I think it's the likes of JLCPCB making it cheap to have a go, and a lack of perspective about the difference between building a board to have a Pico slapped on it, just connecting some peripherals, vs putting the naked RP2040 on there.I am surprised anyone who doesn't have prior experience of designing high-speed digital electronic layouts would even embark on such a project, would expect it to work.
This is a whole different level of ball game to slapping a low MHz microcontroller on a PCB or breadboard. I am not sure what makes them think they would succeed, how they would ever expect to resolve issues without having the appropriate test equipment and diagnostic tools. Some don't even have logic analysers or oscilloscopes. They are ill-equipped to determine issues, are inexperienced, and often don't even know where to start in identifying issues.
Having said all that, it's not in fact extremely hard. It's not the sort of high speed design that needs controlled impedance, tightly matched track lengths etc.
Maybe the issue is RP2040 documentation, a lack of warning that it's not easy, things can be difficult to resolve if it doesn't work, may require skills, equipment, and experience when it doesn't ?
One thing that I think would make a huge difference is a recommendation "if you are new to this, use a 4-layer PCB". Doing a good 2-layer RP2040 layout, especially one that uses all the pins and needs to fit in a limited space, is significant work even for someone competent. But even the worst designs we've seen here in terms of layout would probably have worked if you stuck a groundplane under them.
4-layer is really not expensive nowadays - I sometimes use 4-layers even on commercial RP2040 jobs, because even in moderate volume the cost saving doesn't pay for the extra time spent doing layout (or the extra risk of failing EMC testing).
I was expecting it to be a forum for help in designing RP2040 boards (though I'd hoped for more discussion before the event rather than after!), but I agree that support for running things on other off-the-shelf RP2040 designs is a valid and quite separate topic - it's surprising that there's not more posts on that front.I also think there needs to be a separate "Fix my custom board" sub-forum as my understanding is "Other RP2040 boards" was meant to showcase off-the-shelf commercial boards using RP2040. Those have been drowned in a deluge of "help, I tried and failed" posts. I never expected so many to actually try, but it doesn't surprise me so many have failed.
Statistics: Posted by arg001 — Tue May 07, 2024 4:34 pm