DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Mavic Air fly-away with no luck to find it. SOLVED

alexbutenkofi

Member
Joined
May 25, 2018
Messages
7
Reactions
2
Age
37
I have never thought that something like this will happen to me but it did.
All began with my idea to try out youtube streaming capabilities of my Mavic Air. This was not my first flight. So...
It all began when I went out and tried to find a good spot for take off.
First place was near one of the local schools but there I was getting a lot of magnetic interference and I've tried to calibrate compass few times with no luck.
I moved about 100 meters from the position but still there was something that was making interference, I moved 100 more. Its needed to mention that I have had a flight at this position already before with no problems.
So I found a good spot, calibrated the compass,good GPS signal was also present(13-18 satelites) the home point was updated also, I took off and slowly went up and aprox after 25 second of my flight my drone became uncontrolled and started to speed up even though I was not planning to fly away to shoot something.
During the present connection I've tried "pause" button few times and than "go home" button. No luck.
I have tried to search the drone with last position in menu "Find my drone", after 1 hour of searching I've ended with no luck.
I hope nobody gets hurt. Since there is my name and phone number I also have not received any calls.
I would appreciate if experience pilots could give me approximate conclusion why did it happened.
Here is a flight log. DJI Flight Log Viewer - PhantomHelp.com
 
Last edited:
  • Like
Reactions: SagaM
Very strange. It couldn't be too far away. Did would walk towards the last known position with your RC on. Was it still connected to your MA? If it was, it should tell you how far away it was, and you could trigger it to beep so you can try to locate it via sound. But it's too late for that now.
A lot of recent fly-aways was due to compass heading for some reason. Will have to wait for @sar104 to have a look. A it looked like there was something wrong with the battery.
 
Very strange. It couldn't be too far away. Did would walk towards the last known position with your RC on. Was it still connected to your MA? If it was, it should tell you how far away it was, and you could trigger it to beep so you can try to locate it via sound. But it's too late for that now.
A lot of recent fly-aways was due to compass heading for some reason. Will have to wait for @sar104 to have a look. A it looked like there was something wrong with the battery.
Thanks for your answer, I have been running around and even tried more far direction where the drone flew, and was trying to get my RC to catch the drone signal but no luck(
 
Did the red arrow in the video correspond with the actual orientation of the drone (i.e. facing south-west on takeoff)?

at 22 sec some vibration and drift started in the pitch and roll without controller input. because the seemingly uncontrolled flight it is hard to predict where it went. lets hope you get a call with some good news
 
  • Like
Reactions: MrsTreat
Did the red arrow in the video correspond with the actual orientation of the drone (i.e. facing south-west on takeoff)?

at 22 sec some vibration and drift started in the pitch and roll without controller input. because the seemingly uncontrolled flight it is hard to predict where it went. lets hope you get a call with some good news
The drone took off with no problems, I mean nothing strange, no errors or else...Like you may see on the flight video it started to speed up with changing altitude up an down and facing wrong flight direction.
 
This looks like yet another example of taking off from a magnetically distorted location. The yaw initialized at -147° (approximately pointing SW). However, after takeoff and initial vertical climb, with the yaw still SW (now -135°), at 25 seconds the applied negative pitch and negative roll will have caused the aircraft to move forwards and left - i.e. approximately south. But it headed north instead, suggesting that it was actually pointing NE, not SW.

2018-05-24_[22-41-26]_01.png


That's a 180° yaw error and will cause exactly this kind of behavior if the aircraft remains in P-GPS mode, because the aircraft motion is directly opposite what the FC expects from the indicated yaw and the corrections that it attempts to fix that motion are directly opposite what is needed. It produces a positive feedback loop (unstable) instead of a negative feedback loop (stable).

Once again - the advice to check that the aircraft orientation arrow correctly reflects the direction that the aircraft is pointing before takeoff is absolutely key here. And stop calibrating the compass all the time - it's not the compass calibration that is the problem - it's a magnetically distorted takeoff location. No amount of compass calibration will fix that.

As for where it ended up - no way to know. It was heading west towards the railway tracks at an altitude that was unlikely to clear the buildings beyond. After losing signal it should have attempted to RTH but with that magnitude of yaw error it had no chance of navigating successfully. It should have dropped into ATTI, at which point, with no signal, it would have attempted to autoland.
 
This looks like yet another example of taking off from a magnetically distorted location. The yaw initialized at -147° (approximately pointing SW). However, after takeoff and initial vertical climb, with the yaw still SW (now -135°), at 25 seconds the applied negative pitch and negative roll will have caused the aircraft to move forwards and left - i.e. approximately south. But it headed north instead, suggesting that it was actually pointing NE, not SW.

View attachment 38710


That's a 180° yaw error and will cause exactly this kind of behavior if the aircraft remains in P-GPS mode, because the aircraft motion is directly opposite what the FC expects from the indicated yaw and the corrections that it attempts to fix that motion are directly opposite what is needed. It produces a positive feedback loop (unstable) instead of a negative feedback loop (stable).

Once again - the advice to check that the aircraft orientation arrow correctly reflects the direction that the aircraft is pointing before takeoff is absolutely key here. And stop calibrating the compass all the time - it's not the compass calibration that is the problem - it's a magnetically distorted takeoff location. No amount of compass calibration will fix that.

As for where it ended up - no way to know. It was heading west towards the railway tracks at an altitude that was unlikely to clear the buildings beyond. After losing signal it should have attempted to RTH but with that magnitude of yaw error it had no chance of navigating successfully. It should have dropped into ATTI, at which point, with no signal, it would have attempted to autoland.

Thank you for your answer and explanation. I have sent this case to DJI, lets see what will be their summary. Meanwhile I will try to research the area and hopefully can find the drone on one of the trees. I hope it did not autoland on the roof.
 
For what it's worth, could it be that if the drone kept on flying off with the same heading, that it flew a circle? Perhaps you've already searched here, but it could be that the drone is located somewhere else you would otherwise expect.

I've roughly continued its trajectory to give you an idea. Possibly hitting the buildings, depending on their height.
 

Attachments

  • sdfsdf.jpg
    sdfsdf.jpg
    167.1 KB · Views: 154
For what it's worth, could it be that if the drone kept on flying off with the same heading, that it flew a circle? Perhaps you've already searched here, but it could be that the drone is located somewhere else you would otherwise expect.

I've roughly continued its trajectory to give you an idea. Possibly hitting the buildings, depending on their height.
Thanks for the advice! One of my ideas is to draw a map of possible uncontrolled flight. You have me another good option to try.
 
  • Like
Reactions: MrRobville
@sar104 since I see your advice come up so often, have you let DJI know of this problem?
Hmm, problem? One definition of problem is:
Denoting or relating to people whose behaviour causes difficulties to themselves and others.”
Not sure DJI can fix pilot error!
 
Hmm, problem? One definition of problem is:
Denoting or relating to people whose behaviour causes difficulties to themselves and others.”
Not sure DJI can fix pilot error!

Regardless of what happened the aircraft should have responded to the pause command. Is the pause shown in the log?
 
And stop calibrating the compass all the time - it's not the compass calibration that is the problem - it's a magnetically distorted takeoff location. No amount of compass calibration will fix that.

The last few times I have flown it won't let me take off without calibrating the compass. This as at a variety of locations.
 
The last few times I have flown it won't let me take off without calibrating the compass. This as at a variety of locations.
Sar’s suspicion is that it wasn’t the lack of compass cal, but that local interference caused the drones heading to be 180 degs out.
 
After having read a number of "fly away" posts and the advice to always check the aircraft's orientation on the app to make sure it is accurate, I always check it now. I did find one flight where the orientation was wrong. I turned everything off and back on and all was well.

My question is, and it's mostly for you guys who have read 100's of these posts, how often do you see "fly away" scenarios where the aircraft was actually pointed in the right direction on the screen before takeoff? Or, do you never see the problem if the aircraft was accurately reporting it's preflight orientation. My reason for asking is, is it fair to assume that if the aircraft is reporting it's orientation before flight correctly, that a "fly away" is an extremely remote possibility?
 
  • Like
Reactions: Buckaboy
After having read a number of "fly away" posts and the advice to always check the aircraft's orientation on the app to make sure it is accurate, I always check it now. I did find one flight where the orientation was wrong. I turned everything off and back on and all was well.

My question is, and it's mostly for you guys who have read 100's of these posts, how often do you see "fly away" scenarios where the aircraft was actually pointed in the right direction on the screen before takeoff? Or, do you never see the problem if the aircraft was accurately reporting it's preflight orientation. My reason for asking is, is it fair to assume that if the aircraft is reporting it's orientation before flight correctly, that a "fly away" is an extremely remote possibility?

The "flyaway" scenarios fall into a couple of categories, but the unstable flight version is the commonest, and they almost always result from large yaw / compass errors that were due to magnetic interference at the takeoff location. The aircraft will always be reporting an incorrect orientation in that situation, and the corollary, that if it is reporting correctly it will not lead to unstable flight, appears to be almost always true. The only exception I can recall recently was an aircraft with a broken z-axis gyro, which is a hardware failure that leads to exactly the same kind of yaw error.
 
Ya I follow the things I see on here regarding orientation, interference and gps lock. Never had a single close call or issue.
 
  • Like
Reactions: Robert Droner
@sar104, you are the perfect person to answer my question as everyone looks to you for flight analysis. You've analyzed as many of these as anyone and probably have the best gauge on it. Thank you for the summary.

I've had an Inspire 2 for about 1.5 years and have had my Mavic Platinum for about 5 months. I'm fascinated by the data you interpret for people all the time and want to dive into it myself. Hopefully, it won't take me too long to start wrapping my head around it.
 
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,599
Messages
1,554,243
Members
159,603
Latest member
refrigasketscanada