Uncommanded yaw

[dennyc39]

This afternoon I was puzzled by some strange behavior, where the Mavic was not following my yaw commands. It has had some 150 flights with no violent crashes, and has seemed quite reliable. But in a brief flight this afternoon, there were two points where it was not responding well to yaw inputs (left stick). It was never more than 100 yards away horizontally, and 200-400 ft vertically, and was in a hover in both instances.

The flight record shows incidents where it was yawing more or less randomly, with no stick input. There were other times when it ignored stick input, once when it yawed the wrong way, and similar behavior. Eventually, I activated RTH and it returned. When I executed a manual landing, it behaved as expected.

I check IMU and compass sensors before every flight, and have always found them in the green at the launch point. The only thing unusual about this flight is that I launched twice without ever shutting it down. The first flight was for some photos, and it flew normally at this time. I then brought it inside and set it on a card table to review some recent photos, then went back out to the deck and launched again. The only time I checked sensors was before the first takeoff, but at that time they were good, as always.

Does this sound like compass error or RC calibration?

Was I stupid to engage RTH automation when I could see that something wasn't right?

 

[Lake_Flyer]

Was I stupid to engage RTH automation when I could see that something wasn't right?

I wouldn't call it stupid but I think generally it is better never to issue RTH, and especially when you suspect a compass error. But since you're experienced enough with 150 flights I assume you would not have a problem taking the right action if RTH wouldn't have worked properly.

 

[RobinP]

I would do a RC calibration.

 

[sar104]

This afternoon I was puzzled by some strange behavior, where the Mavic was not following my yaw commands. It has had some 150 flights with no violent crashes, and has seemed quite reliable. But in a brief flight this afternoon, there were two points where it was not responding well to yaw inputs (left stick). It was never more than 100 yards away horizontally, and 200-400 ft vertically, and was in a hover in both instances.

The flight record shows incidents where it was yawing more or less randomly, with no stick input. There were other times when it ignored stick input, once when it yawed the wrong way, and similar behavior. Eventually, I activated RTH and it returned. When I executed a manual landing, it behaved as expected.

I check IMU and compass sensors before every flight, and have always found them in the green at the launch point. The only thing unusual about this flight is that I launched twice without ever shutting it down. The first flight was for some photos, and it flew normally at this time. I then brought it inside and set it on a card table to review some recent photos, then went back out to the deck and launched again. The only time I checked sensors was before the first takeoff, but at that time they were good, as always.

Does this sound like compass error or RC calibration?

Was I stupid to engage RTH automation when I could see that something wasn't right?

There's no way to diagnose the problem just based on that description. If you want to look at the mobile device DAT file for the flight then you can compare the rudder input, IMU yaw and magnetometer data to see what, exactly was happening.