I first tried out FPV on a small airplane called the dolphine but I realized without a stable flight, flying FPV may turn out to be uninteractive and be a poor experience. Learning from the previous project, I decided to go on for a larger and stable platform and with flight stabilization. KK2 is a cheap and easy to use flight controller for multirotors. However, there is a firmware called OpenAero that uses the attitude estimation from the board but to control a fixed wing aircraft.
The KK2 Board was mounted using a custom 3D printed part modified from the 3D Modelfrom thingiverse.com. The 3D model shown in the picture above can be downloaded here. The Camera holder and mounting holes for the KK2 was added. The KK2 was fixed to the 3D printed canopy via a M3 bolt and the canopy is fitted to the airplane with magnets.
Tutorials for setting up Openaero can be found here. I had problems setting up the KK2 board to change modes by the transmitter. I was using the Gear channel for switching modes (Manual mode, Stabilize mode, Stabilize+Autolevel mode) but didn’t work. After fiddling with the control gains, I found out the scale of the receive inputs and the scale of the flight profile values are different. It now works perfectly.
- Aircraft: Hobbyking Bixler 2
- Video Telemetry: Immersion RC 5.8GHz 600MW Transmitter
- FPV goggles: Fatshark Predator V2
- RC Transmitter: Turnigy 9x (mode 2)
First flight was very successful and I was very impressed with the performance changes in each modes. It was a fairly windy day and the plane was rocking around due to the wind. As I triggered autolevel mode, the plane was locked in within it’s attitude and was easier to control. What surprised me was the sluggish pitch response the airplane was showing in autolevel mode. This must be due to the low dynamic gain settings in pitch control, but it was very interesting that the control response gets slower when the aircraft was controlling it’s attitude. I also expected the plane to maintain it’s altitude in autolevel mode but soon found out that while the plane stayed level, altitude changed by the thrust. When trying to land the plane, it was very interesting to see that the high angle of attack due to the descent slowed the aircraft dramatically. Maybe the vertical landing capability shown by Cumulus One was done through autoleveling and reducing speed to push the aircraft into a stall, while still actively controlling the attitude of the aircraft. Had a small crash into a tree but fairly repairable.
The weather was a little bit more windy and it was soon obvious that the pitch gain in stabilize mode was making the airplane oscillate in high speeds. After tuning the gains, the aircraft was locked in to the direction it was flying.