#avr Logs

Feb 22 2017

12:00 AM rue_house: if for while char int long unsigned printf return
12:04 AM Rickta59: thanks
12:21 AM daey_ is now known as daey
12:39 AM _ami_: chisight: able to configure stm32 dev environment yesterday. blink code works well. :)
12:39 AM _ami_: using libopencm3
12:40 AM chisight: great progress.
12:42 AM chisight: libopencm3 looks nice. too bad it doesn't have an m0 version.
12:43 AM chisight: oh, i guess it does.
12:44 AM chisight: will have to look closer at that. been using hal and not thrilled.
03:51 AM l9_ is now known as l9
04:25 AM skz81: _ami_, hello
04:25 AM _ami_: skz81: yo
04:26 AM skz81: in the link you provided, in chapter "Implementing a PWM driver"
04:27 AM skz81: it sounds clear to me that the concept of "chip" maps with a driver implementation
04:27 AM _ami_: its not mapped to a PWM controller instead?
04:27 AM _ami_: i hv a board which has 3 timers and each timer output has 2 pwm outputs.
04:28 AM skz81: yup I noted that
04:28 AM _ami_: so i was wondering should i create 3 pwm chip in driver implementation and add 2 pwms to it
04:28 AM skz81: As I understand things, that's ONE chip with 6 outputs
04:28 AM _ami_: or create a pwm chip and add 6 pwms
04:28 AM _ami_: aha,
04:28 AM _ami_: hmm, ok
04:28 AM Jartza: I just realized I have 128kB SPI ram chips
04:29 AM Jartza: maybe I use those for my attiny85-basic machine
04:29 AM Jartza: they are also nice 8-pin chips
04:29 AM skz81: BUT y'oull need to somehow implemented the fact that output 1-2 have same freq & polarity but CAN HAVE different duty cycle
04:29 AM skz81: erm
04:30 AM skz81: BUT you'll need to somehow implement... rather
04:31 AM skz81: _ami_, that just my own understand at macro level, I MAY be totally wrong, though
04:31 AM skz81: understandING* (ouch I'm struggling this morning)
04:34 AM skz81: _ami_, there are maybe technical reasons to choose rather the other way (it may simplify the implementation, I don't know so far)
04:35 AM _ami_: i think i would go with a pwm chip and 6 outputs
04:35 AM _ami_: should be simpler for user to use.
04:35 AM _ami_: sysfs dir structure for this pwm would look well.
04:36 AM _ami_: skz81: thanks for your suggestions :)
04:37 AM skz81: Yes that the spirit : try to understand what are the side-effect on such subsystems (sysfs, user API, ....) for both choice before actually implementing (or implement both if you have time :p )
04:38 AM skz81: What I understand is the "chip" concept in the drivers allow to have multiple soft/hard implementation cohabiting in the same system
04:38 AM skz81: that was not possible before
04:38 AM _ami_: aha, good point
04:38 AM _ami_: noted down.
04:39 AM skz81: s/in the drivers/in the kernel/ but you understood the idea.
04:39 AM _ami_: yeha
04:40 AM skz81: Note also : a same driver can have multiple "instances" if you have multiple ressources.
04:40 AM skz81: the instance number is the minor number of devices (/dev)
04:40 AM skz81: where the major is the number of the driver "class" in the loaded modules list
04:41 AM skz81: I say that because you ask : "<_ami_> its not mapped to a PWM controller instead?"
04:42 AM skz81: the controller model's maps to the "chip" class, the actual controller maps to an instance ;))
04:42 AM skz81: ;�)
04:46 AM _ami_: aha,
04:46 AM _ami_: makes sense
09:24 AM hetii: Hi :)
09:24 AM hetii: Q: any clue if nand gates like DL003D (probably the same type like 74LS03) are capable to act as a buffer for data line in WS2812b, or they to slow ?
09:24 AM LeoNerd: Check the speed specs on the data sheet
10:09 AM twnqx: i would be more worried about them being LS-TTL and not cmos
10:10 AM twnqx: go fot ACT/HCT/AHCT compatible ones
10:10 AM twnqx: for*
10:10 AM LeoNerd: From experience: the 2812 is quite tolerant of off-spec voltage levels, but it does want fast transition edges
10:11 AM LeoNerd: If you have a long cable with lots of stray L+C factors, you smooth out those edges and can lead to missed clocks
10:12 AM LeoNerd: I'd suggest getting one of those little SOT-23 sized "1G" logic gates to reshape it again
10:13 AM twnqx: yeah, those are nice
10:14 AM LeoNerd: I have a handy stock of 1G 4066 (analog switch), and 125 (active-low tristate buffer) for similar purposes
10:14 AM LeoNerd: A 1G125 is a great way to make a "not quite SPI" chip SPI-compliant, by giving it a hiZ state on the serial output pin
10:15 AM LeoNerd: A 1G66 (the 4066 analog switch) is useful for isolating part of an I²C bus on the SDA line. E.g. because you want to use an ATtiny USI for both SPI and I²C and you're a tight bastard who can't afford a bigger chip ;)
12:22 PM Steffann is now known as Steffanx
01:32 PM mindw0rk: Hey guys, I gots a question. I want to find a uC that would work as a usb programmer for this parallel flash IC (http://www.tme.eu/en/Document/57bcabc230b43ed1bf2ed146dfbfa1b4/SST39xF512_010_020_040.pdf). I guess the hard requirements are that I need 24 I/O pins (16 address, 8 read/write for values) and I don't know whether to choose a hardware usb interface
01:32 PM mindw0rk: or use V-USB. Any ideas from the top of your heads?
02:08 PM LeoNerd: mindw0rk: I'd consider the good ol' 32U4, real USB... stick the 8 IO lines on one of the 8bit IO ports and hang the 16 output-only address lines off two 74'595 shift registers on SPI.. or maybe an I²C expander.. slower though
02:08 PM LeoNerd: But perhaps I'm biased because I'm working on the 32U4 right this moment
02:08 PM bss36504: The 32U4 is pretty sweet
02:08 PM bss36504: Very nice little chip
02:09 PM LeoNerd: It's not *great*... one thing that really annoys me about it is the muxing of the four interrupt lines, with the two I²C and UART pins
02:09 PM LeoNerd: Meaning: if you want to use any I²C sensor that has an interrupt pin, say, you've immediately lost half of the interrupt pins
02:09 PM bss36504: Ok fair enough
02:09 PM LeoNerd: WAnt to talk I²C and async serial? OK then no interrupts for you
02:09 PM bss36504: haha
02:10 PM bss36504: I wonder why they did that...
02:10 PM LeoNerd: You *could* still use the PCINTs, but there's only *one* bank of those on this chip; being muxed on PORTD
02:10 PM LeoNerd: PORTD also has the SPI port
02:10 PM LeoNerd: None of PORTs B, C, E or F support PCINT
02:11 PM crazy_imp: LeoNerd: why the 595s? you got 26 IOs
02:11 PM LeoNerd: Eh.. maybe
02:11 PM LeoNerd: I tend to always move "simple" IOs like that out onto secondary chips.. almost always it turns out to be the right design
02:12 PM LeoNerd: E.g. I've always put a bunch of buttons/LEDs for front panels off a PCF8574 or similar... leaves me more spare IO lines on the chip for faster things
02:12 PM LeoNerd: Plus I get debounce for free by the interrupt => I²C transaction delay ;)
02:13 PM LeoNerd: But then I mostly make small one-off runs of things for personal use.. perhaps if I was designing it as a commercial product, I'd go for a single bigger chip
02:13 PM bss36504: Usually for mass production fewer BOM lines is worth more than the slightly higher cost of a larger mcu
02:13 PM crazy_imp: i would prefer 595s over i2c portexpanders (price and speed)
02:14 PM mindw0rk: yeah don't want to add to much stuff to this design, because it's space constrained and will probably need to have other components on it on later iterations of it.
02:14 PM bss36504: Then a larger MCU is probably a better option
02:15 PM crazy_imp: but if the uC should only do one thing and it needs 24 IOs of 26 and only some code to output the address on the different ports...
02:16 PM bss36504: If that's all it has to do then you'd be fine, sure
02:17 PM mindw0rk: yeah, I just realized it all has to run @ 5v, so that's a few extra dimes to each of those components.
02:17 PM mindw0rk: anyway, thank you, everyone.
03:34 PM Lambda_Aurigae: atmega32 and a usb-serial adapter?
03:34 PM LeoNerd: Depends what USB protocol you want to talk...
03:35 PM crazy_imp: and how fast...
03:35 PM LeoNerd: Yah that too :)
04:32 PM arij is now known as arij_work
05:07 PM plouf: any idea where i can find "avrprog" manual says , for manual upgrading STK500
05:07 PM plouf: is not installed with VRStudio7
05:07 PM plouf: and not found in atmelsite using search functin
05:13 PM Tom_itx is now known as Tom_L
09:27 PM JanC is now known as Guest31516
09:27 PM JanC_ is now known as JanC
10:05 PM arij_work is now known as arij