I’ve swapped the standard DJI motors + plastic blades back to the T-Motor + Carbon fiber blades; sure enough, no rocketting up into the sky, and once more I was able to complete a flight. But…
all the old problems I had returned – excessive drift and yaw.
- the DJI motors are not powerful enough to manage the angular momentum of their heavier plastic blades
- the T-motors are not all equal – from the direction of the drift, I would guess the right-hand pair are damaged leading to drift and yaw.
So I have no choice but to buy new motors T-motors, and see what happens. Hopefully that will sort things out and I can get back to real testing of my control software PIDs rather than diagnosing bizarre behaviour due to dodgy hardware.
So frustrating wasting my very limited testing time on something that turns out to be not my fault!
whatever remains, no matter how improbable, must be the truth.
Although the video from the other day probably showed the best autonomous flight control for a while, there were pieces I couldn’t explain or was unhappy with my speculative explanation.
So it’s time to move into a completely controlled test environment – my kids’ playroom – and check step by step everything that makes up the quadcopter.
- hardware – are the arms aligned – are the top and bottom plates flat? Surprisingly, the answer to the second question is no. I was adding the new sticky feet onto the flight controller (Raspberry Pi + breadboard) and only 3 feet touched the top plate. Yet all four feet touch the table, kitchen work surface and floor. Time for a replacement top plate methinks.
- gravity calibration – it needs to be done as best as I possibly can to avoid it corrupting further test cases
- vertical take-off, hover, and landing from a horizontal platform with only the inner most PID (angular rate with feedback from the gyro) in control to maintain horizontal flight with no drift – this can only be a short flight as gyros are prone to drift
- vertical take-off, hover and landing from a slightly non-horizontal platform with the absolute angle PID also engaged to control the angle when the gyro drifts – again a completely vertical take-off with 0 drift regardless of take-off platform should be achieved even over a longer flight
- finally, add the horizontal speed PID into the mix to make sure it doesn’t change things when the horizontal speed it set to 0.
Now I was planning to do this tomorrow, but now Phoebe has failed at step one, the testing can’t start until she has her new flat top plate.