Closed superjax closed 6 years ago
Hey @drinkdhmo, thanks for volunteering to help with this.
Here is my Revo.ioc file from STM32CubeMX. I am pretty sure that I probably messed it up, so it might be worthwhile to just start over, but if it were helpful I just wanted to make sure you had it.
Okay, I have this working. At the very least, communication seems to be working nicely. Current loop rate is only 45 us. Holy smokes!
Things still to do
Known Issues Unable to flash with debugger while USB is connected. PWM outputs on the RC input pins aren't working (limits us to 6 max outputs for now) Unable to mix PWM outputs (can't mix output rate for servos and ESCs)
Is it planned to move the F4 documentation onto rosflight.org? Just a suggestion because checking out this library I had no idea it supported anything other than F1 controllers until going through the issues.
Hi @atyshka, thanks for the suggestion. F4 support is still under development and hasn't yet been released (should happen soon! Relevant pull request is #278), but once that happens then yep we plan to move the documentation over.
I just wanted to write an update for my progress in building F4 support. I've bought a revo from ready-to-fly-quads link. And I started by trying to strip the relevant drivers from cleanflight, betaflight and raceflight
This was a horrible experience. I can't believe that cleanflight labels itself as ''clean''. There are some absolutely terrible coding practices going on in cleanflight, and after about 20 hours, I gave up, and decided to write my own. In about an hour, I started getting some functional builds.... I feel stupid for even trying to do anything with cleanflight.
The biggest difference, besides the faster processor, is native USB support. I had never worked with a USB peripheral on an embedded controller before, so that was something new. I was hoping that I could bypass learning about how it works by relying on cleanflight, but as I already mentioned, turned out to be a terrible idea. It's not all that complicated, but I need to make sure that it is using asynchronous callback methods or DMA or something like that. I believe that we will be able to have even higher bandwidth on the VCP, (virtual COM port) than we have seen previously.
Anyway, now that I have decided to completely leave the stupid flight XXXXflights behind, I think I'll have a working solution in a couple of hours. I intend on replicating all the BreezySTM32 examples with an F4, but I'll probably leave it as a separate repository just so it doesn't get complicated