Thank you very much, guys. Regarding the noise, the motors are dynamically balanced and the props are statically balanced. I am thinkning I should try inserting some foam inside the arms.
I did several more flights today and I was able to reproduce the bounce, this time in position hold. Here is the log file:
https://www.dropbox.com/s/6qq58l3qvznuv47/014-AQL.zipI've made the following observations.
- Reducing Navigation Altitude Speed IM and OM to 500 reduces the overshoot and in some cases the AQ is able to recover on its own (no manual control intervention). I've also reduced Navigation Altitude Position PM and OM to 1.8, but I can't tell that has any effect. My idea was to "tone down" the loops a bit. If Navigation Speed Altitude I is less than 1.5, then AQ does not hold altitude very well. Vales above 1.7 make the quad go up and down with high frequency, so I ended up with value of 1.5.
- The chance for AQ to bounce increases with the wind speed. I've observed the quad sinking slowly during a gust, and at some point it throttles up, but overshoots, and the cycle repeats with increasing amplitude.
- When running on 3S, I could not reproduce the behaviour. But I use 4S batteries and with them the quad is more "nervous" in general.
Your thoughts?