A few months ago, I posted some video and pictures of a tilt-rotor quadcopter called iQuad. Since then, I've been working on building a new version of iQuad which is lighter, has a cleaner design, and adds FPV. The new iQuad weighs 1600 grams w/o batteries (down from 2260) thanks to a carbon fiber body and custom-built rotation mechanism and landing legs. My goal for this version is to achieve faster forward flight and eventually add 3D-printed wings to get airplane-like flight characteristics.
The previous iQuad used unmodified arducopter firmware, with the tilt servos being directly controlled by one of the channels on my radio, which limited the tilt angle at which the APM was still able to produce stable flight. For this version, I am going to control the tilt servos directly from the APM, for which I'll need to write some custom firmware. (any help would be appreciated as I am new to arducopter development)
Here are some more photos:
Comments
Great result
I ordered the bearings, gears, aluminum tubes, standoffs and screws from servocity. For the "gear hub", servo mounts and legs, I actually used an online service called eMachineShop. I'm not very good with the more professional CAD packages, but they have some free CAD software that is pretty easy to use and you can actually verify, price and order parts from their software. It was surprisingly easy to use. A month later, I had some beautiful custom made parts.
Interesting project ! Where are the alumn components sourced from for the bearing supports and the gear hub?
Nice work. I'd for the first couple of steps where you're just trying to get it to fly well in stabilize mode you'll be ok with an APM2. CPU runs at about 70% ~ 75% in stabilize so you've got some extra room if you're careful.
If you get it working to the point that it's going into master it'll need to be in a separate motors library I suspect but that doesn't necessarily mean you need to do that as a first step. The control problems look pretty tough on their own. best of luck!
Thanks for all the advice guys... it sounds like I'm going to be buying a pixhawk and getting familiar with 3.0 very soon. I also realize doing a full conversion is going to be very complex, but I'm going to take it in baby steps. My first goal is to just recreate what I did in the first iQuad with servo control being performed by the APM. The next goal will be to get controlled flight with progressively higher tilt up to about 45 degrees. If that goes well, I'll try adding wings (which I'm sure will introduce a whole bunch of new control problems).
I don't claim to be an expert on this, but I have looked a bit at this use previously for a quad that could convert to a plane and the reality is that the combined control problems are significant enough to (require) at least a Pixhawk or PX4, the APM simply does not have the memory (or CPU cycles) available to serve as a suitable platform for this.
On the APM the CPU cycles available are seriously marginal for the existing control scheme, compounding the calculation as much as will be required to progressively mix the roll, pitch and yaw according to tilt and velocity will almost certainly break it's back.
The PX's could probably handle this hardware wise, the APM simply cannot.
Also, the current existing examples are simply quads that convert to normal airplane use, they really don't even begin to address the mixing problems resulting from transition, most of them just try to fly up high enough and perform a rapid enough transition where they can simply switch from Quad mode to Plane mode.
Possibly if you were only going to do a very limited swivel forward <10 degrees you could get some small benefit in forward velocity because the frame would still be aligned to the relative wind, but I would be surprised if you picked up even 10 percent in maximum forward velocity and the changes in lift of the tilted rotor moving horizontally would still require rapid and complex real time compensation.
I guess with a complex enough control code you could take of normally and transition in a hover to straight vertical and then continue the transition till you were completely upside down if you wanted, but that is going to be one serious piece of code and of course it gets a lot hairier if you actually want to move while you are in any of those intermediate points.
3.0 is better in every conceivable way. :) I wouldn't even consider using anything else.
I do not think you'll get adequate roll control using torque. While you may be able to minimize the moment of inertia of the frame, the aerodynamic loads remain.
And in condition such as, 30° pitch forward, you will have significant issues where, roll control via thrust differential will produce massive amounts of yaw torque, I really don't think you'll be able to counter it with torque.
And the math we are currently use to do the Earth-Frame=>Body_Frame rotation is not that complicated, it is simplified by the fact that, the motors are rigidly aligned in the frame, and you have 3 mutually exclusive torque outputs which rotate with the frame.
What you will need to do is Earth-Frame=>Body-Frame=>Motor-Frame. And that final step is going to be fairly complicated. I actually can't figure it out myself, maybe Leonard or somebody else with an understanding of advanced math will.
@Emin and @Marcus, thanks for suggesting wingcopter. I knew about their project, but I don't know how helpful it will be since they seem to be using traditional control surfaces in forward flight. That being said, I think it's worth a closer look to see how they went about modifying the code, how they set up their radios and lessons learned along the way.
@R_Lefebvre, it's funny you mention having four independent servos. That's actually what I had designed in my first prototype. The mechanics were much more complicated and heavy, but the bigger issue I had was that the servos were very difficult to synchronize. Even with the programmable servos that I am using, they moved at slightly different rates, and I felt it would be very difficult to get around the problem. So, I ended up tearing out that system and using just the single front and back servo.
So, as you pointed out, this design will have to rely on using cross-axis torque for roll control as the props tilt forward. I am hopeful that this will work because most of the weight of the model is on the center axis, so the angular moment on the roll axis is quite low.
The math for mixing yaw and roll is not that complicated. There's already and example of it in the arducopter codebase since the same thing happens when a regular (non-titlrotor) quad pitches forward. The math for controlling pitch will be slightly more complicated since I will need to compute the desired forward and vertical thrust vectors for the two front and two rear motors and then use an inverse tangent calculation to compute the servo angle. To be honest, I'm not too worried about that part, I'm actually more concerned about working with the rest of the codebase since I'm so unfamilar with it.
What are the advantages, in my case, of using version 3.0 or later?
Email this guy from wingcopter,he always answerd all my questions in detail...his project is somewhere on this forum as well...
-
2
-
3
-
4
-
5
-
6
of 6 Next