I have been practicing Auto-mission using Pixhawk, Arducopter firmware and mission planner from quite some time. In the last flight, however, in an attempt to check the failsafe, it was observed that the quad started gaining altitude even after entering the RTL and later LAND mode. The quad started drifting away from the launch site ultimately leading to a crash. A discrepancy was observed in the value of RC3 in and RC3 out channel values also. The logs made on autopilot happen o have stopped abruptly even before the actual crash occurred. The quad was recovered from a place around 300m away from the place where the last entry in the log was found.
We had already performed alt-hold, auto-flight, landing, and RTL checks the previous night and everything were working perfectly fine. The quad automatically went to RTL and landed perfectly when the transmitter was turned off.
My project deadline is approaching soon and I cannot risk another crash.
I am uploading the autopilot log too.
I have no idea what might have caused this crash.
In cases where the copter climbs out of control it is due to Vibrations. That is the case in your flight as well.
The thing about vibrations is that wind at the right direction can increase them or sudden changes in motor speed.
Since the unit was bad to start they increase enough to become uncontrollable.
Also turning the radio off during a failure is not going to shutdown the copter.
You should switch to Stabilize mode and reduce the throttle.
Since the unit flew towards people it could have ended badly.
Land and RTL only work if all the sensors are working correctly.
Z vibrations were way too high now-and-then during the flight, looks like something is unbalanced/loose and starts to resonate.
The last 12 seconds have absolutely crazy vibrations , plot IMU.AccX,Y,Z , most likely something fell apart midair, a prop fragmented, motor came loose etc..
go for higher machanical quality (when building)
Did the same thing eventually. Used new balanced props , isolated the autopilot using a 2 tier anti-vibration mount and tightened the loose arms . Carried out a few tests and the z vibrations are gone . Haven't faced the same problem after that incident. Check out the anti-vibration mount I made for the autopilot .Used anti-vibration foam along with vibration isolation balls.Total Vibration killer XD