[hwh]

I couldn't find a thread discussing support for non-OP designed hardware or criteria for considering it so I thought I'd make one.  The only concrete mentions of work on support for other hardware I could find was that someone (TheOtherCliff?) is working on support for the Sparky2.

I can think of four criteria for considering support for a board:

• It has to be practical to port to, nothing that's not based on STM32 chips with at least the capabilities of the STM32F103 on the cc3d.
• It should be relatively widely used and available.
• Enough technical information needs to be available to code for it.
• And last and maybe most important, a developer needs to be willing to work it.

In general, I'd like to discuss what the rest of you think of my proposed list of criteria and what other criteria you think should be added to the list. Or deleted from it.

In specific, the NAZE32 and it's clones meet all my criteria and if no one else is already working of it I intend to.  Initially I want to support the Micro Scisky/Quantum Pico brushless controller I've ordered to build a 100 size microquad. Since it's essentially the same as a Rev4 Naze if I put in support for it we get support for the widely available Naze4 clones as well.  If the majority agree with the idea of pulling it into the main code when it's done I'll order a couple of clone Naze4s to test on.

What do the rest of you think?

[mazevx]

Hello, Im not the great specialist but I dont think that the work get paid in anyway if the devs port it to some old an weak hardware such as Naze or Flip or whatever just because the revo clones get more an more available in other countrys, quality gets better with every batch and smaller versions of the revo getting available to.
If we or they (the developers of Librepilot firmware) do so much work to port to some other FC we should take care that its reliable for future releases.

F3 and F4 based FC`s will get cheaper and better and some of them were layouted with serious skills in routing like the "seriously dodo FC"
I think and thats only my personal opininon, we should focus on modern hardware, look to the future and not to past.

[hwh]

I agree with you that newer hardware is better.  Maybe the criteria should be nothing less than F3 or F4.  I proposed F1 because that's the minimum that it's practical to support. TauLabs already has NAZE32 support in their code, we could refer anyone with a NAZE32 to them or Dronin (a fork of Tau).

What additional hardware do you think should be supported?  The only other FC that I've seen mention that someone is working on is the Sparky2.

[mazevx]

Hello, as posted above, the seriously dodo and SPracing F3 would be a worthy goal as of its proven performance. Another thing thats might be interesting is the pixracer v1 from arsovtech which was designed to be the newest pixhawk derivate but needs to get his own Software.

[Brian]

As was mentioned, I think the biggest hurdle is finding a dev that is motivated to do the port.  I think the Sparky 2 came about because it was so similar to the Revo and it was hard to get Revos at the time.  Several Revo clones have come out now that make the Sparky 2 port a bit less necessary, but it is still nice to have support for it.

As for the other boards, I'm not sure what the motivation would be.  Certainly it would be hard to find someone motivated to port to most of the F1 board unless it has some features that other boards don't have (e.g. it's very tiny, has builtin support for brushed motors, etc).  The Pix* hardware could likely be supported, and they are now coming down in size to the same size as the CC3D/Revo, but they still have an odd mix of sensors and no clear (to me) benefit over Revo/Sparky2.

I have a Pixfalcon that I bought before all the clones started coming out that I thought I might port to.  It's a nice board, but I'm not sure what it has over the Revo/Sparky2 that would motivate me to spend the time to port to it.

I think if I did another port, it would likely be to the Craziflie 2, which I think is really cool.  The AutoQuad M4 also seems like a very unique piece of hardware that would be interesting to port to, although I don't know how open it is.

[hwh]

@mazevx I agree with Brian that I can't find any advantage to the boards you mentioned.  They're all nice boards but to me there's no advantage over the current Revo clones.

I originally started considering porting to Naze32 as much out of boredom than anything.  Most of my working life (I'm retired now) I almost always had some project or another that I was programming.  In the course of looking at it I've found that TauLabs and it's fork Dronin are actively developing for the Naze32 so there really isn't much purpose to me doing it.

For now I think I'll stick to working on my little GPS/Mag adapter board and see what else comes up after that.

[xfce]

I think we can try to port some other MCU like Freescale's and NXP's

[hwh]

Is anyone making a FC based on either of those companies chips?  Most of them I can think of are based on STM32 series chips and I think I've seen a few with Atmel chips.  What flight controller were you thinking of?

[mazevx]

The only advantage is the very good layout from the dodo, nothing more.
I agree with you in that way.
The question was wich FC should be supported...
Im waiting for revo mini, be curious about the layout, Standard revo clones are very simplyfied wich causes some well known problems.
Thats the only negative point on that.

[TheOtherCliff]

I recall that the F3 series of processors doesn't have a lot of computing power and would probably only make sense for CC3D sorts of things.

Being that LibrePilot is hardware agnostic and we have some decent cheap boards available, the main reason I can imagine new boards being supported are:
- board popularity
- we need more/better hardware (CPU/sensors/ports) for things like vision processing, Linux-on-board, and sub-250 gram quads.

and as said, once there is a reason, a developer needs to be motivated to do the work.  The group of developers is not an exclusive club.  Basically any good developer is welcome to join us and will continue to be unless we get flooded with new devs (not likely).

A minor problem is that currently, adding new hardware increases the work load for everyone.