Drone crash after reassembly

Our drone was previously functioning well—we even completed a stable autonomous flight (not perfectly tuned, but sufficient for our needs). After a crash, we reassembled it exactly as before (same components, same configuration) only replaced the Pixhawk Cube Orange carrier board (kept the original Cube flight controller). and properly recalibrated both the GPS and accelerometer.

The first problem occurred when attempting takeoff in Loiter mode. The drone flipped and crashed immediately. The logs from this incident showed several concerning errors:
2025-06-08 16:08:38.62: EKF3 IMU1 emergency yaw reset
2025-06-08 16:08:38.62: EKF3 lane switch 1
2025-06-08 16:08:38.62: EKF primary changed:1
2025-06-08 16:08:38.63: EKF3 IMU0 emergency yaw reset
2025-06-08 16:08:40.11: Crash: Disarming: AngErr=166>30, Accel=0.2<3.0
2025-06-08 16:08:40.11: EKF primary changed:0
2025-06-08 16:08:44.52: PreArm: Compass not healthy
2025-06-08 16:08:44.52: PreArm: GPS 1: Bad fix

Initially we thought this might be a GPS or compass calibration issue. The next day we replaced the GPS with a new one and recalibrated it. We tested in Stabilize mode and it flew normally - takeoff and basic flight control worked fine. We even switched to AltHold during flight, though we noticed it was difficult to control (we had planned to retune it later).

However, when we later attempted takeoff directly in AltHold mode, the drone immediately flipped upside down and crashed (video attached). This seems related to the previous Loiter mode failure, suggesting a deeper issue than just a one-time glitch.

We’re including the following in the attached google link drive:

  • Flight logs from the initial Loiter crash
  • Logs from the Stabilize flight and mid-air AltHold switch
  • Logs from the AltHold crash
  • Video footage of both crashes and the successful Stabilize flight

Any ideas what’s going wrong? we’re not sure where to focus. Any help would be greatly appreciated!

Why don’t you tune it now? You have some PID parameters at default, others appear to be fiddled with. The notch filter isn’t set. I can’t believe this craft actually ever flew well with this level of tune.

we didnt tune it because it worked before with these pid parameters we just want to know what could possible be the issue it flips like that directly after it takes off. We were planning to tune it after we test the althold mode

It makes no sense to have the Roll Rate PID’s at default and the Pitch PID’s at a fraction of that. This looks like some poor incomplete attempt at tuning making this craft unstable. Those Pitch P&I values are too low for a craft this size.

Reset to default and start over. If you are inexperienced at Arducopter tuning use The Methodic Configurator

1 Like

Thank you for the advice — I’ll reset everything to default and start over as you suggested. I’ve also attached a short part of the video showing the drone flying autonomously for about 5km.

We’ll go through the setup from scratch and focus on proper PID tuning this time. That said, we’re also trying to understand why the drone flipped like that multiple times — could it be purely due to the PID values, or does anything else in the logs suggest a deeper issue? Also, any advice on how to spot critical errors or indicators in the logs that could help guide our debugging process would be greatly appreciated. We just want to make sure we’re not missing anything major — whether it’s a hardware fault, configuration mistake, or something else.

Any guidance would be really helpful, especially since we’re working under a very tight timeline.