#avr Logs
Mar 01 2017
#avr Calendar
12:13 AM daey_ is now known as daey
01:33 AM Thrashbarg_ is now known as Thrashbarg
02:17 AM Tordek: I'm having the weirdest problem... I have a circuit that works unless the LCD display is plugged in; at that point it.. hangs? but the pwm keeps outputting its 50% duty cycle wave
03:38 AM avrdude: >tfw my nickname gets highlighted a lot in this channel, but its never anyone actually talking to me
03:41 AM _ami_: avrdude: :)
03:42 AM _ami_: are you really THE avrdude!? :)
03:44 AM Haohmaru: avrdude does that surprise you?
03:44 AM avrdude: not really :(
03:53 AM Haohmaru: no one talks to you, they only try to command you to do weird stuff
03:53 AM Haohmaru: they never say "please" nor "thank you"
04:03 AM Lambda_Aurigae: well, avrdude is the name of the program many of us use to upload programs to avr...AVR Downloader Uploader
04:50 AM JanC is now known as Guest86538
04:50 AM JanC_ is now known as JanC
06:01 AM hetii: Hello :)
06:10 AM skz81: <avrdude> >tfw my nickname gets highlighted a lot in this channel, but its never anyone actually talking to me
06:10 AM skz81 is now known as avr-gcc
06:10 AM avr-gcc: I guess I will also, now
06:11 AM avr-gcc is now known as atmega
06:11 AM atmega: or even this one
06:11 AM atmega is now known as skz81
06:11 AM skz81: ^^
06:19 AM Emil: the heck is the point of doing that?
06:19 AM Emil: Having nicks that you dont certainly have anything to do with
06:20 AM Emil: ohwell, it is not my loss :D
07:44 AM skz81: <polprog> if they go this fast there will be more pi versions than atmel micors >> If you count banana pis, orange pis and so on with them, yup, probably !
07:53 AM Emil: skz81: ?
07:54 AM skz81: Emil, !
07:54 AM skz81: about the number of rasp.pi models
07:55 AM skz81: i'm a bit backscrolled, though
08:04 AM Lambda_Aurigae: I'm liking the look of the new rPI0M
08:04 AM Lambda_Aurigae: should make a great hacking tool.
09:04 AM enhering: morning
09:05 AM Ad0: hi
09:29 AM enhering[irssi] is now known as enhering
09:29 AM bss36504: Lambda_Aurigae: you mean the W? I was just talking about that yesterday...looks cool
09:30 AM Emil: Imho CHIP is still much better value
09:30 AM Emil: Though you do get a better supported OS with raspi
09:31 AM bss36504: does CHIP have wifi and BLe built in?
09:31 AM Emil: Yeah
09:32 AM Emil: and built in memory
09:32 AM Emil: and built in lipo charging and usage
09:32 AM Emil: The fact is that not having to deal with micro sd cards is a god send
09:32 AM Emil: gift
09:38 AM bss36504: Looks pretty cool
09:44 AM Ad0: how much power in the CHIP?
10:03 AM enhering: anybody here uses irssi?
10:08 AM LeoNerd: Probably many do; it's a popular client. If you had a question, best just to ask it
10:08 AM LeoNerd: (Though is #avr *really* the most appropriate place?)
10:09 AM enhering: is there a way for it to notify me when someone joins or quits?
10:10 AM LeoNerd: Do you mean join a particular channel, or just turns up on the server as a whole?
10:10 AM LeoNerd: In the /specific/ case of freenode, they have disabled the WATCH IRC extension
10:10 AM LeoNerd: So what you can do on most networks doesn't actually work on freenode
10:11 AM enhering: i mean on a server
10:11 AM enhering: ok
10:12 AM enhering: thanks, LeoNerd
10:13 AM LeoNerd: It's mostly why I've not really got around to implementing WATCH in my IRC client yet.. kindof little point when the one big network I actually use doesn't support it
10:16 AM enhering: makes sense
11:38 AM Ad0: enhering: I do
11:52 AM antto: enhering /ison <nick>[,nick2,nick3,...]
11:53 AM antto: should work in most networks
12:13 PM Emil: If join is enough for you
12:13 PM Emil: Then its visible by default
12:13 PM Emil: and you can always poll whois
12:17 PM Emil: https://emil.fi/jako/linkit/onoff do you guys see a problem with this?
12:19 PM Ad0: hey antto :)
12:19 PM antto: Emil poll /ison
12:19 PM antto: Ad0 wub
12:20 PM Emil: antto: if ison works then yeah, obviously one should use that
12:20 PM antto: that's what mIRC uses for its notify
12:21 PM Emil: Anyone streaming video with a raspberry pi?
12:23 PM cehteh: for what definition of streaming? .. kodi acts as media server, i tried that once and worked
12:25 PM Emil: cehteh: streaming from the camera module to a website
12:25 PM cehteh: ok never tried
12:27 PM Ad0: hehe
12:27 PM Ad0: antto: you probably know a thing or two about ADCs
12:27 PM Ad0: is there one with buffering and buffering frequency?
12:27 PM Ad0: that's less expensive than an AVR
12:27 PM Ad0: 1024 uint16_t
12:28 PM antto: buffering?!
12:28 PM antto: like an audio buffer?
12:28 PM Emil: Ad0: eh?
12:30 PM antto: Ad0 they don't seem to work like that, you trigger them, they sample.. then you read the value.. and you can make your code do that periodically at an interrupt or so
12:30 PM antto: for something that works more like audio you should be looking more at these so called "audio codecs"
12:30 PM antto: they typically need a clock and they do the ugly stuff
12:34 PM Ad0: antto: yep
12:34 PM Ad0: it's not audio
12:35 PM antto: what then?
12:35 PM Ad0: it's just sampling some ADC input - 1024 samples across 200 ms, not too intense. the mcp3008 was too annoying to work with. 1024 spi requests per channel through linux on PI used 30-40% of one core
12:36 PM Ad0: I just want to avoid needing to flash the chip so I either wanted a nice ADC
12:36 PM Ad0: or some chip that could deliver data that has been buffered at a given rate
12:36 PM Ad0: because chips to it more realtime and aren't affected by thread scheduling etc by the OS.
12:36 PM Ad0: do it*
12:37 PM antto: hm
12:37 PM antto: so what's the sampling rate and what bitdepth?
12:38 PM antto: 5120Hz?
12:38 PM Ad0: yeah
12:39 PM Ad0: or something that just has a 3k buffer at all times
12:39 PM Ad0: ring buffer
12:39 PM Ad0: hehe
12:39 PM Ad0: but no MCU has that big of a memory I guess
12:39 PM antto: you could use an 8bit mcu with ADC and USART maybe?
12:39 PM skz81: <Emil> Anyone streaming video with a raspberry pi? >> have coworked that did
12:39 PM Ad0: yeah antto
12:39 PM Ad0: I do that today and I even make it calculate the stuff I need
12:40 PM antto: there's atmegas with like 8kB, 16kB, even 32kB memory
12:40 PM Ad0: but it's not really fast enough
12:40 PM antto: are you sure the cpu consuption isn't from your app?
12:40 PM antto: like.. do you block while waiting?
12:40 PM Ad0: no
12:40 PM antto: i mean, busy wait
12:40 PM Ad0: the SPI you mean?
12:40 PM antto: weird then
12:41 PM antto: yes, where you said you got high cpu usage
12:41 PM antto: how do you use SPI from that thing even ;P~
12:42 PM antto: maybe whoever wrote the spi functions was lazy.. or is it bitbanged spi?
12:42 PM Ad0: it's hardware SPI
12:42 PM Ad0: I studied the driver
12:43 PM Ad0: seems legit
12:43 PM Ad0: it doesn't use DMA before the transfer size >= 96 bytes
12:43 PM Ad0: doesn't use IRQ before the time spent is > x
12:43 PM antto: what happens if you run the program, but the actual spi device (the thing with the ADCs) is not there?
12:43 PM Ad0: you mean like disconnect it ?
12:43 PM antto: yiz
12:43 PM antto: but don't tell anyone ;P~
12:44 PM Ad0: hm not sure gonna try that tomorrow haha
12:44 PM Ad0: not sitting at the office
12:44 PM Ad0: but afaik there's no way around it
12:44 PM Ad0: either switch to a dac where you can just request continous
12:44 PM Ad0: or program an AVR to do it
12:44 PM Ad0: leaning against the latter because I can actually control it to deliver what I want
12:45 PM Ad0: which AVR could support that throughput?
12:45 PM Ad0: 5k-6kHz
12:45 PM Ad0: or should it be USART or SPI
12:45 PM antto: i guess you prefer to get the data packed tightly (more samples together, less headers/footers) and at bursts?
12:45 PM Ad0: my idea was this:
12:46 PM Ad0: request what channel to sweep continous sample
12:46 PM Ad0: in one byte query
12:46 PM Ad0: channel 0-7 in bits
12:46 PM Ad0: turn bits on for channel enable
12:46 PM Ad0: and then the avr just reads the ADC x that is set to 1 in that byte
12:46 PM Ad0: and do them sequentially
12:46 PM Ad0: until you tell it to do something else
12:46 PM Ad0: that is you tx 0x0 after the request to do clocks.
12:47 PM Ad0: forever and ever :)
12:47 PM antto: hm
12:47 PM Ad0: the SPI master decides the clock speed
12:47 PM Ad0: so you can sample at that SPI speed
12:47 PM Ad0: that regulates the speed at which you sample
12:47 PM Ad0: has to work right? :)
12:47 PM antto: no idea
12:48 PM Ad0: do X clocks till the ADCs are set up
12:48 PM Ad0: so enable channel 0 is sending 0x1 0x0 0x0 0x0
12:48 PM antto: you want the master to drive the sampling speed by the communication speed or by sending periodic clock packets?
12:48 PM Ad0: 3 other bytes are just to let the AVR get it's shit together
12:49 PM Ad0: the master to drive the sampling speed
12:49 PM antto: that makes things complicated
12:49 PM Ad0: I thought it made it simpelr :)
12:49 PM antto: well, i don't know what happens with the data once it reaches teh comput0r
12:50 PM Ad0: the big question is it fast enough to get the ADC value during the clock cycle to return the byte
12:50 PM Ad0: or maybe I should have clock gaps between each sample.
12:50 PM Ad0: I think it would work
12:51 PM Emil: skz81: how did he do it?
12:51 PM Emil: Ad0: wtf
12:52 PM Emil: Ad0: the atmega328p can stream 48000 8 bit samples per second through uart just fine
12:52 PM Ad0: nice
12:52 PM Ad0: without a crystal?
12:52 PM Ad0: I want to avoid it
12:52 PM Emil: Ad0: the specification says that you get the 10Ks samples at the 10 bits just fine
12:52 PM Emil: Ad0: do you even datasheet, mate?
12:52 PM Ad0: I do
12:52 PM Emil: No, apparently you don't
12:52 PM Ad0: but it's a lot easier to ask here :)
12:53 PM Emil: 2017-03-01 20:10:42 +0200 < Ad0> but it's not really fast enough
12:53 PM Ad0: if I were to say "read the datasheet" for every question it would get fucking boring here don't you think
12:53 PM Ad0: Emil: it does calculations as well
12:53 PM Ad0: let me show you.
12:53 PM Ad0: also it was an attiny45
12:53 PM Ad0: at 8 mhz
12:53 PM Ad0: soft uart
12:54 PM Ad0: http://lechacal.com/RPICT/3CT/RPICT3_v1_4.ino
12:54 PM Emil: a) wtf are you using soft uart when you have USI b) pls
12:54 PM Ad0: I am getting into USI now
12:54 PM Emil: And wtf are you using arduion
12:54 PM Emil: arduino*
12:54 PM Ad0: I'm not
12:54 PM Ad0: I ported it to arduino
12:54 PM Ad0: FROM*
12:55 PM Ad0: anyway I want to do the calculations after I have the samples from the avr so I don't have to flash it every time I want to change the logics there
12:55 PM Emil: If that was your code no wonder you couldn't do shit with it
12:56 PM Ad0: yeah wonderful code isn't it
12:56 PM Ad0: lots of double etc lol
12:56 PM Ad0: just what the AVR likes
12:57 PM Emil: Why the heck would you do that calculation on the avr and not on the raspi you are streaming to?
12:58 PM Ad0: well...
12:58 PM Ad0: that's what I am trying to do
12:58 PM Ad0: but how to get all those samples over in a timely fashion
12:58 PM Ad0: is an attiny84 with USI enough?
12:59 PM Ad0: in case I have 4 chans I want to send over
12:59 PM Emil: depends entirely on the sample rate
12:59 PM Ad0: sample rate is 5-6kHz
12:59 PM Ad0: 10 bit
12:59 PM Emil: And if you sample the IRMS of if you need to continuously sample the whole thing
12:59 PM Ad0: exactly
01:00 PM Ad0: so who can garantuee even spacing between the samples?
01:00 PM Emil: wat
01:00 PM Emil: Like everyone has said: timer interrupt
01:00 PM Ad0: yeah that's on the AVR
01:00 PM Emil: ...
01:00 PM Emil: How silly are you?
01:01 PM Ad0: ehm
01:01 PM Emil: You push the samples to the tiny as you please
01:01 PM Emil: to the raspi*
01:01 PM Emil: with even intervals
01:01 PM Ad0: I don't want your help
01:01 PM Ad0: you can stop now
01:01 PM Emil: The raspi gets bytes out at even intervals
01:01 PM Ad0: I would rather have help from someone a bit more polite
01:01 PM Emil: Ad0: no wonder you can't get things done if you can't even listen
01:02 PM Emil: Go to #arduino if you can't handle being told how it is
01:03 PM Emil: Was already coding an example that would have done what you are after but since you are not interested eh
01:04 PM Ad0: hehe
01:04 PM Emil: Ad0: this is not a customer support channel
01:04 PM Emil: you either stop whining or move somewhere else
01:04 PM Ad0: I know but be nice.
01:04 PM Ad0: it pays off in the long run
01:05 PM Ad0: thanks for your help so far
01:05 PM Ad0: we agree on that the .ino file is ... not optimal :P
01:05 PM Emil: Ad0: I have no interested in sucking anyones dick. And the words I used to insult you were "how silly are you"
01:06 PM Emil: If that's too much for you, then too bad
01:06 PM Ad0: yeah
01:06 PM Ad0: there's something between sucking someone's dick and being one
01:06 PM Ad0: :D
01:06 PM Emil: I can also call you a fucking idiot if you prefer
01:06 PM Emil: If "how silly are you" is being a dick nowadays then huehuehue
01:07 PM Emil: But like I was saying
01:07 PM Emil: Don't recalculate the whole sum of the values everytime you get a new sample
01:07 PM Emil: Instead keep in memory the last n samples and their sum
01:08 PM Emil: That way you don't have to do the for i number_of_samples
01:08 PM Ad0: yeah
01:08 PM Ad0: I just realised that if you use a timer like you said
01:08 PM Ad0: you know that the samples coming over are at a timed rate
01:08 PM Ad0: even if the UART is not perfect timing
01:08 PM Emil: That's what I said ;)
01:08 PM Ad0: yes
01:08 PM Ad0: I just had to process it
01:09 PM Ad0: because in my head I was thinking oh no they are coming over UART in a non-timely manner
01:09 PM Ad0: but that doesn't matter....
01:09 PM Emil: Here's a point that pays off in the long run
01:09 PM Ad0: silly me, yeah you're right
01:09 PM Emil: Don't go 2017-03-01 20:32:05 +0200 < Ad0> I don't want your help
01:09 PM Emil: 2017-03-01 20:32:08 +0200 < Ad0> you can stop now
01:09 PM Emil: If you want someones help
01:09 PM Emil: Even if you are super god damn right, there's a chance you arent
01:10 PM Emil: Everyone should swallow their pride when asking for help
01:10 PM Emil: But here's a thing, too
01:11 PM Emil: You must implement some sort of framing if you are sending binary data
01:11 PM Ad0: yes
01:11 PM Emil: One way to do it is to send for example two \n\n or some other reserved character every n measurements
01:11 PM Emil: Or have some other framing
01:11 PM Emil: Losing sync is rarely an issue
01:12 PM Emil: (that is when you send binary data and you for some reason lose an entire byte)
01:12 PM Ad0: this is a lot better than hammering an MCP3008 with SPI
01:12 PM Emil: SPI would be a lot better, though
01:12 PM Ad0: how would that be regards to timing?
01:13 PM Emil: You said you wanted to send ~4 channels, each at 5kS/s?
01:13 PM Ad0: yes
01:13 PM Emil: Well, SPI much faster and way more effective
01:13 PM Emil: And timing is not an issue there, either
01:14 PM Ad0: but to get it at that rate I had to clock it way down or is there some other way?
01:14 PM Ad0: the clue here is to not send many small transactions
01:15 PM Ad0: but have a big one
01:15 PM Ad0: small transactions burned CPU
01:15 PM Ad0: because it was busywaiting / polling on the PI
01:15 PM Emil: If you use uart then you need a baudrate of at least 440000 if you send 20k 16-bit samples per second and assume 10% overhead
01:16 PM Emil: With SPI you can send at F_CPU/4 (or even /2 if you don't care about receiving)
01:16 PM Emil: So if you want to use uart then choose 500kBaud as the speed
01:17 PM Ad0: yeah
01:17 PM Emil: and calibrate your clock for precise 8MHz
01:17 PM Ad0: is there an automated way to calibrate clock from linux for example?
01:18 PM Ad0: I saw the atmel studio can suggest a clock tune
01:18 PM Ad0: like run a self tuning on startup perhaps
01:18 PM Emil: Most AVRs are already tuned to 8MHz but that's not a very good tuning iirc
01:18 PM Emil: You only really need to tune it once
01:18 PM Ad0: yeah
01:18 PM Emil: If your temperature doesn't change that much
01:18 PM Ad0: I've done it with OSCCAL
01:18 PM Emil: Correct
01:19 PM Ad0: with SPI I wouldn't need framing would I ?
01:19 PM Emil: Yes I would still implement framing there
01:19 PM Emil: but it can be a lot more lightweight
01:19 PM Ad0: maybe 0x00 gaps
01:20 PM Ad0: the main thingis that I want to offload the PI and using SPI with 3 byte transfers used a lot of CPU
01:20 PM Ad0: if it's > 96 byte it goes into dma mode
01:20 PM Emil: Also
01:21 PM Emil: If you need to sample at 20k
01:21 PM Emil: you might as well only use 8 bits
01:21 PM Ad0: also - I don't require all channels to be sampled exactly at once
01:21 PM Emil: 10 bits ADC resolution is not guaranteed anymore at that samplerate iirc
01:22 PM Ad0: they are treated separately
01:25 PM Ad0: 40960 bytes a second
01:26 PM Ad0: I am just scared if I start doing this with SPI I might run into some issues on linux with CPU usage etc
01:28 PM Lambda_Aurigae: bss36504, yeah..the W, not the M...dunno where I pulled the M out of.
01:30 PM Emil: Ad0: no, you wont
01:30 PM bss36504: Lambda_Aurigae: Upside down W?
01:30 PM bss36504: Either way, yeah, it looks cool
01:31 PM bss36504: I had an edison that cost me 7x as much and it inexplicably stopped working one day. Very disappointing. The price of entry on the new rpi is much more enticing than buying another edison.
01:31 PM Ad0: Emil: I talked to people invovled with the SPI driver
01:31 PM Lambda_Aurigae: edison was waaaay pricy....and intel.
01:31 PM Ad0: you have to choose the right size of tx/rx for it to not drain CPU
01:32 PM Lambda_Aurigae: the rRI0w at 10 dollars is in that impulse buy range where I might buy 4 or 6 of them for the fun of it.
01:32 PM bss36504: haha, good luck finding even one to buy.
01:32 PM Ad0: I manage to sample 3 x 1024 x 3 bytes using ~35% of one core
01:32 PM bss36504: but yes, once they are in stock again I'll be getting some
01:32 PM Emil: bss36504: DO NOT. TOUCH. THE EDISON
01:32 PM Emil: ABORT ABORT
01:33 PM Emil: Lambda_Aurigae: buy chip instead ;)
01:33 PM bss36504: I was going to roll my edison into my reflow oven just for giggles, but then it crushed my dreams by not booting anymore for no particular reason
01:33 PM Emil: Ad0: don't do 3 byte reads
01:33 PM bss36504: Cant even get it to enumerate to re-flash it
01:33 PM Emil: do 256 byt reads
01:33 PM bss36504: thing's fucking FUBARd
01:33 PM Ad0: Emil: yeah this was on the MCP3008
01:33 PM bss36504: Emil: Yeah gonna cut my losses and move on to another SBC.
01:34 PM bss36504: I dont even care about SBCs that much, I just wanted to tinker.
01:34 PM bss36504: The edison worked great out of the box. Got it ssh-ing in about 30-60 minutes.
01:34 PM _ami_: https://www.aliexpress.com/item/WS2812-5050-RGB-LED-Module-for-Arduino/32310855531.html -> Is 5v good supply (1.5A) good enough to drive this eight WS2812 LEDs strip ?
01:34 PM Emil: bss36504: Pine64 has treated me okay (though I am particularly biased because I absolutely adore and love the 1Gbps etho <3333)
01:34 PM _ami_: https://pbs.twimg.com/media/C52pVxAXQAAm8C9.jpg:large -> only 6 out of 8 leds are working fine.
01:34 PM bss36504: Yeah I want wifi
01:34 PM _ami_: seems like other 2 leds are dead.
01:34 PM bss36504: Minimalist design, wifi, enough I/O to do something useful
01:34 PM bss36504: Thats what I want
01:34 PM Emil: _ami_: never ever fucking trust ebay/chink shit powers
01:35 PM Emil: They are _ALWAYS_ underperforming
01:35 PM _ami_: this A+A+A+ aliexpress seller is really is bad :/
01:35 PM _ami_: Emil: yeah :(
01:35 PM Emil: bss36504: that's CHIP for you
01:35 PM _ami_: bad experience this time
01:35 PM Lambda_Aurigae: Emil, can't buy the chip, make a board, and populate it for 10 dollars myself.
01:35 PM Emil: Lambda_Aurigae: eh?
01:35 PM Emil: Lambda_Aurigae: how can you do that with a raspi, then?
01:36 PM Lambda_Aurigae: 10 dollars for a rPI0 with 512MB of ram, wifi, bluetooth, usb, hdmi, composite video out
01:36 PM Emil: Lambda_Aurigae: you were talking about the Pi0W afterall, right?
01:36 PM Lambda_Aurigae: yes.
01:36 PM Lambda_Aurigae: you said buy chip instead.
01:36 PM Emil: Lambda_Aurigae: CHIP
01:36 PM Emil: C.H.I.P.
01:36 PM Lambda_Aurigae: oh.
01:37 PM Emil: nexthingco
01:37 PM Emil: :D
01:37 PM Emil: I agree: they named it horribly!
01:37 PM bss36504: Chip looks pretty cool. The Chip Pro seems like my kinda thing
01:37 PM Emil: Googling it was an absolute pain
01:37 PM Emil: In the beginning
01:37 PM bss36504: Might just get one of them first
01:37 PM Emil: bss36504: the pro chip doesn't seem that interesting, really
01:37 PM Emil: it's more expensive and has less features
01:38 PM bss36504: Yeah but castellated pads <3
01:38 PM bss36504: I hate .1" pin headers with a passion
01:38 PM Emil: Well, sure :D
01:38 PM Emil: That's true
01:38 PM bss36504: One of the selling points of the edison for me was that sexy low profile mezzanine connector with 70 IO
01:38 PM Emil: That connector was absolutely horrible
01:39 PM Emil: Want to detach? Lol imma break on you
01:39 PM Emil: They were designed for like a maximum of 4 attach detach cycles
01:39 PM Emil: And that was if you _knew_ how to take it off
01:39 PM Emil: properly
01:39 PM bss36504: yeah but it's soo small
01:39 PM bss36504: Maybe thats whats broken...
01:40 PM Lambda_Aurigae: chip pro is 6 dollars?
01:40 PM Emil: Lambda_Aurigae: 16
01:40 PM Emil: Lambda_Aurigae: CHIP is 9
01:40 PM Emil: Oh they have a blooper on the site :D
01:40 PM Lambda_Aurigae: oh...seems you can buy the processor yourself for 6 dollars.
01:40 PM Emil: Ohwait
01:40 PM Emil: no
01:40 PM Emil: The actual chip is 6 dollars, yeah
01:42 PM Lambda_Aurigae: so, does the pro have 256MB or 512MB ram?
01:43 PM Emil: You can select
01:44 PM Lambda_Aurigae: umm.
01:44 PM Lambda_Aurigae: ok.
01:44 PM Lambda_Aurigae: it says the processor has 256 but the board has 512
01:44 PM Emil: But the pro is not available
01:44 PM Emil: Yet
01:44 PM Lambda_Aurigae: oh.
01:44 PM Lambda_Aurigae: nevermind then.
01:44 PM Lambda_Aurigae: it's 512MB flash.
01:44 PM Lambda_Aurigae: and 256MB ram.
01:44 PM Lambda_Aurigae: that's kinda low.
01:45 PM Emil: Well, sure
01:45 PM Emil: But it runs linux and shell just fine for most applications
01:45 PM Emil: And the fact that I don't have to deal with usd cards is a win
01:49 PM Lambda_Aurigae: interfacing to monitor is a bit more difficult though.
01:49 PM Lambda_Aurigae: no hdmi port on there...the signals are there I guess but no socket to plug into.
01:51 PM Ad0: https://github.com/acropup/ATtiny-OSCCAL
02:18 PM _ami_: Lambda_Aurigae: do check review of omega2 too. not bad if you plan to do light stuffs: https://www.youtube.com/watch?v=uI_8G2t2ZLg
02:26 PM julius: hello guys
02:27 PM julius: this mosfet driver is killing me, the input is 25% duty cycle at 25khz (probably, have to check about the khz) for 5 seconds and than about 10% duty cycle...but the output from the mosfet driver is always 5v. i checked with a scope
02:27 PM julius: shoildnt the mosfet go to 0v output when the pwm signal is 0v?
02:30 PM Emil: Lambda_Aurigae: whatcha need a monitor for?
02:30 PM Emil: But sure
02:30 PM Emil: if you really want a monitor, then chip might not be for you
02:32 PM Emil: Omega2 has a teeeeensy amount of storage and ram :D
02:34 PM bss36504: Emil: If i get a chip, how hard is it to set up if you have no screen? Can you get a terminal via USB?
02:34 PM bss36504: Chip*
02:35 PM bss36504: C.H.I.P** :)
02:39 PM julius: do i need a 1mohm or something like that between mosfet input and gnd?
02:42 PM bss36504: Emil: I also really like that they broke out the USB connections to pads on the C.H.I.P. Very cool
03:07 PM Emil: bss36504: you get a serial terminal through the same usb :)
03:08 PM Emil: julius: wtf?
03:08 PM Emil: julius: do you mean gate?
03:08 PM Emil: And by 1mohm do you mean 1Mohm?
03:09 PM Emil: If yes to both: yes, you should have a pulldown resistor on the gate if your mosfet is a n channel one
03:13 PM bss36504: Emil: I figured it out after I read the docs :P Slow day at work so I had time. I just ordered a Chip Pro Dev kit. Seems just like what i need.
03:17 PM Emil_ is now known as Guest39186
03:19 PM bss36504: Emil: Ok here's something that wasnt immediately obvious to me, what sort of APIs do they have? Is it all shell based stuff? Or do they have "real language" APIs?
03:20 PM Ad0: probably a small python library that will use 99% cpu to read 10 bytes of UART
03:21 PM antto: that many?! u need more layers of virtual machines
03:21 PM antto: u need to run the python lib on an arm-linux running on an arm-emulator on the attiny
03:22 PM Ad0: lol
03:23 PM Ad0: I tried pi4j to read SPI
03:23 PM Ad0: lol...
03:23 PM Guest39186 is now known as Emil
03:24 PM Emil: bss36504: hm?
03:24 PM bss36504: What?
03:24 PM Emil: bss36504: you asked me a question
03:24 PM Emil: I didn't understand it
03:24 PM Emil: I also seem to have disconnected for some reason (fucking freenode)
03:24 PM Emil: so the last message I saw was at 2017-03-01 22:39:47 +0200 < Emil> If yes to both: yes, you should have a pulldown resistor on the gate if your mosfet is a n channel one
03:24 PM Emil: and then after that your question was the next I saw
03:24 PM Emil: 2017-03-01 22:49:54 +0200 < bss36504> Emil: Ok here's something that wasnt immediately obvious to me,
03:25 PM Emil: What happened in between?
03:25 PM bss36504: huh, odd that you only got half of that since I sent it as one message...
03:25 PM Emil: bss36504: No I got the whole thing
03:25 PM Emil: 2017-03-01 22:49:54 +0200 < bss36504> Emil: Ok here's something that wasnt immediately obvious to me, what sort of APIs do they have? Is it all shell based stuff? Or do they have "real language" APIs?
03:25 PM bss36504: oh
03:25 PM bss36504: sorry gotcha
03:25 PM Emil: bss36504: but what were you referencing with that?
03:26 PM bss36504: Ok so on the edison and I assume the rpi and in the CHIP examples, they show that you can say turn on a GPIO by writing to a particular location in the file system. That's all well and good if you want your stuff to be controlled by shell scripts. What if I want to write a C program? Are there already APIs for "real languages" ie not shell script?
03:26 PM Emil: Oh yeah sure
03:27 PM bss36504: Ok, I figured.
03:27 PM Emil: You don't have to use the slow filesystem api (which is convenient, though)
03:27 PM bss36504: Yeah it's great to mess around for sure
03:27 PM bss36504: but If I want to actually do something meaningful and/or fast, I'll need something better than that haha
03:27 PM Emil: https://docs.getchip.com/chip.html#gpio
03:28 PM bss36504: Ah, ok. I was looking at the Chip Pro docs before, and they kind of just glossed through the Filesystem API demo and left it at that
03:30 PM bss36504: Is the python lib they are referencing just a wrapper around the commandline stuff?
03:30 PM Emil: Dunno
03:30 PM * CORDIC ROFL
03:30 PM Emil: Easy to check out, probably
03:31 PM bss36504: I guess it says it's "like the RPi.GPIO library, but for CHIP"
03:31 PM bss36504: However, I've never tooled around with the RPI lib so I dunno. I also don't really know python because it's a bastard language ;)
03:33 PM julius: Emil, well usually they out 1Mohm between mosfet gate and gnd...but what if you have a driver in front of it
03:33 PM julius: should the driver get the same resistor?
03:37 PM Emil: Doesn't matter if driver or not (unless the driver itself includes a resistor already)
03:37 PM Emil: One should "always" have a pulldown on a mosfet
03:39 PM bss36504: Why? If the driver is actively pulling the gate down you don't need a resistor I wouldn't think
03:43 PM Emil: bss36504: if you lose power
03:43 PM Emil: bss36504: or if your driver fails
03:43 PM bss36504: Eh, safety is for squares :P
03:45 PM bss36504: Well, in just 14 business days I'll have myself a new Chip Pro eval kit. Noice. I really like that for $49 they include the breakout board with a pro on it PLUS another pro just on it's own. That's a really nice price point. the fuckin edison with the mini breakout was $75
03:49 PM Emil: :D
03:49 PM Emil: You went all in, it seems
03:49 PM Emil: I would have just a bought a normal CHIP first
03:53 PM bss36504: Out of stock. I'm too impatient
03:54 PM bss36504: Besides, I would have just used the regular one headless anyway
04:20 PM Duality_ is now known as Duality
05:34 PM arij is now known as arij_work
06:53 PM Tachyon` is now known as TachAway
07:14 PM NoHitWonder__: what would be the fastest way to communicate between two microcontrollers?
07:23 PM Lambda_Aurigae: gigabit ethernet
07:23 PM Lambda_Aurigae: or multi-gigabit fiber connection
07:24 PM Casper: NoHitWonder__: most likelly uart due to the fast bitrate possible (you can use non-standard one) and the low overhead
07:35 PM Lambda_Aurigae: you can often get 2MB/s with usart on avr...specially running at 20MHz.
07:37 PM Tom_L: Lambda_Aurigae do you remember the original default lpt1 port address?
07:38 PM Tom_L: was it 378 or 3bc?
07:38 PM Lambda_Aurigae: think i was 3bc
07:39 PM Lambda_Aurigae: hmm.
07:39 PM Lambda_Aurigae: 378 was motherboard based
07:39 PM Lambda_Aurigae: 3bc was hercules card based.
07:40 PM Tom_L: docs i can find say 3BC was for video cards with parports
07:40 PM Tom_L: and usual lpt1 was 378
07:41 PM Tom_L: do you happen to recall the default irq?
07:42 PM Lambda_Aurigae: 7
07:42 PM Tom_L: for lpt1?
07:42 PM Tom_L: at 378
07:43 PM Tom_L: sounds right as i recall it
07:43 PM Lambda_Aurigae: yup
07:43 PM Tom_L: 5 & 7 were for pp
07:43 PM Lambda_Aurigae: yup.
07:43 PM Lambda_Aurigae: 1 and 2 often shared irq7
07:44 PM Tom_L: are the irq's that critical?
07:44 PM Lambda_Aurigae: not usually.
07:44 PM Lambda_Aurigae: more for ECP than regular parallel work.
07:44 PM Lambda_Aurigae: where the printer could send an interrupt back to the pc and stop printing and such in an "emergency"
07:45 PM Lambda_Aurigae: if you have 2 ports configured with the same IRQ then the second usually gets switched to polled mode.
07:46 PM Lambda_Aurigae: which takes more processor time and slows down the communications.
07:54 PM Tom_L: what was spp mode?
07:54 PM Tom_L: i suppose ecp is better
07:54 PM Tom_L: or epp
07:56 PM Lambda_Aurigae: epp is step up from the base parallel..that's what is needed for stk200 style programmers...it has handshaking capability.
07:56 PM Lambda_Aurigae: ecp used dma for super fast comms and such.
07:56 PM Lambda_Aurigae: also had hardware data compression built into the parallel port interface chip.
07:57 PM Lambda_Aurigae: or maybe it
07:58 PM Lambda_Aurigae: or maybe it's the bitronics bidirection that makes for the programmers to work right.
07:58 PM Lambda_Aurigae: something in there.
07:58 PM Lambda_Aurigae: spp is standard parallel port.
08:20 PM Tom_L: mmm
08:21 PM Tom_L: i'm not finding a solution to my problem
08:26 PM Lambda_Aurigae: dynamite
08:26 PM Lambda_Aurigae: or tnt
08:26 PM Lambda_Aurigae: either is effective.
08:27 PM Lambda_Aurigae: nitroglycerin works too but is a bit more dangerous.
08:28 PM Tom_L: C4
08:29 PM arij_work is now known as arij
09:17 PM xentrac: Lambda_Aurigae: wow, 2MBps sounds like a lot; I don't think I've ever run a UART at over 119200. SPI is quite a bit faster, though, isn't it?
09:19 PM xentrac: Tordek: congratulations on debricking your ATMega!
10:21 PM Tom_L: mmm that's wrong
10:22 PM Tom_L: so is that
10:24 PM Tom_L: mmm nevermind