#avr Logs

Sep 21 2020

01:29 PM LeoNerd: Sooo.. after a lot (a *lot*) of searching, I have finally found myself a dual-simultaneous ADC chip that doesn't cost more than just buying two single-channel ADCs of equivalent spec
01:35 PM LeoNerd: The AD7356; Two simultaneous channels of 12bit 5MS/sec; cheap at only £13
03:05 PM cehteh_: uhm LeoNerd why dont you use a µC that has those build in?
03:05 PM cehteh_: for $5 or less :D
03:42 PM LeoNerd: cehte... bah. pinged out
03:42 PM LeoNerd: Anyway for posterity: I don't know any MCU that has dual-simultaneous ADCs
03:45 PM LeoNerd: cehteh: [repost] I don't know any MCU that has dual-simultaneous ADCs
03:45 PM LeoNerd: There's a few larger ATmega chips that have two ADCs but I don't know if you can trigger them simultaneously
03:46 PM vmt: soooo, you'll get two chips!
03:53 PM vmt: i'm pretty sure there aren't a lot of cheap mcus that can hit 5 MS/s at 12 bit
03:54 PM vmt: so i'm guessing that can kind of limit your options:D
03:54 PM LeoNerd: No.. I don't need it that fast
03:56 PM vmt: stm32s should have dual-simul adc, just a hint
03:56 PM vmt: for a far smaller price tag
03:56 PM LeoNerd: Yeah but I know nothing about STM32s
03:57 PM LeoNerd: whereas I have written lots of AVR code and generally know and can feel my way around it just fine
03:57 PM LeoNerd: and I have all the tooling (aaaalll the tooling) including lots of bits I've made myself for AVR
03:58 PM vmt: true, but that adc is £13 in singles, yeah?
03:58 PM vmt: that's almost an order of magnitude higher than an stm32, heh
03:59 PM LeoNerd: It'll cost me a lot more than £13 to entirely retool myself and my bench ;)
04:00 PM LeoNerd: anyway I've found other ones lately too... MAX11190 for £5 or just two single SPI chips at £3.50 each
04:00 PM vmt: LeoNerd: got a reflow oven? for a small batch even a heatgun works for smd
04:01 PM vmt: or well, just solder it with an iron, too
04:01 PM LeoNerd: I heatgun all my stuff, yah
04:01 PM LeoNerd: By "retool" I meant... programming tools, etc..
04:01 PM LeoNerd: Right next to me on my desk I have three AVR programmers, and in the next window over I have a bunch of AVR C code, avrdude, avr-gcc, ...
04:02 PM LeoNerd: I have /nothing/ for STM32
04:02 PM vmt: well, c code is quite trivially portable. st-link clones cost pennies at fleabay. the software learning curve for setup is somewhat steep from an avr standpoint, though
04:04 PM vmt: by trivial i don't mean arithmetic necessarily. just that avr programs don't tend to have massive codebases, really
04:06 PM LeoNerd: Ahah! A careful staring at datasheets tells me why a dual-sampling ADC is "better" than two single ones.. i.e. why you'd buy this even though it costs more
04:07 PM LeoNerd: The two channels are much better matched to each other, than the overall repeatability specs between chips (and thus between two different single-channel chips)
04:07 PM LeoNerd: e.g. in terms of gain/offset/linearity, or sampling time
04:08 PM LeoNerd: Critically in my case, a single channel ADC sample apature time is measured in ns, but the matching between channels of a dual one is quoted in ps
04:09 PM LeoNerd: Either is fine in my usecase (so I'll probably use two singles), but it does answer the question on /why/ these chips exist. So that's good
05:10 PM twnqx: hm. wonder if i can somehow implement dual CDC ACM into a candleLight firmware for an STM32...
05:17 PM twnqx: also, did any of you ever try to connect two CAN devices back to back, without their trasceivers?
05:25 PM cehteh: LeoNerd: some cortex or risc-v 32 bit can do, maybe not 5mbs ...
05:29 PM cehteh: LeoNerd: http://www.gd32mcu.com/data/documents/yingyongbiji/GD32_Series_of_MCUs_Selection_Guide_(Aug_2020).pdf
05:30 PM vmt: risc-v, oh no
05:30 PM cehteh: shaddup :D
05:36 PM cehteh: LeoNerd: https://www.seeedstudio.com/SeeedStudio-GD32-RISC-V-Dev-Board-p-4302.html that price and 2x adc 12bit iirc at 4mbs, but rtfm about the details
05:37 PM cehteh: hey even the redpill has 2 ADC's https://www.seeedstudio.com/Sipeed-Longan-Nano-RISC-V-GD32VF103CBT6-Development-Board-p-4205.html ..
05:38 PM cehteh: including hardware doublebuffering and DMA
05:42 PM cehteh: oh only 1MSPS ...
05:59 PM vmt: I think with at least some st's cm3s you can get more than when you use the absolute minimum sampling time
06:00 PM vmt: however, with what accuracy, ...
07:19 PM LHLaurini2 is now known as LHLaurini
10:02 PM [1]MrMobius is now known as MrMobius
11:14 PM day_ is now known as day