cool, but unstable! Here’s what the drone looks like now:
Upgrades include new blades (lightweight, stiff carbon), new motors (higher power, lower weight), and new back legs (black). I’ve also swapped the RPi model B for a model A in my PiBow variant case showing in a previous blog entry.
Why? The model A swap was for using less battery power and lowering the total weight and the centre of gravity; the motors for increased spin power for the same battery power usage and lowering weight; the blades for lower weight and greater lift; and the legs cos they’re cool!
Even though it’s still not stable, things are getting better – primarily due to fixing a couple of bugs in the code when changing from take-off to hover mode, but also because of setting PID gains. My starting point was here, and this article does a great job of explaining how the PID works.
I’d recommend watching his video.
I screen captured the PID diagram as it’s so helpful showing the quadcopter double tier PID control that allows stability to be controlled directly from the gyros by the inner PID, and the attitude (stable angle of lean in any direction) to be controlled from an operator (me) to be controlled by the outer PID. Certainly, the lack of low pass filter for the gyro feedback is something I’ll be investigating, although the MPU6050 does have one built in. If that doesn’t work, it’s fast fourier transforms (FFT) all the way!
P.S. You ain’t seen nothing yet!