Fly Away and crash during RTK GPS waypoint mission, curious CTUN.DALT drift

Hello Everyone,

We experienced a flyaway during an RTK GPS waypoint mission. Looking at the log, we see that CTUN_DALT goes down rapidly to a negative value just before the GPS glitch warning is triggered. This caused the drone to rise rapidly until it hit the geofence, disarmed, and fall down.
We use two levels of geofencing, the first fence is set to RTH, and the second one which is 15 meters away is set to disarm. We were flying multiple drones on that mission, and this glitch affected a few of them, during the 10 minutes mission, at different times.

The log file can be downloaded here: https://drive.google.com/file/d/1rGL4Ovdq53_V-UB877zLawhCyNv-9WLA/view?usp=sharing

The above log and graph are typical for the other drones that also failed, we always see a discrepancy between CTUN.DALT and BARO.ALT a few moments before the GPS glitch happens and the drone flies away.

Our drones were calibrated on site, just before the mission.

Any insight into what may have caused this would be greatly appreciated.

Sincerely,

Patrice

There was some clipping in that log, that could be the cause. It’s not showing up directly in the vibrations, the clipping means the IMU has exceeded the limits of what it can measure so it likely wont show in the recorded vibrations.
Lots of people say “I see the clipping but there’s almost no vibrations, so it cant be that”

Hello Shawn and thanks for the comment. How can I see the clipping in the log? On what parameter would it show up?

Thank you

Under the VIBE section in log analysis tools.
Check for the same trend on the other aircraft.

Also try GPS_GNSS_MODE,65 or 67 and see if that improves reliability.

Continuing the discussion from Fly Away and crash during RTK GPS waypoint mission, curious CTUN.DALT drift:

Thank you, I appreciate your input, it’s very helpful.