CubeOrangePlus reboot/brownout in the air

Hello guys,

I had a weird unexpected crash with a large octacopter UAV running 4.6.3 arducopter firmware. The flight controller is a Pixhawk Orange Cube + from Cubepilot. After a successful flight in Position Hold, i tried an autonomous mission. During the last phase of the route all of a sudden the drone pitched forward and flipped, which resulted in a catastrophic drop from 30m altitude.

Unfortunately when i tried to download the logs afterwards, the download froze midway and when i canceled the download and retried all the files, they were overwritten and split in 500 small (kb size) files, containing data only after the spin while the drone was freefalling towards the ground. As a result of this corruption of the SD card, I don’t have any log files prior to the instability. I will share the only valueable log i could save which is during the fall.

All the small bin files contain no useful info and are all in stabilize mode. Also, when I first plugged a usb into the FC i saw 3 log files in the minute of the crash, which indicates, along with the instability that for some reason the FC rebooted in the air, causing the drone to fall.

The FC has dual, redundant power inputs and the battery voltage was at 3.6V per cell at the time of the crash. No error messages were shown in mission planner.
Any ideas why this could happen ? Could it be a memory allocation issue due to the large number of peripheral devices or a brownout ?

This is the log :

As I’m sure you saw there isn’t much in this log to review. No battery status or RCoutputs. Vcc is low. Questionable PID’s and a mis/unconfigured Notch filter…

This is after the reboot, so i am not sure the FC had any time to show real data. What do you think is questionable about the PID ? The notch filter is now disabled, but this is not the case, the drone was flying fine before this sudden spinout. I am positive the FC rebooted in the air, but i don’t know why

ATC_ANG pitch/roll below default is usually a sign of a bad tune. The rate filters look too low as do the rate P&I gains. Hard to say about anything else as there is little reference data in the log to review and work from. You need attitude, throttle out and servo outputs data to determine much else relative to tuning.

Maybe, but this log isn’t going to tell you that.