I think the best approach is to exhibit good engineering and document things thoroughly. I don’t believe in the educational v.s. whatever else divisions. A well made system is the most educational system.
The main problems arise when things are not documented well. Good documentation also can actually reduce the pressure on the engineering side, because difficult choices about usability vs whatever else don’t have to be made. As long as the documentation is good, the system can be complicated or have experimental features or whatever, and still be useful.
SimpleFOC is relatively good compared to many libraries or code bases, but there are still gaps that require a lot of trawling through the forum to patch up. The lowest hanging fruit is to focus on common use cases and features first, which is done but I think it could be taken further. One small step here is to just have a quick discussion on what the best recommendations for most people are, here in the forum, for instance on recommended compatible hardware, then put that on the drivers page.
I think we are probably on the same page with the B-G431B-ESC1 board being one of the top options, although it certainly has issues that should be mentioned (very hard to access the pads, not many pins broken out, tends to get hot/waste power without good reason, major licensing and sourcing issues).
Thoughts on recommending the G431 MCU nucleo board?like this one NUCLEO-G431KB STMicroelectronics | Development Boards, Kits, Programmers | DigiKey
It uses the KB variant, while the ESC1 board uses the CB variant, IDK what the difference is.
I don’t see much point in using the nucleo board instead of the ESC1 unless you needed more pins or SPI, but as a general purpose replacement for blue pills, picos, and unos it makes sense.
I have never actually tried it it just seems advisable, more so than a blue pill or a Pico. Darlington transistors for the power output seems good up to 5 amps, I tried that and it went well. I don’t think this TMC6200 boards are good value.
The problem with any mosfets is they sort of need higher voltages etc. to switch, so you need a board and there is no triple half h bridge board that I know of.
Regarding sensors, I don’t actually think te as5600 is that bad as long as the smoothingsensor stuff is used, it’s clearly not the best chip I would grant but at the same time there appear to be no other boards available for magnetic angle sensors, at a reasonable price. The boards for the AS5048 or whatever were like $30 each or something on digikey. They both come with the little magnet, at least.
If someone is going to get boards made, the SDC6022 or whatever appears to be the best SPI magnetic angle sensor, in the balance, a number of people have tried it out and it appears to be good and readily available. I have like 15 complete boards with them if anyone wants one or two for testing.
All the optical encoders are totally overpriced, and mounting/interfacing them could be even harder. At the same time these magnetic sensors are not like a compass, they do have a lot of error issues and need calibration etc… I don’t see any viable other option for most people most of the time tough.
For me the as5600 is actually working out ok when I need a sensor although I am definitely going to have to go sensorless for long term.
Edit: maybe this is kind of a useless discussion and the main thing is just to get on with the third gen lepton which is bound to be cheaper and better in basically every way for our purposes than the ESC1 board.