#avr Logs

Feb 27 2017

12:16 AM daey_ is now known as daey
02:19 AM Polprog: what c version does avr-gcc use by default? c98?
02:35 AM ub|k_ is now known as ub|k
02:43 AM xentrac: c89
02:56 AM Polprog: xentrac, thx
04:27 AM NoHitWonder: does avr-gcc support boolean type?
04:29 AM malinus: NoHitWonder: #include <stdbool.h>
04:29 AM NoHitWonder: yeah, its same with amrs
04:30 AM NoHitWonder: *arms
04:30 AM specing: sad
04:30 AM specing: if you must use a shitty language then at least use the latest standard of it
04:31 AM NoHitWonder: C is shitty?
04:31 AM NoHitWonder: what would you prefer?
04:32 AM specing: Ada
04:32 AM NoHitWonder: okay
04:32 AM specing: and yes, C is shitty
04:33 AM NoHitWonder: if i remember correctly, there was some embedded design competititon with Ada last summer
04:33 AM specing: yes, makewithada
04:34 AM NoHitWonder: did you participate?
04:34 AM specing: no
04:36 AM NoHitWonder: sad
04:36 AM specing: yes. But I did other embedded work in a HDL based on Ada (VHDL)
04:36 AM NoHitWonder: nice
04:37 AM NoHitWonder: i have done a simon clone with VHDL at school
04:37 AM NoHitWonder: back in the day when i was studing electronics
04:38 AM NoHitWonder: i didnt know vhdl is based on ada
04:40 AM specing: well, now you do.
04:40 AM specing: and knowing Ada helps tremendously when writing VHDL
04:41 AM NoHitWonder: ok
04:43 AM NoHitWonder: my co-worker is planning to do a SID chip with VHDL
04:48 AM NoHitWonder: specing http://papilio.cc/ is this a good dev kit?
04:50 AM specing: no
04:50 AM specing: its nowhere close to open source due to Xilinx fpga
04:51 AM specing: and the chip they use is not even supported in their latest IDE
04:51 AM specing: get the ICEstick if you want fully open toolchain
04:52 AM specing: or a terasic DE*SoC for great cheap Altera SoC-FPGA devkits
04:53 AM NoHitWonder: okay
04:55 AM NoHitWonder: i do have this https://www.terasic.com.tw/cgi-bin/page/archive.pl?No=30
04:56 AM NoHitWonder: but im looking for some lighter option
04:56 AM NoHitWonder: possibly open source
04:56 AM NoHitWonder: and i dont need all that extra hardware
05:02 AM NoHitWonder: well, i need to go, good day to you, specing ->
07:45 AM julius: hi
07:48 AM julius: got a MCP1407 here (mosfet driver) it says that anything belo 0.8v as input is "low" and high is 2.4v and upwards. im feeding it a pwm signal from a atmega32 which is between 0,47v and 1.1v each 5 seconds long. but the output is always on 5v? datasheet: http://ww1.microchip.com/downloads/en/DeviceDoc/20002019C.pdf
07:49 AM julius: ive been using these pwm values successfully to drive the mosfet directly for a simple led between + and drain of the mosfet
07:49 AM julius: results in two different brightness levels
07:54 AM LeoNerd: Why is your PWM otuput only going up to 1.1V?
07:54 AM LeoNerd: You want to be swinging rail-to-rail for a good clean signal into the FET driver
07:55 AM julius: haven really thought about that yet
07:56 AM LeoNerd: I'd suggest that's the first thing to look into. If your PWM output from the microcontroller isn't able to rail-swing, that suggests a heavy loading on the output
07:56 AM julius: also the mosfet driver does limit my output values, while without the mosfet i could generate anything between 0-5v (depending on what voltage the mosfet goes on) now im limitted to 2.4v up to 5...right?
07:57 AM LeoNerd: Usually a FET driver is a fairly binary switching affair; either off or on. Not really a linear thing
07:58 AM julius: ah yes, it will always output 5v when its on
07:58 AM julius: but when it is driven by pwm the voltage over time on the output can still be lower than 5v right?
08:00 AM dunz0r: julius: Yes, it'll even out to less than 5V.
08:01 AM LeoNerd: The /average/ over some time, sure
08:01 AM julius: ok
08:01 AM julius: so i could attach a led and "dimm" it?
08:01 AM LeoNerd: But at any instant in time it ought to be close to GND, or close to VCC
08:01 AM LeoNerd: Yes that's the entire point of PWM
08:02 AM dunz0r: julius: With a LED it will probably be GND or VCC, but it will look like it's dimmed, since the switching is done so crazy fast.
08:03 AM LeoNerd: (Though I'd suggest you probably don't need this MCP FET driver chip, if you're just controlling an LED)
08:03 AM dunz0r: Depends on how much current the LED uses
08:03 AM * dunz0r has a really beefy LED that he plans to make a bikelight out of
08:03 AM dunz0r: But I need to build a constant voltage/constant current-driver first.
08:04 AM LeoNerd: Right - but the AVR could just directly drive the gate of whatever actual switching FET you're using to switch the LED
08:04 AM LeoNerd: You wouldn't need a FET driver chip in front
08:04 AM avrdude: is anyone familiar with FatFS?
08:05 AM * LeoNerd 's written AVR code for FAT16 before
08:05 AM avrdude: why does http://elm-chan.org/fsw/ff/en/dwrite.html say "Remarks: Application program MUST NOT call this function, or FAT structure on the volume can be collapsed."
08:05 AM LeoNerd: Oh, that's a particular implementation thereof.. notsure then - I wrote my own ;)
08:06 AM avrdude: say the application did use disk_write, would it be faster than using f_write?
08:06 AM avrdude: oh :x
08:10 AM Lambda_Aurigae: avrdude, one of them is a direct write to a block on the disk...not a good idea if you are using fat filesystem.
08:11 AM Lambda_Aurigae: disk_write writes directly to a sector on the drive.
08:11 AM avrdude: why not? here he even shows an example on how to allocate space for a file, then write to it using disk_write: http://elm-chan.org/fsw/ff/en/expand.html
08:11 AM Lambda_Aurigae: if you use that you best make sure you know what you are doing.
08:11 AM LeoNerd: The way I did it, was to keep two blocks of disk buffer cache
08:11 AM LeoNerd: (Two blocks because that's half of the RAM I have on a 328P ;) )
08:12 AM Lambda_Aurigae: you can corrupt the filesystem if you write to the wrong block and overwrite something that's already used by the filesystem.
08:12 AM avrdude: but i guess if i have used expand first, like in the example, it should be safe right?
08:12 AM Lambda_Aurigae: f_write writes to a file..using an open file handle which is handled through the filesystem handler/driver.
08:13 AM Lambda_Aurigae: if you know what you are doing in writing to a drive with an active filesystem, then go for it.
08:13 AM Lambda_Aurigae: if you have to ask such questions I would assume you have no clue.
08:14 AM Lambda_Aurigae: yes, it is possible to write to the file using low level disk function disk_write
08:14 AM Lambda_Aurigae: it is dangerous
08:14 AM Lambda_Aurigae: if you write to the wrong sector you could be writing to another file or to something in the directory structure.
08:14 AM Lambda_Aurigae: which can cause corruption of the filesystem.
08:15 AM avrdude: but nothing that will permanently damage the disk..
08:15 AM avrdude: i have no idea what i'm doing, but i'll try anyway :^)
08:15 AM Lambda_Aurigae: no..you can always reformat it.
08:16 AM Lambda_Aurigae: you see that line where they calculate lba?
08:16 AM Lambda_Aurigae: do you understand that line?
08:19 AM julius: LeoNerd, indeed i dont need this driver for a simple led
08:19 AM julius: but i actually want to drive a 10w led
08:20 AM julius: the 30ma led is just for testing
08:20 AM julius: still, the issue remains. my multimeter says its 0.47v at the input for ~5 seconds (thats what i set in the code) but still the drivers output is always at 5v
08:21 AM julius: dunz0r, thanks
08:21 AM LeoNerd: Ah, a multimeter might be averaging the result though
08:21 AM Lambda_Aurigae: julius, most multimeters aren't going to read a pwm output properly.
08:21 AM LeoNerd: If it's PWM at some kHz frequency, the meter won't see that
08:21 AM Lambda_Aurigae: I like to use a quick responding analog meter for those.
08:21 AM avrdude: Lambda_Aurigae: i don't even know what lba is.
08:22 AM Lambda_Aurigae: avrdude, then don't be using disk_write or you will likely end up formatting the drive multiple times...
08:22 AM julius: yeah its a cheap one
08:22 AM Lambda_Aurigae: in this case it's logical block address...
08:22 AM Lambda_Aurigae: julius, it probably won't read a PWM output properly.
08:23 AM julius: so i might be over +2.4v which the driver accepts as high
08:23 AM Lambda_Aurigae: cheap 8 dollar analog meters work well for getting a good approximation though.
08:23 AM Lambda_Aurigae: or an oscilloscope so you can see the pulses.
08:23 AM julius: is about 40€ i guess
08:24 AM Lambda_Aurigae: you are seeing an average..not the specific pulse max voltages.
08:24 AM Lambda_Aurigae: to see those you will need an oscilloscope.
08:25 AM julius: ok thanks
08:25 AM julius: need to get this hantek going with linux first :(
08:25 AM Lambda_Aurigae: avrdude, I know this is going to sound mean and evil and whatnot,,,before starting to use the low level fatfs routines I recommend reading the entire website, every link, every page. Also read the fat file system specifications.
08:26 AM LeoNerd: I've often considered making a little low/avg/high voltmeter that has a temporal resolution up into the 100s of KHz at least, if not MHz
08:26 AM Lambda_Aurigae: heck, for that matter, before starting with fatfs, read the entire page, all links, all pages.
08:26 AM Lambda_Aurigae: LeoNerd, 100KHz shouldn't be too difficult really.
08:27 AM avrdude: ive been using it for a while, but i want to improve the write-speeds
08:27 AM LeoNerd: Lambda_Aurigae: couple of fast-decay peak detectors on opamp+diode configuration, and a 3-channel ADC
08:27 AM Lambda_Aurigae: avrdude, on your sd card, what interface are you using?
08:27 AM avrdude: spi
08:27 AM Lambda_Aurigae: there's your first problem.
08:28 AM Lambda_Aurigae: that is the slowest possible method to read/write that sd.
08:28 AM LeoNerd: SPI is slow for SD cards, but easy to implement
08:28 AM Lambda_Aurigae: LeoNerd, yup.
08:28 AM LeoNerd: I only needed a few bytes per second in aggregate for my datalogger, so it was fine fo rme
08:28 AM avrdude: ok great. so maybe it would be a good idea to atleast maximise the write-speed then?
08:28 AM LeoNerd: I store a line of text once per second, and a line of battery log once per 30. and that's it
08:28 AM Lambda_Aurigae: avrdude, maximize the speed of your interface first.
08:29 AM avrdude: ..how
08:29 AM Lambda_Aurigae: using disk_write isn't going to be noticable to most human interfaces....granted, this depends on what you are doing too.
08:30 AM avrdude: quick question: what happens if i call disk_write(drv, buffer, lba, 2048); twice? does it overwrite?
08:30 AM Lambda_Aurigae: run it in 4 bit sd mode maybe?
08:30 AM Lambda_Aurigae: yeah.
08:31 AM avrdude: hm?
08:31 AM Lambda_Aurigae: you are writing the same buffer data to the same drive at the same logical block address.
08:31 AM Lambda_Aurigae: might I suggest you read the page that tells exactly what disk_write does?
08:32 AM avrdude: how about.. disk_write(drv, buffer+2048, lba+2048, 2048); ?
08:33 AM hetii: Good morning :)
08:33 AM Haohmaru: hetii morning? it's 4pm
08:33 AM * Haohmaru overwrites avrdude
08:33 AM avrdude: ;_;
08:34 AM hetii: Hmm, right but depend on timezone :)
08:34 AM Lambda_Aurigae: oh goog glub.
08:34 AM Lambda_Aurigae: sorry, but if you can't figure that out on your own I'm not going to be able to help.
08:35 AM Lambda_Aurigae: read the manual...in this case the webpage.
08:35 AM hetii: I have small challenge, have some amplifier that is controlled over some infrared protocol. but the question is how to guess whit one is used there
08:37 AM hetii: I think about using IRMP project, to use some brute force method and iterate over all protocols address and commands
08:37 AM hetii: but the amount of the combination make it not much usuable
09:29 AM julius: Lambda_Aurigae, i did setup the hantek usb oscilloscope on my windows machine now. this is the output: http://junghans.online/public/1.small.jpg and these are the settings: http://junghans.online/public/3.small.jpg i got some trouble with the duty cycle, it is set to 10us/div but the raster behind the data isnt really aligned to it
09:31 AM julius: a wait, let me read up about the triggering feature
09:32 AM xentrac: julius: what software are you using on the Linux side to drive the scope? sigrok?
09:37 AM julius: currently windows
09:37 AM julius: just wanted to get going
09:38 AM julius: xentrac, is sigrok on linux good?
09:38 AM Jartza: hallo
09:41 AM _ami_: Jartza: hey
09:42 AM _ami_: Jartza: have you done i2c-slave bitbang on attinys?
09:42 AM julius: ah...one can move the signal to match the raster with the mouse cursor. so time per division is 10us, signal goes over ~4 divisions and is +5v(high) in 1 division as can be seen here: http://junghans.online/public/1.small.jpg (did not move the signal in that one) looking at the picture i would guess that the duty cycle is about 25%?
09:44 AM LeoNerd: _ami_: I've done USI, but not /raw/ bitbang
09:47 AM _ami_: LeoNerd: cool. i have attiny13as.. and i was able to do i2c-bitbang master: https://raw.githubusercontent.com/amitesh-singh/amiduino/master/avr/attiny13a/i2c-master/i2c-master.c
09:47 AM * LeoNerd nod
09:47 AM _ami_: since t13a does not have interrupts.. so i guess for i2c-slave, i need to poll
09:47 AM _ami_: for scl reads
09:47 AM LeoNerd: No PCINT? Yah.. it's a pretty tiny chip
09:49 AM * _ami_ checks t13a DS again. surely there was no INTX.. need to verify if PCINTs are there.
09:50 AM _ami_: LeoNerd: PCINTs are there.. cool!
09:51 AM LeoNerd: :)
09:57 AM LeoNerd: Yah - I'm not aware of an AVR chip that totally lacks any sort of PCINT ability
09:57 AM LeoNerd: Even the 32U4, subject of many of my complaints, does have /some/
09:57 AM LeoNerd: Granted they're all in a stupid place and collide with other peripherals and there's nowhere near enough, but it does have some
10:39 AM xentrac: julius: I don't have any experience with sigrok. it sounds super awesome from the website, but I don't have experience with it, so I was hoping to hear from someone who did :)
10:40 AM xentrac: (or with an alternative)
10:40 AM xentrac: LeoNerd: what are your complaints about the 32U4?
10:41 AM LeoNerd: Hah.. many
10:41 AM LeoNerd: The one I'm referring to /here/ is that the 32U4 has very few interrupt-capable pins and they're all in stupid places
10:41 AM LeoNerd: INT0/1/2/3 are muxed onto the UART and I²C module pins
10:41 AM LeoNerd: If there's ever one situation where I most want interrupt pins, it's when I'm dealing with I²C sensors.. So whythefuck do we mux interrupt and I²C on the same pins?
10:42 AM LeoNerd: But that's OK, ignoring the dedicated INT lines, there's also the PCINT port. Of which the 32U4 has only 8 PCINT-capable pins on one port (D), where most ATmega chips have them available on most of their ports
10:42 AM LeoNerd: PORTD also carries the SPI module - again, if I'm using SPI I'm quite likely to want some interrupt pins
11:13 AM Jartza: _ami_: umm, I've done some simple stuff yes
11:13 AM _ami_: i2c slave purely bitbang?
11:13 AM Jartza: umm, no
11:13 AM Jartza: using USI
11:16 AM LeoNerd: I can't think of a problem I'd be solving that was so small I could get away with a tiny13, and yet need I²C
11:16 AM LeoNerd: The only thing I've made on a tiny13 so far is a little WS2812 LED driver that makes different colours depending on analog levels presented to it
11:17 AM LeoNerd: https://www.tindie.com/products/leonerd/4-way-identifying-continuity-tester/ <== one of these
11:17 AM LeoNerd: (OK so that LED is actually an APA-106, but it's electrically identical)
11:22 AM xentrac: have you seen the NeoPixels?
11:22 AM LeoNerd: Those /are/ WS2812s
11:22 AM xentrac: uh
11:22 AM xentrac: sorry
11:22 AM xentrac: so I se
11:22 AM LeoNerd: Also the dotStar is an APA-102
11:23 AM xentrac: what I was trying to say is that if you wanted to make a thing that was similar to a NeoPixel, starting with an RGB LED, an ATTiny13 with I²C might be a reasonable way to do it?
11:23 AM LeoNerd: This latter is driven by a signal that's a bit like SPI, in that it has separate serial data and clock lines. A lot easier to drive (because you can literally write it from an SPI port)
11:24 AM LeoNerd: I can't imagine why I'd want to make one of those when I can already get 2812s and similar :)
11:24 AM xentrac: maybe instead of red, green, and blue, you want infrared, violet, and ultraviolet
11:25 AM LeoNerd: OK, so in that case I'd get a WS2811 driver chip and stick my own LEDs on it
11:25 AM xentrac: oh, you can do that?
11:25 AM xentrac: that would be better!
11:25 AM LeoNerd: Yes; the WS2812 is the driver chip embedded inside an LED package with three actual LED junctions attached to it
11:25 AM LeoNerd: The WS2811 is the driver chip on its own, in a SOIC(?) package..
11:25 AM Chillum: good idea
11:25 AM Chillum: I never thought of that
11:25 AM xentrac: I thought they were all in the same package
11:26 AM xentrac: like https://learn.adafruit.com/assets/10668
11:26 AM LeoNerd: Yeah, you can clearly see the controller die and the three LED junctions on this photo
11:26 AM xentrac: yeah
11:27 AM LeoNerd: APA-106 is the same thing, but built into a 5mm leaded LED package instead of a 5050 SMD
11:27 AM LeoNerd: I use those on the continuity probe
11:27 AM xentrac: cool!
11:30 AM xentrac: still it seems like the Tiny13 gives you four GPIO lines after you spend two on I²C, and you can talk to 128 of them over a two-wire I²C bus
11:32 AM xentrac: with ADC and PWM
11:33 AM xentrac: that seems like it could be useful for sampling data or for driving things taht aren't LEDs
11:35 AM _ami_: t13a has a int0 too.. what i am smoking these days! huh! should have known.. now its a pleasant surprise for me!
11:37 AM LeoNerd: :)
11:41 AM xentrac: This might be naïve, but for anything other than mass production, wouldn't it make sense to use an ATTiny40 these days for anything you might have considered an ATTiny13 for? They're both 3 mm square in VQFN, but the 40 costs $1.08 to the 13's 54¢ and has three times the GPIO lines, plus SPI and TWI hardware
11:42 AM xentrac: maybe that's kind of what you were saying about "so small I could get away with a tiny13"
11:43 AM LeoNerd: Yah; for anything production-aimed there's sortof a size limit, below which AVR chips actually get more expensive just due to low volume
11:43 AM LeoNerd: So you end up finding bigger chips than you need are /still/ cheaper
11:43 AM xentrac: oh, I hadn't even thought about that
11:44 AM LeoNerd: Also compare tiny85 to tiny25 or 45
11:45 AM xentrac: I meant that if you're making like 1–200 of something, the more capable chip will probably be less headache, and using the less capable chip instead will only save you like $25 or something
11:46 AM xentrac: I guess this is the reason that a lot of people are just using LPC11xxs instead of AVRs nowadays and we're stuck with people coming in here to troll about Ada
11:46 AM specing: What does Ada have to do with your choice of MCUs?
11:49 AM xentrac: huh, I didn't realize the 85 came in a 20-pin QFN. with seven "DNC" pins!
11:51 AM xentrac: still only six GPIOs
11:52 AM LeoNerd: five-and-a-half
11:52 AM LeoNerd: (RST)
11:53 AM xentrac: heh
11:54 AM xentrac: it looks like the ATTiny20 is the same price as the ATTiny13, but has 12 GPIOs like the 40
11:56 AM xentrac: and I²C and SPI hardware. I guess "TWI" is Atmel's name for I²C so they don't have to pay a trademark license fee to Philips?
11:57 AM LeoNerd: Hrm.. I wonder how the tiny20 compares to the tinyx4
11:57 AM LeoNerd: Ah, very similar. identical pinout
11:58 AM LeoNerd: Hm.. it appears to basically be a tiny24, but with real SPI and I²C modules instead of that USI
11:58 AM xentrac: it's kind of crazy that they pay for all those extra pins in the VQFN package and then don't bother to add a couple of transistors on the chip to drive the pins
11:59 AM LeoNerd: Hah! but they've muxed the I²C *and* the SPI modules onto the same physical pins
11:59 AM LeoNerd: How silly
11:59 AM xentrac: oh shit
11:59 AM xentrac: really?
11:59 AM xentrac: that's hilarious
12:00 PM LeoNerd: SCL and SCK both share PA7; SDA and MOSI both share PB1
12:02 PM xentrac: so if you want to interface an SPI bus and an I²C bus, this is not the part fo ryou
12:02 PM LeoNerd: Still.. if you have a 74HVC1G66 to hand, you can mux SPI and I²C off the same physical pins easily enough
12:02 PM xentrac: or use the SPI for the other silly things we can use SPI for
12:02 PM xentrac: hmm, I don't know what a 74166 is
12:03 PM LeoNerd: Put the 1G66 (an analog switch) between the SDA pin and the actual bus... with the pullup on the bus side
12:03 PM xentrac: ah
12:03 PM LeoNerd: Gate the switch from the SS line of the SPI slave (or diode-OR them if multiple). Upshot is that if no SPI slave is selected, analog switch is closed and you can talk I²C across it; SPI slaves will ignore you
12:04 PM LeoNerd: Assert the SPI slave by pulling SS low, and analog switch will open, cutting off the SDA line, so no I²C slave will see a START condition on the wire, and hence will sit idle while the SPI slave is used
12:04 PM _ami_: xentrac: attiny13 and attiny20 have price difference.
12:04 PM xentrac: Hmm!
12:04 PM xentrac: Awesome!
12:05 PM _ami_: 10 t20 - 15$ , 10 t13a - 5$
12:05 PM xentrac: This is all assuming you're the SPI master and the I²C master?
12:06 PM LeoNerd: Yah, true
12:06 PM LeoNerd: Obviously if either module is to be a slave, you do need dedicated IO pins for those
12:06 PM xentrac: _ami_: Digi-Key will sell you ten ATTiny20s for US$5.60 right now
12:06 PM _ami_: wow!
12:06 PM LeoNerd: What package size?
12:06 PM xentrac: http://www.digikey.com/product-detail/en/microchip-technology/ATTINY20-SSU/ATTINY20-SSU-ND/2357326 for example
12:07 PM xentrac: either VQFN or SOIC
12:07 PM xentrac: actually I think they charge 1¢ more for the QFN
12:08 PM LeoNerd: Huh.. How odd
12:08 PM xentrac: Am I the only one who feels sad every time they see "Manufacturer: Microchip Technology"
12:09 PM LeoNerd: A break from usual AVR naming - the ATtiny40 is *not* just a 4Ki version of the ATtiny20
12:09 PM LeoNerd: It's a totally different kind of chip
12:09 PM LeoNerd: 20 pins instead of 14; has EEPROM which t20 lacks
12:10 PM xentrac: Yeah, that was confusing me too
12:10 PM LeoNerd: https://en.wikipedia.org/wiki/Atmel_AVR_ATtiny_comparison_chart is super-handy by the way
12:14 PM xentrac: oh awesome!
12:17 PM avrdude: >Attiny4
12:17 PM avrdude: what in the...
12:18 PM LeoNerd: Oh the tiny4.. Now /that/ chip is truely deserving of the name 'ATtiny'
12:18 PM avrdude: it comes in a sot23 package lmao what is this
12:18 PM LeoNerd: Not like those pretenders like the tiny88 or the 1634
12:25 PM xentrac: even so, the 88 in a 32vqfn is 4mm × 4mm, quite a bit smaller than the 4.65mm × 4.65mm × 3.54mm 2N7000 I have here
12:28 PM xentrac: like, you could fit three of them in the space of this single transistor
12:51 PM NoHitWonder: is KiCAD worth learning? i used to use eagle in the past but it seems it's autodesk product now
12:51 PM cehteh: yes
12:52 PM Chillum: I think I will try kicad out right now
12:52 PM Chillum: I have been using eagle but I really should try out the competition
12:56 PM NoHitWonder: i downloaded autodesk eagle 8.0.2
12:56 PM NoHitWonder: im gonna try it first
12:56 PM NoHitWonder: if i can avoid learning new software
12:57 PM cehteh: the eagle free version sux
12:57 PM xentrac: it's proprietary; sooner or later it will fuck you over
12:57 PM cehteh: took me 5 minutes :D
12:58 PM cehteh: discovered that some feature i wanted was not enabled .. blarg .. bye
12:58 PM cehteh: cant remember what, was something simple not brazillion layers or huge pcb .. only 2 layer, thumbnail size, something different
12:59 PM xentrac: that's not too bad. at least it didn't cut off your access to designs you'd already made until you paid them, which is the kind of thing that proprietary software vendors do at times
01:04 PM Chillum: cehteh: I think it is just dimension limits and 2 layer limit
01:04 PM Chillum: oh, and the annoying thing about not being allowed to place parts outside of the dimensions
01:04 PM Chillum: even for a moment
01:09 PM Tom_L: for what software?
01:10 PM Tom_L: oh gawd... i forgot autodesk took over that product. i bet they ruin it
01:10 PM Chillum: eagle free version
01:10 PM Tom_L: yeah 2 layers and size
01:11 PM Chillum: I have not ran into those limitations myself. I can't afford large boards
01:11 PM Tom_L: plenty big for most hobbyists
01:11 PM Chillum: I do tiny runs and oshpark is really the only option for 3 copies of a board
01:11 PM Chillum: and large boards cost a fortune from them
01:11 PM Tom_L: tried mitch yet?
01:11 PM xentrac: mitch?
01:12 PM Chillum: mitch?
01:12 PM Tom_L: i forget his #channel
01:12 PM Tom_L: yeah he does china boards
01:12 PM Tom_L: an aussie
01:12 PM Tom_L: quite pleased the times i used him
01:12 PM xentrac: oh, I thought maybe you meant Altman
01:12 PM Tom_L: no
01:13 PM Tom_L: i've used laen a few times for one off boards
01:14 PM Tom_L: (oshpark)
01:14 PM Tom_L: just don't like purple personally
01:15 PM Tom_L: yeah that's it... #hackvana
01:15 PM Tom_L: (mitch)
01:15 PM Chillum: love the purple
01:16 PM Tom_L: wanna see the first purple run?
01:16 PM Tom_L: wasn't even close
01:16 PM Chillum: sure
01:16 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/boards/atmega32u4/atmega32U4_3.jpg
01:17 PM Tom_L: so that was a while back
01:18 PM Chillum: not as pretty as today
01:20 PM Tom_L: i know
01:20 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/boards/atmega32u4/atmega32u4_1.jpg
01:20 PM Tom_L: i'm a little partial to blue or the standard green or red
01:21 PM xentrac: the ODrive guy is enthusiastic about how his latest run of boards is black
01:23 PM Chillum: as long as I can see the traces clearly I am happy with any colour
01:57 PM julius: xentrac, you used a hantek6022be under linux?
01:57 PM xentrac: I have not yet done so, no
01:57 PM julius: ah ok
01:58 PM xentrac: apparently it's supported by sigrok but I would like to hear other people's experiences trying it before buying one myself
02:03 PM plouf: hi
02:04 PM plouf: is there any newset/modern HighVoltage Programmer than STK500 ?
02:05 PM NoHitWonder: did you brick your avr or why would you need one?
02:07 PM plouf: to have it
02:07 PM plouf: i have a STK500, for about 15 yeras
02:07 PM plouf: needed twice for unbrick
02:08 PM plouf: and now (15 years after) for a project that i need to deactivate reset
02:08 PM plouf: and because STK500 is old anyway , i was searching and found , actually NO official/full support HV programmer
02:08 PM plouf: STK and dragon only but...
02:10 PM plouf: i may order a dragon ... if no better solution exist
02:10 PM xentrac: to use the reset pin for regular I/O?
02:11 PM plouf: yes
02:18 PM malinus: heh I destroyed my old STK500
02:18 PM malinus: a wild resistor somehow got under the board
02:18 PM malinus: pumff
02:19 PM malinus: [magic smoke intesifies]
02:19 PM specing: malinus: sounds like an ideal opportunity to bail to cortex-m
02:20 PM malinus: specing: that's what I'm using, yeah.
02:20 PM xentrac: haha
02:24 PM malinus: specing: I'm only using coretx-r :)
02:25 PM specing: core transmitter receiver? :D
02:25 PM specing: good typo
02:25 PM malinus: haha
02:59 PM Jartza: lol
02:59 PM Jartza: http://www.plzi.com/koodia/lander.asp
03:00 PM xentrac: ?
03:18 PM NoHitWonder: this autodesk eagle seems pretty much the same as the old one
03:23 PM NoHitWonder: and for some reason it says: EAGLE Premium License for commerial use
03:24 PM NoHitWonder: although i didnt buy anything
03:30 PM NoHitWonder: apparrently students get free premium license
03:32 PM xentrac: of course
03:32 PM xentrac: otherwise all fresh graduates would be evangelizing kicad or something
03:33 PM xentrac: I mean some companies try to slap a "for non-commercial use" restriction on their student licenses, but they're generally full-featured
03:54 PM polprog: is 4n7 enough decoupling or 100n is better?
03:54 PM polprog: on an atmega8
03:54 PM polprog: with moderately good psu
03:56 PM xentrac: 4n7 means 4700pF? for a power supply bypass cap? I don't know much of anything but that sounds like a lot
03:56 PM polprog: too much? it's the smallest one
03:56 PM polprog: i have in smd storage
03:56 PM xentrac: no, it's not too much
03:57 PM xentrac: I mean the only possible downside of having big bypass caps afaik is that you have potentially a high inrush current at power-up time
03:57 PM polprog: the case is that i have no idea about calculating cmos bypass caps and i never ran into problems, i just sticked a whatever cap i had for prototype builds
03:57 PM xentrac: and thus you get sparks when you plug stuff in
03:57 PM specing: xentrac: add an inductor before it
03:57 PM polprog: im now designing the thermocouple pcb so i will probably ask stupid questions
03:57 PM xentrac: but I don't think you have to worry that it's too small. maybe someone who actually knows stuff will comment though
03:58 PM polprog: ill go with 4n7
03:58 PM specing: just go with what the datasheet says
03:58 PM specing: I think it actually specifies that
03:58 PM polprog: the funniest part is that i just found out about max31856 which is the beefy version of the 31855
03:58 PM polprog: but the one i have is ok
03:59 PM xentrac: specing: I think there may be some possible downsides to putting an inductor in series with your power rails :)
04:00 PM polprog: sparks at 100 nano :P it's not designed to be used in flammable gas environments
04:02 PM xentrac: really you shouldn't plug and unplug electrical circuits in flammable gas environments :)
04:02 PM polprog: ask the intern to do it
04:02 PM polprog: :^)
04:03 PM polprog: theres nothing in the datasheet on the bypass caps for general supply, only about AREF caps
04:04 PM specing: so look what arduino boards use
04:04 PM specing: copy&paste
04:04 PM xentrac: https://en.wikipedia.org/wiki/I%C2%B2C#Derivative_technologies says I²C isn't a registered trademark, the patents have lapsed, and that often "TWI" means an incomplete I²C interface implementation
04:04 PM specing: and learn mandarin along the way
04:04 PM polprog: oh, great idea :)
04:04 PM polprog: thanks
04:05 PM xentrac: Arduinos dont use ATMega8s
04:05 PM polprog: whatever
04:05 PM specing: but they use ATCloseEnoughs
04:05 PM xentrac: heh, yeah
04:06 PM polprog: 100n, i think their design consideration is people who know nothing about electronics
04:07 PM xentrac: tonight I am going to see if I can finish getting this voltmeter on the adjustable power supply working
04:07 PM xentrac: I scavenged an LCD display from a clock radio, which has this shitty 2×11 common-cathode pinout
04:07 PM xentrac: uh, LED
04:08 PM polprog: 2x11 common cathode? that's interesting
04:08 PM xentrac: yeah, I really expected 4×7
04:09 PM xentrac: the datasheet says max 25mA per LED, or 100mA at 10% duty cycle
04:09 PM xentrac: which is kind of a bummer because the most I can drive directly from an ATMega is 40mA per pin, which is 20mA per LED
04:10 PM polprog: how is 40mA per pin 20mA per led...? sinking from two at once?
04:10 PM xentrac: right, if it's an anode both of whose segments are lit
04:10 PM xentrac: and I don't really want to haul out a couple of ULN2805s to drive the 11 anodes
04:11 PM polprog: never heard of that chips, look cool thanks
04:11 PM polprog: use BC547 :^)
04:12 PM xentrac: this means I'm going to have, at best, 1.8mA average per LED. really more like 1.3mA with the current-limiting resistors I've put on the cathodes
04:12 PM polprog: barely lit
04:12 PM xentrac: yeah. I'd like the output not to be so dim
04:13 PM xentrac: BC547s would be the same thing as the ULN2805s, only eight times worse ;)
04:13 PM polprog: but hey, nostalgia
04:14 PM polprog: wow
04:14 PM xentrac: in the sense that I would need eight times as many of them
04:14 PM polprog: i ust found 3 TQFP mega16's stashed in a box
04:14 PM xentrac: (and also their hFE is smaler, and their total current is smaller, but those are details)
04:14 PM polprog: yeah, i was joking :D
04:15 PM xentrac: oh! sorry
04:15 PM xentrac: that sounds awesome! what are you going to do with them?
04:17 PM polprog: probably i wont have to buy anymore, cos i have 6 in tqfp and two in dip, they will be used for prototyping
04:17 PM polprog: im ordering some attiny85's to play with
04:18 PM specing: xentrac: buy a CREE XHP70 mounted on a copper star if you want bright
04:18 PM specing: $12 from china
04:19 PM specing: 4000 lumens
04:19 PM NoHitWonder: polprog http://www.atmel.com/images/atmel-2521-avr-hardware-design-considerations_applicationnote_avr042.pdf this uses 2 decoupling caps on the example design
04:19 PM NoHitWonder: maybe because of this http://forum.arduino.cc/index.php?topic=120796.0
04:20 PM NoHitWonder: 100n and 4u7
05:06 PM polprog: what is the technical name for theese 4 resistor combo elements?
05:07 PM polprog: i tried to search resistor pack but it gives me resistor kits
05:07 PM specing: bridge?
05:07 PM specing: here we go
05:07 PM chatter29: hey guys
05:07 PM chatter29: allah is doing
05:07 PM specing: again.
05:08 PM polprog: what was that lol
05:10 PM polprog: specing: thanks, thats it
05:14 PM specing: chatter is promoting islam all around freenode. It seems it is an actual human and not a bot
05:14 PM polprog: does this channel even have an op?
05:15 PM specing: I keep trying to mess with it, but it gets banned way too soon
05:15 PM specing: I tried countering it with the flying spaghetti monster in gnuradio channel
05:15 PM polprog: dont even waste your time
05:16 PM specing: this is practice
05:16 PM specing: so when the time comes for some real quality trolling, I'll be prepared
05:16 PM polprog: kek
05:17 PM polprog: anyways, i m halfway thru the board design
05:17 PM Tom_L: rev 1 of 3
05:17 PM polprog: probably the most relaxing process
05:17 PM Tom_L: it's like painting copper
05:18 PM polprog: actually i had a board with a second rev, but it was more like version two
05:18 PM polprog: upgraded
05:18 PM polprog: the new board is all smd so probably i could etch it
05:18 PM polprog: if i wanted to and had time
05:18 PM Tom_L: i got tired of the mess
05:18 PM Tom_L: and lazy
05:19 PM polprog: im afraid the mess too
05:19 PM Tom_L: persulphate isn't quite as bad as fecl3
05:19 PM polprog: ive never done it before but i think i knoe the theory
05:19 PM xentrac: ever tried copper chloride?
05:20 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/etching/etch_index.php
05:20 PM Tom_L: xentrac no but i hear it works
05:20 PM polprog: i think i will use sodium sulphate, thats what is used here in pl
05:21 PM polprog: this reminds me i need to get a domain
05:21 PM Tom_L: i forget which one, ammonium persulphate or sodium persulphate will remove etch marker ink so be careful. i did mine with laser toner transfer
05:21 PM polprog: ill go the toner method if any
05:21 PM Tom_L: you don't need a domain
05:21 PM Tom_L: my server is sitting right here
05:21 PM polprog: but i want
05:22 PM Tom_L: and the dynamic ip stuff is free
05:22 PM polprog: ill set myself up a vps
05:22 PM Tom_L: i don't care if they gripe about the way the link looks
05:23 PM Tom_L: my isp blocks 80 so i gotta use another port
05:23 PM polprog: ive used no-ip.org when i was setting up minecraft servers
05:23 PM Tom_L: i had dyndns but they got flakey
05:24 PM xentrac: I used dyn.ml.org
05:24 PM xentrac: then Aveek Datta came back and shut down Monolith
05:24 PM polprog: maybe... i could setup a site on a spare raspi
05:24 PM xentrac: free dynamic DNS domains have a long history and aren't hard to set up but have a tendency to disappear
05:24 PM Tom_L: if mine goes down, i know where the problem is :D
05:25 PM xentrac: I think the shutdown happened in 1998 or 1999
05:25 PM * Tom_L goes back to setting up his virtual box
05:25 PM polprog: the only thing that is bad in that kind of setup is that you need to take the fact that theres an open webserver on your lan and thats a security issue
05:26 PM polprog: open to tge whole internet
05:26 PM polprog: anyways
05:26 PM polprog: its late here, and i have school tomorrow
05:26 PM polprog: see you
05:26 PM specing: < Tom_L> my isp blocks 80 so i gotta use another port
05:26 PM Tom_L: it's been up for quite a few years with no issues so far
05:27 PM specing: Tom_L: use 443
05:27 PM Tom_L: specing, why?
05:27 PM specing: TLS
05:27 PM specing: 443 is https fyi
05:27 PM Tom_L: hmm
05:27 PM Tom_L: i always just used 81 82 or such
05:27 PM xentrac: I noticed recently that my ISP gives me a globally routable IPv6 address
05:27 PM specing: Tom_L: I mean
05:28 PM specing: Tom_L: switch to port 443 and enable TLS on your website with a proper certificate
05:28 PM Tom_L: i'd have to figure all that out..
05:28 PM Tom_L: i'm not super linux savvy
05:28 PM specing: It takes an hour
05:29 PM Tom_L: that's 3-4 for me
05:29 PM specing: half of which is reading the letsencrypt service agreement
05:29 PM Tom_L: you read those?
05:29 PM specing: Yes
05:30 PM Tom_L: anyway, back to this project..
05:30 PM specing: No, TLS :D
05:30 PM Tom_L: i have an old dos box i gotta maintain
05:30 PM Tom_L: trying to get it working on new hardware
05:31 PM Tom_L: all that's left is fixing a pci parallel port remap issue
05:31 PM Tom_L: it's a mess trying to combine legacy stuff with current
05:32 PM Tom_L: someone suggested a virtual box as an alternative so until some hardware gets here i'm messin with it
05:36 PM xentrac: virtualbox is often a good bet, but I've often used dosbox (and previously dosemu) with success for that
05:36 PM xentrac: none of those will work for real-time control though
05:36 PM Tom_L: that's kinda what i figured
05:36 PM Tom_L: i'm just messin with it. never used virtual box
05:37 PM specing: Tom_L: ?_?
05:37 PM Tom_L: surprised?
05:37 PM specing: why do you depend on an old dos box?
05:37 PM Tom_L: it's an old business app that would take too much to replace
05:38 PM Tom_L: i could rewrite it in windows but kinda shudder to think about that
05:38 PM xentrac: because some days you just gotta live on the edge, specing
05:39 PM specing: Tom_L: so run it in dosemu?
05:39 PM specing: yes there is a purpose built emulator for dos...
05:39 PM Tom_L: still won't fix the hardware remap issues will it?
05:39 PM Tom_L: it seems to be a quite common problem
05:40 PM specing: How are you supposed to have hardware remap issues if you have no hardware?
05:40 PM Tom_L: mid age motherboards let you still map stuff, this bios has nothing on peripherals
05:40 PM Tom_L: i have a pci card but remapping it doesn't work
05:40 PM Tom_L: i've got all but one parport working
05:41 PM Tom_L: btw, dos blisters on a ssd
05:41 PM specing: ssd in your case is a 128 MB usb key?
05:42 PM Tom_L: no
05:42 PM Tom_L: 500gb samsung
05:42 PM specing: 0.o
05:42 PM specing: so 499 GB free? :D
05:42 PM Tom_L: it dual boots
05:42 PM Tom_L: you didn't think i'd just let that sit and rot did you?
05:42 PM xentrac: hah, that's awesome
05:43 PM Tom_L: i've got a TB and 500gb ssd on my -> other pc
05:43 PM Tom_L: err 2 500 and a tb
05:44 PM Tom_L: too bad they're not pcie
05:45 PM Tom_L: this'll be a decent setup for a dos box if i can get it all to work
05:45 PM arij is now known as arijwork
05:46 PM Tom_L: could use an itx i've got but this is a mini atx
05:46 PM Tom_L: same MB just different form
05:46 PM Tom_L: couple less slots
06:20 PM Emil: mmmm
06:21 PM Emil: This Raspi + 1 wire is too god damn easy
06:28 PM Lambda_Aurigae: Jartza, got the OctaPentaVeega!
06:28 PM Lambda_Aurigae: it's sooooo cute!
06:59 PM Tordek_: I'm having some trouble getting this to work https://www.instructables.com/id/GPSDO-YT-Disciplined-Oscillator-10Mhz-Reference-Fr/
06:59 PM Tordek_: the oscillator issue is fixed, and I got my ICs back with the bootloaderusing that
07:00 PM Tordek_: so it would seems to work... but... when I try to use this, the pwm remains at 0.5 and nothing is displayed on my lcd
07:00 PM Lambda_Aurigae: ask the guy who wrote the instructable.
07:00 PM Tordek_: right, but one more thing
07:01 PM Lambda_Aurigae: being as I see no code there,,,,,no flippin clue.
07:01 PM Tordek_: I set the CLKOUT flag,so I should be getting a clock on xtal2, right?
07:01 PM Tordek_: and that's not happening
07:02 PM Tordek_: any ideas a la "your reset is stuck on" or something?
07:02 PM Lambda_Aurigae: no clue without testing the circuit.
07:03 PM Tordek_ is now known as Tordek
07:07 PM Tom_L: don't happen to have a msclient floppy .iso file do you?
07:07 PM Tom_L: disk1 & 2
07:07 PM Lambda_Aurigae: not handy.
07:08 PM Lambda_Aurigae: might have the disks somewhere but they are probably moldy.
07:08 PM Tom_L: i got the NT server iso but need an image to get them in the stupid v box
07:10 PM Tordek: ~
07:10 PM Tom_L: does ms ftp server even work anymore?
07:11 PM Lambda_Aurigae: no clue
07:11 PM Lambda_Aurigae: I don't do mickysoft unless absolutely necessary for my job.
07:13 PM Tom_L: i may have found them
07:17 PM Tom_L: wonder how i can make a img file now
08:28 PM arijwork is now known as arij
10:51 PM Thrashbarg_ is now known as Thrashbarg
11:27 PM Tordek: yay I may have set the "disable reset" flag on my atmega
11:30 PM Casper: congratulation! you now need to make an HV programmer!
11:30 PM Casper: new project!