#avr Logs

Aug 16 2017

01:32 AM _ami_: enh: cool
01:54 AM day_ is now known as daey
07:36 AM remkooo1 is now known as remkooo
10:37 AM day_ is now known as daey
12:18 PM KungFunk is now known as Funklord
12:21 PM kre10s: say I want to encode/decode SPI as RTZ or NRTZ is there a chip for that? Like a transmitter and reciever that will do the clock extraction?
12:22 PM LeoNerd: I happen to know of a chip for doing SPI over a single DC-balanced twisted pair, actually
12:23 PM polprog: interestin
12:23 PM antto: jackpot
12:23 PM LeoNerd: The LTC6820
12:23 PM polprog: i was about to suggest writing an NRZ encoder snippet
12:23 PM antto: i was going to say "that's perverted"
12:24 PM polprog: a bit of assembly never killed nobody
12:24 PM antto: polprog look at this graveyard, do ya see it? ;P~
12:25 PM polprog: i was going to say that if you need isolation you better go rs232/485, spi is not designed to be sent over long distances
12:25 PM polprog: imo
12:26 PM kre10s: hmm. neat! now make it to 100Mbits/s
12:26 PM antto: don't forget that SPI is synchronous
12:27 PM LeoNerd: SPI is fine over long distances, provided your master has a receive clock
12:27 PM kre10s: I actually only need one direction.
12:27 PM LeoNerd: Annoyingly the LTC6820 doesn't output a receive clock :/
12:28 PM antto: a wat?
12:28 PM polprog: 3 differential pair drivers for 3 lines /s
12:29 PM antto: if you only need master transmit, you could use a synchronous USART on some chips
12:29 PM LeoNerd: antto: To account for timing latency along a long cable, it's often useful to refect the clock signal back down the cable along with the MISO signal, and have the master use that reflected clock, rather than its own generated one, to clock in the receive bits
12:30 PM LeoNerd: That usually requires two SPI modules
12:30 PM antto: ah, that's a cheat
12:30 PM LeoNerd: But you can get away without it if you're only doing transmit-only, or at sufficiently slow clock speeds with respect to the cable length
12:31 PM kre10s: I really only need to get data from a to b at high speed. I have a data and a clock line but want to transmit with a single differential pair or fiber etc.
12:31 PM LeoNerd: Consider UART on RS-485 then
12:31 PM polprog: ^^
12:32 PM LeoNerd: E,g DMX512 uses that; clocked at 250kBaud, and that works totally fine across noisy stage equipement over maybe hundreds of meters
12:56 PM Lambda_Aurigae: abuse ethernet...just send packets, no response needed...
12:56 PM Lambda_Aurigae: one pair.
12:57 PM bss36504: Trying to remember who I was talking to about eagle 8.3 the other day...anyway, sparkfun has a nice writeup on it: https://www.sparkfun.com/news/2456
12:57 PM bss36504: Neato features
01:06 PM Lambda_Aurigae: but is it still as good as before it got bought out?
02:50 PM day is now known as daey
03:43 PM MrFahrenheit: 80% of that writeup seems to advertise sparkfun
03:44 PM bss36504: MrFahrenheit: Well it was written by sparkfun....so I guess i don't see your point
03:44 PM bss36504: At any rate, it does explain some of the new features, and showcases the interoperability with other autodesk products, which is nice.
03:45 PM MrFahrenheit: you called it a nice writeup, I see an ad, that was my point
03:46 PM MrFahrenheit: I don't see why eagle is so popular, it's very mediocre software
03:47 PM bss36504: Yeah i mean i guess if you want to distill out all the stuff they talked about eagle, then sure, it's an ad.
03:47 PM polprog: eagle was usable for me, till i hit the fre version size constraint...
03:48 PM bss36504: I understand that for sure. Definite pain. However, I try to make my boards much smaller than that to cut down on mfg cost, but that's personal preference. I understand sometimes it isn't possible/desirable.
03:48 PM Emil: There is simply no reason to using KiCAD
03:49 PM Emil: to not use KiCAD*
03:49 PM Emil: There is no reason to use Eagle
03:49 PM bss36504: For me, not learning a new CAD package is a plus
03:49 PM bss36504: I have shit to do, and eagle gets me there faster
03:49 PM Emil: It's rather intuitive
03:49 PM bss36504: eh
03:49 PM bss36504: debatable.
03:49 PM polprog: what i definitely hate is altiums schematic look. their color scheme and SERIF fonts on a technical drawing look so awful my eyes bleed
03:49 PM bss36504: I did not find it intuitive
03:49 PM bss36504: polprog: yeah, the schematics are dumb
03:49 PM bss36504: but the layout tool is quite nice
03:50 PM polprog: indeed
03:50 PM polprog: i watched some altium tutorials, hoping ill work on that at work
03:50 PM polprog: but they said use whatever you want
03:50 PM polprog: so kicad it is then, probably didnt want to pay for another license
03:50 PM bss36504: Yeah, the price tag really limits it to large-dollar companies
03:51 PM bss36504: FWIW I really have nothing concrete against KiCad other than "I know how to use eagle and i 'just didnt like' kicad when I tried it"
03:51 PM MrFahrenheit: kicad still needs a lot of streamlining
03:51 PM bss36504: But that doesnt mean I wouldn't check it out again. I'd really like a good package that can do board level parasitic extractions.
03:51 PM polprog: i know, it is a pain to learn a whole new cad package
03:52 PM MrFahrenheit: the 3 render engines that can't do the same stuff is really weird
03:52 PM MrFahrenheit: smells of super leaky abstractions
03:52 PM polprog: the biggest issue for me in kicads 3d engine is that it goez nuts when there's a 2 board pcb
03:52 PM polprog: also that
03:53 PM bss36504: right yes, the different render engines. Real big WTF when I was learning about that
03:53 PM bss36504: still don't fully understand the point
03:53 PM polprog: i dont see the point of cairo engine tbh
03:53 PM polprog: it looks exactly like opengl but slower :P
03:54 PM MrFahrenheit: it used to look better at one point
03:55 PM MrFahrenheit: it's also software rendered which can be useful sometimes
03:55 PM polprog: funny though i nver got gEDA to work...
03:56 PM polprog: wtf pulseaudio is taking 1,2 GB of ram
03:56 PM polprog: lol
03:57 PM bss36504: lol
04:00 PM antto: bss36504 legacy mode .. it should be self-explanatory
04:00 PM antto: then someone made the so-called "GAL" which is a new thing, changes many things hopefully for the better
04:00 PM antto: and with GAL you have two flavours which work the same way but are rendered differently - opengl and cairo
04:01 PM antto: cairo is good to have because not everybody has a fancy video card
04:01 PM antto: is there anything unclear so far?
04:02 PM antto: there really is difference between legacy and GAL, and the legacy mode is supposed to vanish one day
04:02 PM antto: then the sh*t is gonna be consistent
04:55 PM kre10s: hmm. I could solve my problem with a bunch of logic chips. or a CPLD. Is there a CPLD manufacturer where there exists a complete open source toolchain?
04:56 PM kre10s: or at least all the parts needed to make it work. a simulator wouldn't be stricktly necessary...
05:25 PM Emil: kre10s: unlikely
05:25 PM Emil: kre10s: but you could use the project icestorm
05:25 PM Emil: for fpga
05:28 PM MrFahrenheit: although apparently the altera quarus prime works on linux https://spinorlab.files.wordpress.com/2016/05/quartusprime2.png
08:37 PM enh: hi
08:56 PM grazfather: hey guys, i am disassembling a attiny45 hex and it looks like there are calls outside of the address range, e.g. 0x8xx, what would this do?
08:57 PM grazfather: the attiny45 is 0x1000 flash, but it's word addressable so only 11 bits are used, meaning addresses only go up to 0x800
08:59 PM tpw_rules: i bet your disassembler is exposing byte addresses
08:59 PM tpw_rules: like it would be an error to have it be odd
08:59 PM tpw_rules: (please don't tell me that xx is sometimes odd :P)
09:05 PM enh: no idea
09:16 PM grazfather: tpw_rules yeah i thought that might be the case, but other calls are fine, and YES there are odds ones
09:17 PM tpw_rules: oh wait
09:17 PM grazfather: would any of you be able to disassemble raw bytes?
09:17 PM tpw_rules: i remember
09:17 PM tpw_rules: look in the datasheet for the memory mapp
09:17 PM tpw_rules: flash does not start at 0x0000
09:17 PM tpw_rules: i think it might be 0x0060 or 0x0100 for that part?
09:17 PM grazfather: that's where sram starsts
09:18 PM grazfather: reset vector is at 0
09:18 PM tpw_rules: ugh i'm all confused
09:18 PM grazfather: and as expected at that address it jumps to a function
09:18 PM tpw_rules: yes you're right
09:18 PM tpw_rules: don't listen to me
09:18 PM grazfather: f7e1e010 seems to disassemble to call 0x887
09:18 PM grazfather: lol :p
09:19 PM grazfather: know of any non-IDA disassemblers? I could just compare. I'll try gcc-avr
09:25 PM enh: avr-gcc
09:25 PM grazfather: yeah that's what i meant
09:26 PM grazfather: but unfortunately it seems i can't just load/map a binary to get it to disassemble, i'll have to load simavr or something
09:29 PM enh: I don't know the right workflow, but what you wrote seems right
09:29 PM enh: Ate least the code you have must be translated to machine language of the right microcontroller
09:30 PM enh: if you use mac, there is an avr simulator for it which is pretty straightforward
09:30 PM enh: unfortunately i do not remember the name
09:35 PM grazfather: yeah i don't have any code, i just have a flash dumb unfortunately
10:00 PM JanC is now known as Guest6447
10:00 PM JanC_ is now known as JanC
11:39 PM day_ is now known as daey
11:52 PM Casper: rue_shop2 / rue_shop3 / rue_bed / rue_more: ya there?
11:58 PM rue_shop3: of course