I have been flying Reptile twin-engine for some time using F-765 FC. One morning after ~24th flight, I launched the plane in auto takeoff mode as I always do, the plane reached its destination and started to circle but before it can complete one circle it dropped banked to the right, and hit the ground. I tried taking control by switching to FBWA mode but it was too late.
Please review and comment on the log file if you see anything odd.
Nothing jumping out at me. I think it’s a mechanical or C of G issue. From the time the plane starts to lower the nose it seems relatively unresponsive to the control outputs the FC are trying. I don’t think the plane stalled because the progression is smooth.
The moment the plane seems to stop responding there is a bump in the z-vibration. Any chance something came loose inside the plane?
I did noticed the battery voltage dropped very low for the duration. Down to 13.2 during the take off. (or are you using LiIon?)
Set SERVO_AUTO_TRIM=1. Since the plane is always trimming during flight with this on, if you notice extreme or unusual control positions this may be a clue that your C of G is not right.
Thanks Allister for reviewing the log.
We can rule out CG since I always test my planes in manual mode during maiden flights.
Could the FC came loose? It was held down with 3M VHB tape.
I was flying with 4s Lipo. I think at takeoff initially motors draw insane amount of current so voltage drop is expected, but 13.5 volts is way too low.
Not sure what Servo_auto_trim is, but I’ll go to the document to learn about it.
Last request, can you share which log parameters do you display on the graph to determine what is happening during flight. Teach me how to fish
I don’t think it was the FC that moved, but something else in the plane. Most likely the battery since it’s typically the biggest weight. I wondered if something happened with one of the motors like a prop came loose, that’s what got me looking at the current and voltage. But the current stayed pretty consistent and the vibrations only had that bump, not the rise a motor issue will cause.
When I look at logs I start with the desired pitch and actual pitch (and same for roll). On a properly tuned plane the desired and actual should move pretty close together. When those don’t line up, then I start looking for why. Often I’ll look for what the servo outputs are doing. Are they moving appropriately? Is the plane reacting to the servo movements?
If you want to hear it from the real experts (and you have some time): ArduPilot Log Analysis Seminar 16th August 2021 - YouTube