The objectives of our sixth test flight were:
Having enabled the DSP to boot from SPI, the set-up and launch process was considerably quicker than for test flights 4 and 5. After enabling CPU logging to SD and running through the (optional) magnetometer calibration process, we launched into the moderate wind. Take-off was uneventful and handling was fine, although we noted a region of stronger wind just above the treeline near one end of the field.
Shortly after launch, we lost telemetry data (but retained signal), which suggested a DSP thermal failure. Following a smooth landing, we observed that the CPLD was resetting continuously, which indicates that the DSP’s PMIC was in thermal shutdown (> 75°C). At that point we realised that the location of the main battery would block most airflow over the heatsinks, so we re-positioned the battery and added some additional weight to the nose to maintain CoG location.
We powered back up and enabled logging, then re-launched; telemetry remained active during the whole flight and after landing the DSP and CPU were well within the acceptable temperature range. We’re looking at more robust thermal solutions, and will likely incorporate a small fan into the enclosure; surprisingly little airflow is required, but in the absence of that the board will only run for a couple of minutes before shutting down.
The modifications we made to ensure the CPU’s SD card remains in place worked, so we obtained sufficient log data to validate AHRS performance against the horizon extracted from the on-board footage.
Playback of the 50Hz AHRS output via X-Plane demonstrated good position, velocity and attitude accuracy, however the altitude data was very poor. It’s likely that this is due to airflow over the barometric pressure sensors, which would be much higher than in the previous tests.
Our learnings were: