Mid Autotune Lane switch 1,0,2 GPS Glitch

Curious why autotune needs to look at “GPS”?

Luckily this drone comes with a cage, knocked the indoor structure, and did not crash.
This is not the first time I use this indoor place for autotune, in fact, many successful no-wind autotunes at this location.

Vibrations look within the limit before the knock.
Yaw 180 degrees suddenly and drifted. Everything happened very fast for a proper reaction.

Flight log.
CubeOrange
4.3.3

mid outdoor Hereflow inflight calibration lane switch, pitch high oscillation and crash. same folder.

82371 08/03/2023 10:34:04.468 MSG 294783423 FlowCal: x:100% y:72%
82645 08/03/2023 10:34:04.870 MSG 295185357 EKF3 lane switch 0
82648 08/03/2023 10:34:04.870 MSG 295185516 EKF primary changed:0
82790 08/03/2023 10:34:05.073 MSG 295388214 EKF3 lane switch 2
82793 08/03/2023 10:34:05.073 MSG 295388312 EKF primary changed:2
84405 08/03/2023 10:34:07.473 MSG 297788332 FlowCal: x:100% y:98%

4.3.2 firmware

mid outdoor Hereflow inflight calibration lane switch, pitch high oscillation, and crash. same folder Hereflow_cal_lane_sw_2.

drones are all autotuned.

144892 09/03/2023 14:37:09.633 MSG 356741046 FlowCal: x:84% y:30%
145095 09/03/2023 14:37:09.936 MSG 357044137 CAN[124] lpwork_thread : 612 free
145096 09/03/2023 14:37:09.936 MSG 357044532 CAN[124] idle : 80 free
145097 09/03/2023 14:37:09.937 MSG 357045074 CAN[124] i2c2_thread : 72 free
145098 09/03/2023 14:37:09.938 MSG 357046058 CAN[124] spi3_thread : 80 free
145099 09/03/2023 14:37:09.938 MSG 357046270 CAN[124] can_thread : 72 free
145100 09/03/2023 14:37:09.938 MSG 357046774 CAN[124] exceptions : 64 free
146906 09/03/2023 14:37:12.637 MSG 359745716 FlowCal: x:84% y:30%
148925 09/03/2023 14:37:15.642 MSG 362750970 FlowCal: x:100% y:30%
150933 09/03/2023 14:37:18.643 MSG 365751151 FlowCal: x:100% y:30%
152929 09/03/2023 14:37:21.618 MSG 368726028 FlowCal: samples collected
152930 09/03/2023 14:37:21.618 MSG 368726091 FlowCal: no better scalarx:0.950 (fit:0.004 > orig
152931 09/03/2023 14:37:21.618 MSG 368726130 FlowCal: no better scalary:0.991 (fit:0.003 > orig
154195 09/03/2023 14:37:23.457 MSG 370565683 EKF3 lane switch 2
154198 09/03/2023 14:37:23.457 MSG 370565842 EKF primary changed:2
154280 09/03/2023 14:37:23.615 MSG 370723372 EKF3 lane switch 0
154283 09/03/2023 14:37:23.615 MSG 370723463 EKF primary changed:0
156009 09/03/2023 14:37:26.220 MSG 373328628 FlowCal: Stopped
160888 09/03/2023 14:37:33.722 MSG 380830562 EKF primary changed:1

4.3.3 firmware.

You have some high levels of vibration and clipping which on all IMU’s which will trigger the IMU changes.

that was caused by the crash or pitch oscillation.

My apologies - I had not realised there was a crash and should have looked a bit more closely.

The vibrations are creeping up before the first EKF3 lane switch but I’ve seen higher levels without any lane switches so I’m not sure it is the cause of the lane switching.

One thing I have just noticed is the high motor PWM values…especially just before when things start to get out of hand where they max out…around the moment the first lane switch occurs.

When inflight optical flow calibrate, the intention is to roll fast and pitch fast to create the “rock” in the air and I have done that a few times before on this drone design and other drones. It started the high pitch oscillation upon first pilot pitch fast forward.

All inflight optical flow calibrate are done after autotune.

Inflight calibrate using 4.3.x first time, this problem is repeatable for pitch, roll was perfectly executed.