Hello, I'm designing a hexa to carry a 1.5 kg load. I'm trying to create the absolute lightest configuration possible, so it's mostly made from CF. My question is around CG. My intuition tells me that for horizontal stability, the CG should be as low as possible. I would also think that the motion sensors would be most effective if they were as close as possible to the thrust plane. Can of any of you that have built or modified multicopters educate me on whether my assumptions are correct? My current diameter from arm-to-arm is 675mm.
I plan on using my copter for aerial video so to compensate for the weight hanging underneath, I plan on putting the battery on top. Basically keeping the CG as close to the center of the airframe as possible. I'm not sure if this is correct or not, but I will be testing this weekend. I'll try a few scenarios of CG and report back.
I plan to use my GoPro Hero2 as well as my Canon T2i. I would think that balancing the hexa is a similar problem to balancing a waiter's tray in your hand. If I stack stuff up on it, it becomes very difficult to balance. If I hang the same weight underneath it, it balances itself. Being that every copter I see appears top heavy, I must be missing something.
Everyone here is sorta on the right track, except that we're relying on automated fast control response to stabilize the craft. My experiments prove there is no way to ever get the response fast enough for manual piloting without either variable pitch or an inertial feedback system (both would be ideal, but the complexity of the former ought to give one pause). Of course, I am far from alone in this conclusion.
Therefore, the goal for a multicopter design is fast response, not necessarily inherent stability. Reducing rotational inertia of the entire airframe about the control axes is of paramount concern. A center of mass, (and as much of the mass as possible) nearest the pitch and roll axes intersection, or perhaps slightly below, is the ideal design target. Dr. Paul Pounds does a good job of providing the mathematical justification for this in the following paper (if you don't mind staring at a bit of calculus):
Thank you very much, that makes a lot of sense. It supports my desire to eliminate fancy (read heavy) motor mounts and move the ESC's to the center vs. out along the arms. It would also support why heavy lifters with big motors and props appear not to be as stable when flying without their loads.
i made a large hexacopter and kept all esc in the centre making a clean airframe this was very stable and this is flying on a cheap 20 pound kk board:
If you're getting the idea that there are quite a few variables involved here, you're absolutely right. :-)
Moving the CG away from the lifting plane will slow everything down, but as the paper attempts to illustrate, there is a point where control oscillations are likely to occur. If a large lifter appears more stable with a slung payload, it is because the control system is tuned for it (perhaps a larger P that will overcompensate with no load).
Having the ESC as electrically and physically close to the battery as possible is a good thing for many reasons. Keeping parasitic inductance to a minimum is essential when switching such large currents, so keep the wires between the ESCs and the battery large, short, and preferably twisted.