Since ballcopters come together quickly, this will be a quick update. I assembled the foamcore puzzle from last time, and with my package of parts having finally arrived (just in time!), BallCopterTwo is now all assembled minus the code that makes it do the copter thing.
….which I will finish. I promise.
Here are the 1000% less shitty control flaps, attached directly to mini-servos. The servo “horns” are custom 3d-printed affairs that have a 5mm side channel into which the foamcore panel slips and is retained with hot glue. This arrangement is so much more response and just… better than the previous failure of a linkage. This alone will probably make the difference. However, given that all the flaps are independent, they will need to be software controlled.
And the bottom side. The hinge is just Kapton tape wrapped around the structural carbon fiber rod. There is one catch, though – I attached a small section of Kapton tape, sticky-sides together, to the center of the hinge so it can move without sticking. Otherwise, the tape wrapping around the carbon fiber would make it too stiff.
The puzzle piece fully assembled.
I did alot of this on what amounts to autopilot (heh, autopilot), so I forgot to take pictures of the battery installation as well as the super-slick 12-servo (4 inputs, 8 outputs) interface board I whipped up using some perfboard, an Arduino Nano, and a Sparkfun 6DOF Razor IMU; but they are forthcoming. That contraption has been tested for basic functionality and mounts immediate above the blue servos. For now, though, Ballcopter is mechanically and electrically done – not electronically yet, since there’s codes to be written.
The final all-up weight is 19.5 ounces or 550 grams. Much heavier than the J-pop ball drone (350 grams), but this thing is also bigger and has 8 chunky miniservos (14 grams instead of 5 or 6) and a massively oversized 1Ah battery. Yes, I’m still used to sizing batteries up for electric vehicles.