don't think any AVRs have as much as 16KiB of RAM. I think you'd be looking into something like an STM32 at that point
11:13 AM vmt: 1284?
11:13 AM vmt: how much does the 1284 have?
11:13 AM vmt: however, they are so expensive you really are better off getting an stm32
11:13 AM LeoNerd: 1284 has 16KiB RAM
11:13 AM vmt: 1284 has 16 k ram, 128 k flash
11:13 AM vmt: and the rest, obviously. however, they are quite expensive and in general not worth it
11:24 AM twnqx: well, i'd prefer 32bit (arm) anyway :P
11:25 AM twnqx: and well, 44 pins are roughly 36 more than i want
11:25 AM twnqx: i really meant an 8 pin device :P
11:31 AM LeoNerd: To a rough degree, more RAM comes with more flash and more IO pins. I think it's unlikely to find a 16Ki RAM chip with only 8 pins
11:34 AM twnqx: i agreee, i found a 14pin so far
11:35 AM nohit: use st's STMCUFinder
06:36 PM [1]MrMobius is now known as MrMobius
07:08 PM rue_mohr: https://www.nongnu.org/avr-libc/user-manual/group__avr__interrupts.html
10:07 PM rue_mohr: so, todays question
10:07 PM rue_mohr: is "can I do serial recieve on a tiny13, AND control 2 hobby servos"
10:07 PM rue_mohr: :) or even 4
10:08 PM rue_mohr: there is time in the pulse timing to do 4
10:08 PM rue_mohr: but
10:08 PM rue_mohr: the timer is occupied doing the serial
10:08 PM rue_mohr: a servo pulse is 24 bits at 9600
10:10 PM rue_mohr: its palusable that I could use the adc interrupt for timing the servo pulse, but I dont know how stable it is
10:10 PM rue_mohr: at 9.6Mhz,
10:11 PM rue_mohr: for 8 bit I'd be targeting ~ 102Khz
10:12 PM rue_mohr: if the adc clock is divided by 8 its close
10:12 PM rue_mohr: tho I'm not sure If I'm getting /13 or /12 from the adc
10:12 PM rue_mohr: it came out a bit higher than expexted
10:13 PM rue_mohr: I suppose a test is in order
10:15 PM rue_mohr: the lack of lightoutside is really demotivational
10:15 PM rue_mohr: and the lack of life in here
10:15 PM rue_mohr: I wonder if Emil would like to have all the rues in his channel too
10:15 PM rue_mohr: would that be rued?
11:02 PM rue_shop3: well thats odd
11:02 PM rue_shop3: according to this, the actual clock is about 6.5Mhz
11:05 PM rue_shop3: ah no
11:09 PM rue_shop3: 9 adc interrupts per timer overflow, with the adc clock set to /8
11:09 PM rue_shop3: and the timer set to /1
11:12 PM rue_shop3: 4.5Mhz
11:13 PM rue_shop3: why is an avr thats supposed to be running at 9.6Mhz running at ~4.8
11:17 PM day__ is now known as day
11:25 PM rue_shop3: well I must have made a horrid error, and the serial code is *just* getting by at 0.5x the clock rate I expected I was using
11:29 PM rue_shop3: timing doesn't seem any less stable than using the itmer
11:29 PM rue_shop3: +- interrupts knocking heads
11:31 PM rue_shop3: so its closer to 14 cycles/conversion
11:37 PM rue_shop3: hah, ok, so, I can use the adc to time my uart bits, maybe
11:38 PM rue_shop3: if I use /8 and clock off 4 interrupts per bit
11:38 PM rue_shop3: then I can use the timer for the servo pwm and not have jitter
11:39 PM rue_shop3: unless that internal 4.8Mhz is so unstable it shows thru
11:58 PM rue_shop3: hmm, the only way I see of making the timer work out for the servo pulses is to screw severly with osccal
11:59 PM rue_shop3: :/ or more overflow counting