Flight control issues with coaxial octorotor

Hi there,
Me and my team are at loss. We are having severe flight control issues with our custom coaxial octorotor.

We used to have smooth flight, but after some modifications, changed parameters, etc., it does not fly smoothly anymore. Honestly, I am not sure which modification exactly that caused the issues, but we have tried resetting all parameters and flashing with new firmware. Currently firmware v.13.3 v5 RTPS.

The issue is that the drone randomly jerks with high roll/pitch rate. Additionally, it sometimes aggressively attempts to change its position without stick input, even though it was fling smoothly moments before. This hinders us in continuing our work, and it has caused a number of crashes, luckily not damaging any expensive equipment.

The two most recent crash logs:
https://review.px4.io/plot_app?log=1262c2d8-ffda-4054-825b-5e3df792e237
https://review.px4.io/plot_app?log=06ba975c-1bf3-4a40-9ef0-f315cb14e377

As seen, the estimated position drifts severely, and the estimated flight distances and max velocities are way off. However, on the shown map, the plotted trajectory is exactly as we actually flew. Can anyone help me decipher this? Because I unfortunately don’t really understand it.

We reflashed the firmware with v1.13.3 and resetted all parameters related to the controller, and performed a flight where the same weird behavior was present but we didn’t crash:
https://review.px4.io/plot_app?log=83af14fd-28f3-4674-95a4-2cb9e89c9325

Here, the jerks are visible from the roll/pitch angles and rates, especially around 5:24.

Could it be something related to the hardware? I mean, could our flight controller be damaged? We are using CUAV v5+.

Might anyone be able to offer us any insights in how we can go about solving these issues?

Hi,
Your FC(PX4_FMU_V5) has overloaded.
Check CPU Load Logger.

Hi, thank you very much for your reply. I see that the CPU load is a little higher than 0.8, not exceeding 0.9. Is this a sufficiently high load to cause the described kind of flight behavior?

Hi,
It has over 0.9 before stop logging.

Ah, yeah, however this was after crashing, so it would not have caused the behavior.