Pixhawk F450 Frame
Brief Description and Motivation
This page covers projects related to Unmanned Aircraft System (UAS), the ultimate goal is to gain deep understanding of systems related to UAS.
1 - Event analysis | Crash
While flying in Autotune mode the UAV pitched down causing the vehicle nose to tip and crash.
UAV information:
Vehicle: Generic self-build F450 quadcopter frame
Flight Controller (FC): Pixhawk 2.4.8 F
Motors: Generic A2212/6T 2200KV
Propellers: Gemfan Carbon Nylon Propeller 8045 205mm
Electronic speed controller (ESC): Generic 40A Brushless ESC
Battery: 3S HRB Li-Po 3300mAh 60C
Autopilot: ArduPilot
Version: ArduCopter 4.2.1
Crash flight log
Due to the crash the GNSS mast mount was damaged in addition to the two front motors propellers. As per the Pre-Flight log, there was no visible damage to the UAV, in addition, the UAV flew for a about 7 minutes between 21:38:00 - 21:45:00 before crashing which indicates that the reason for the crash was not there at the time of taking off .
Failure Analysis
First of all to eliminate the possibility of a hardware failure I used a servo tester and tested all four motors individually, to find out that all four of them were ok.
Digital servo tester used
Furthermore,
Second of all, I transferred the DataFlash logs from the FC and generated an Auto Analysis report from Mission Planner which indicated that there was a failure in the compass with a " Large change in mag_field (275.40%)"
Mission Planner generated report
Moreover,
Third of all, after further digging there appears to be a spike in the magnetic field of the internal and external compasses of the UAV at around the hour 21:45:10. Given that the UAV was flown at the same area, therefore, an external environment source of interference was eliminated. Moreover, a battery level buzzer was installed on the UAV in order avoid depleting and damaging the battery and after flying for about 7 minutes and around the hour 21:45:09 the battery level reached the set limit of the loud buzzer (10V) trigging a loud beeping noise and the engaging of the speakers within the buzzer created a magnetic field that interfered with the magnetic field of the compasses and because the UAV was in Autotune mode at that time and the heading of the UAV is controlled by the compasses this resulted in the crash.
Battery level buzzer used
Actual flight internal (0) and external (1) compasses magnetic filed in the UAV X axis
Actual flight internal (0) and external (1) compasses magnetic filed in the UAV Y axis
Actual flight internal (0) and external (1) compasses magnetic filed in the UAV Z axis
Moreover,
Fourth of all, to further test the crash conclusion. I simulated the crash set up, I started the UAV with no propellers on and applied a stable throttle of 20% to avoid the UAV disarming, next I used a discharged battery to trigger the battery level buzzer. It was verified that indeed the buzzer is interfering with magnetic field of the internal and external compasses causing the crash.
Simulated experimental flight internal (0) and external (1) compasses magnetic filed in the UAV X axis
Simulated experimental flight internal (0) and external (1) compasses magnetic filed in the UAV Y axis
Simulated experimental flight internal (0) and external (1) compasses magnetic filed in the UAV Z axis
Resources
https://ardupilot.org/copter/index.html
Skills Involved
Failure Analysis
ArduPilot
Mission Planner
Pixhawk