I armed the quad in stabilize mode, made a little Hop, checked the GPS lock and then switched to LoiterMode.
And this thing goes instantly FULL THROTTLE and does not react to any RC inputs at all…
I tried to switch to RTL than LAND but it did not do anything but GOGOGO HIGHER
After like 20 sek. i lost telemetry Connection.
The last Data says:
878m alt, 75A Current (full throttle), totally Leveled AND: -69,8m/s climb rate (!) does the nagative climbrate means it thinks its falling ?
It’s probably caused by high vibration levels on the copter. If there is a tlog it will show the VIBRATION message and values that are regularly above 30m/s/s are dangerous and levels above 60m/s/s are almost guaranteed to have problems. the “Clipping” number is also a good sign of problems - when an accelerometer clips it means the vibration levels are above the 16G limits of the sensor and the EKF has a hard time determining it’s climb rate.
There are some things coming that will improve the situation:
a new line of Invensense IMUs has a 30G limit (nearly double the 16G limit of the MPU6000) which means the EKF will be able to withstand nearly twice the vibration we currently can before becoming confused
we are working on a vibration failsafe for Copter-3.7 to reduce reliance on the accelerometers if high vibration levels are detected.
Until the above is complete though, the best we can say:
ensure the flight controller is properly isolated from vibration as described here and here on the wiki or buy a flight controller with built in vibration isolation like the Hex Cube or CUAVv3.
keep your eye on the VIBE message on the ground station HUD, be ready to re-take control in stabilize mode if it goes red or the vehicle loses altitude control
Thank you all for the Input.
Today I will try to find my Copter again… But this will be hard… i lost connection in almost 1km alt.
But i will show you the tlog when I find it!
i had very similar experience just yesterday on Matek f405 wing mounted on velcros. (AC 3.6.8)
It was in loiter mode for at least 15 minutes prior to that. Then, when i started to push it harder (sharper and more abrupt turns flying quite fast) it just took off at a slight angle.
Once the copter took off, it stopped reacting to yaw, throttle, roll or pitch.
It did react to my arm/disarm switch so luckily it did not fly that high before i made it crash.
unfortunately the log file shows size 0.
its the 3rd time it happened for me on this matek based copter. Never had anything like that with the Cube.
unfortunately with modes on a 6 position pot i usually don’t have time to switch it back to stab under the pressure of the situation. it is much easier for me to disarm.
I had a similar experience last year due to high vibrations and found the only way to bring it back down was to switch it back into stabilize mode, which made it respond to RC inputs. I always have stab mode on the flight modes switch for maiden flights and test flights.
I have had exactly this problem too, and the log showed high vibrations in fast forward flight just before the quad flew to the moon. There’s a Paul Riseborough video which explains this very well. This should be mandatory viewing for all copter pilots!