I looked some more at this flight. Although there was a "compass error" at 9.4 it was not caused by a bad calibration or the AC coming near ferrous material. Although I had raised the possibility of a propulsion issue and the pilot stated that "sounded like three of the motors was at full rev " there were no propulsion issues evident in the .DAT data.
The pilot stated that the AC went straight into the ground and that the battery came out on impact.. However, the .DAT ended at 12.063 secs showing a speed of 15 M/S and 20 Meters above the home point. This shows that the battery came out before the AC hit the ground. The descent rate was 6 M/S implying the AC impacted 2 or 3 secs after the battery came out.
The interesting part of this flight is the incident at 8.8 secs. Leading up to this full elevator had been applied at 6 secs and then removed at 8 secs. The AC pitched up in response and just about slowed to 0 M/S when there was an acceleration spike along the Z axis at 8.8 secs.
The acceleration was -2 G spiked to -4 G then resumed at -2 G. This data was extracted at 200 HZ instead of the default 30 HZ. The spike is not a 1 sample spike; there were 14 samples in this interval. I.e., this is not some data glitch. Looking at the other two acceleration axes.
This implies that
if the AC were hit with a projectile then it would've come from below and to the left of the AC.
Following this event the AC finished braking at about the same time the data says the AC pitched down, rolled left, and yawed CCW. I compared the rate at which it pitched down to a flight I had where I was getting data about braking maneuvers just like this. In comparison this flight was more aggressive in it's pitching down. There doesn't seem to be any reason for the AC to roll left.
The compass error came 9.4 secs with a switch to ATTI mode. Normally, the AC will just hover but in this case it flew backwards. The Yaw is always suspect when there is a compass error so the direction of travel from the Google Earth display and the thrustAngle was used to determine that the AC was flying backwards. thrustAngle is a diagnostic that uses motor speed data to determine the direction relative the AC.
magYaw is a DatCon diagnostic that computes yaw independent of what the AC's flight controller thinks the yaw is. It's been used several times to explain the cause of a compass error. In these cases Yaw (from the FC) will become different from magYaw with one being correct and the other incorrect. In this flight Yaw and magYaw become different but both appear to be incorrect.
At 8.9 secs Yaw is 32 degrees, magYaw is 157 degrees, but the heading (from the direction of travel and thrustAngle above) is 116 degrees. Note that 9.4 secs the FC declared a compass error and switched to ATTI mode.