Test flight 5

Photo of the X8 in flight

The objectives of our fifth test flight were:

After configuring logging and confirming we were seeing similar levels of attitude accuracy in the field as we were in bench tests, we launched. Take-off was much more comfortable this time, and while the competition weight will be approx 1kg heavier than the test weight was, we should have enough power with the current motor and prop configuration.

The wind on the day was intermittent, with gusts of around 10m/s. It soon became apparent that our CoG was about 1cm too far back, as the wind tended to catch the plane during turns. Daniel managed to regain control each time, and flew a 30-minute flight with fairly low dynamics (< 90° roll, < 30° pitch).

After a smooth landing, we swapped main batteries, copied the GoPro recordings, moved the CoG forwards slightly and re-launched. Unfortunately, a few minutes after launch the telemetry cut out; once we got the plane back on the ground and checked the FCS status manually it turned out that the 2S / 2.4Ah battery we were using for the electronics had dipped below the drop-out threshold for the FCS regulator, so the system had shut down due to brownout.

Unfortunately, after switching batteries and logging in to the FCS CPU, we saw that the logs for the first flight only made it up to a couple of seconds after launch—the SD card contacts had shifted enough to cause some filesystem corruption and prevent further data being written. As a result, the only logs we have from that flight are the 1Hz telemetry feed—too slow and unpredictable in latency to compare with the footage.

Even worse, while investigating the problem I stepped on my JTAG adapter for the FCS DSP, so we were unable to re-program it in time for another flight that evening. Thus, we got quite a bit of footage (useful for the video component of Deliverable 2), but very little actual data.

Our learnings were:

The next flight will occur shortly, to validate our solutions to the issues we identified and to evaluate the missed objectives in flight 5: AHRS performance and accuracy based on the GoPro footage and raw sensor logs.

github.com/sfwa twitter.com/sfwa_uav youtube.com/user/sfwavideo