This is about as good as it gets – 10s with minimal horizontal drift, but a slow steady climb, and hence the 20cm drop landing. Previous flights showed a horizontal drift of about the same amount but with no vertical drift – it’s a bit of a lucky dip.
This is with the IMU filling the FIFO at 1kHz with 12 byte batches of sensor data with minimum accelerometer low pass filter of 460Hz. She’s running at ±4g resolution which corresponds to drift of about 6cm minimum drift over the 10 seconds of the flight in a perfect world; for the imperfect world we live in, I’d say Zoe is doing her very best.
The code is on GitHub.
Just one problem: I flew identical code but with the CF tri-blade props this morning, and she can’t get off the floor and I have no idea why the difference in props has such a radical affect on the flights:
- the plastic props flown in the video have two 3.25″ blades with a low pitch angle weighing about 4g
- the CF props have three 3″ blades with a higher pitch angle weighing about 6g
That’s similar enough not to explain such radical difference in behaviour. I need to understand this if I’m ever going to get Phoebe up in the air with her URF installed to measure vertical height and thus control the vertical drift Zoe showed in the video.
Onwards and upwards!