Proven aircraft with hundreds of hours on it, flying an AUTO figure 8 pattern fell out of the sky. No indication from LOGs that any error was populated. Log stops the second the plane locked up. Was flying fine for about an hour. Batteries test good, not full drained still plenty of charge. Power was still on, Pixhawk just completely stopped responding. GCS operator also saw no connection to plane instantly. Airplane is gas powered and engine kept running, servos kept last known position and pilot indicated no ability to control aircraft on the way down. Attempted to take plane into MANUAL but log stops well before that attempt. Any help on why this might have happened is greatly appreciated, and, if I have this in the wrong area please let me know and I can move it
Also, if there is any more information or details you would like feel free to ask I should have them.
I would agree with you Karl but still had power on the radio which was in the same harness as the power for the Pixhawk. It does not seem like power was lost to the system. Motor continued to run all the way to the ground, the aircraft seemed to just stop taking servo inputs and completely locked up.
The log file just ends. I agree with @Karl_S . That’s pretty consistent with a power failure to the flight controller. Look on the SD card to see if there’s any crash dump or other files that might give clues.
SD card shows no additional files other than the log provided. We run redundant power into the top of the Pixhawk (PIXHAWK 2 ORANGE). Dual batteries with different regulators on PWR1 and PWR2. Also, if our rail lost power, we should not have also lost logging. Loss of logging would have to be loss of power on top of the Pixhawk in PWR1 and PWR2 which was redundant. Engine ran all the way to the ground and RC radio did not report loss of receiver on the way down, which is also powered by the rail.
We did have RC passthrough setup so if power was lost on the Pixhawk, it should go into passthrough mode and allow the pilot to recover the aircraft.
To answer your other question, the rail is powered with raw battery voltage (2S 5000). Servos are not sensitive to this voltage change. I do find it interesting that the voltage dips that much near the end, will investigate that further.
From looking at the other power log data I do see some anomalies. Looks like jumps on the Vcc is correlated to a BRICK_VALID going to 0. This seems to ramp up as the flight goes on, then gets the most violent right before the loss of the aircraft. Not sure what caused this and in initial testing, have not been able to replicate it on the bench. The voltage drop is now much though, only down within 4.9v… this is suspect to me.
Also, I can correlate the loss of GPS with the power drop on Vcc in the data logs.
