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

Mavic 2 Pro Crashed into the sea without apparent reason

Isaino

New Member
Joined
Sep 9, 2019
Messages
4
Reactions
10
Age
34
Location
France
Hello Everyone,

I was flying my drone today (Mavic 2 Pro) around the coast to film the sunrise and after 5min of flight the drone just started to be uncontrollable for no apparent reason.

When I lost control, the battery was at 80% and the drone had been hovering since 20sec because I was changing the video settings. After another 20sec of uncontrollable movement it fell into the sea close to rocks in a place with a lot of current so I couldn't dive to get it back.

I would like to understand what happened, I have about 15 hours of flight with this drone and I've never seen such a weird behaviour.

Here is the log of the flight: DJI Flight Log Viewer - PhantomHelp.com

Before the take off I had a compas error due to magnetic interference so I did the calibration process. I was next to a big cross (about 5m high) probably made of metal so the calibration failed at the first attempt. I tried to do it again a few meters away of the cross and it worked, no more warning messages so I thought I was good for take off.
Another guy was flying his Mavic Pro 1st Gen at the same time. Could it be some disturbance caused by the signal of the other remote control? I didn't have any warning and the drone's behaviour during the first 5min of flight was ok.
It was very early in the morning and I didn't see any birds around, I watched the video cash of the flight and I can't see any birds so I don't think that something hit the drone but it might be the reason of the crash ?


I'd really like to understand what happen and if I've done something wrong, It's a shame to lose a drone without understanding why...

I would be very grateful for any help !
 
Last edited:
  • Like
Reactions: Van Shteffy
Hello Everyone,

I was flying my drone today (Mavic 2 Pro) around the coast to film the sunrise and after 5min of flight the drone just started to be uncontrollable for no apparent reason.

When I lost control, the battery was at 80% and the drone had been hovering since 20sec because I was changing the video settings. After another 20sec of uncontrollable movement it fell into the sea close to rocks in a place with a lot of current so I couldn't dive to get it back.

I would like to understand what happened, I have about 15 hours of flight with this drone and I've never seen such a weird behaviour.

Here is the log of the flight: DJI Flight Log Viewer - PhantomHelp.com

Before the take off I had a compas error due to magnetic interference so I did the calibration process. I was next to a big cross (about 5m high) probably made of metal so the calibration failed at the first attempt. I tried to do it again a few meters away of the cross and it worked, no more warning messages so I thought I was good for take off.
Another guy was flying his Mavic Pro 1st Gen at the same time. Could it be some disturbance caused by the signal of the other remote control? I didn't have any warning and the drone's behaviour during the first 5min of flight was ok.
It was very early in the morning and I didn't see any birds around, I watched the video cash of the flight and I can't see any birds so I don't think that something hit the drone but it might be the reason of the crash ?


I'd really like to understand what happen and if I've done something wrong, It's a shame to lose a drone without understanding why...

I would be very grateful for any help !


So i only have 1 question did you Restart the Drone and the controller after the compass error and move your position.
If not than this is why your drone went down.

This is one of the reasons why i created the Sea Lander jackets so you test flight an area before going mainstream over the water as this can happen to anyone. Every pier , bridge, metal grate and Cross, lol is a a potential threat when it comes to magetic interference.

The Data should show the Yaw Errors>
 
I'm SUPER paranoid about the compass errors / magnetic interference. As ^^ he says, if you get the error, pull battery and restart somewhere else. If the IMU is messed up, you have no control. It seems that when I've seen the error (e.g., when I launched from near a large metal machine), the error went away after a bit and I think it would have let me take off.
 
Would not a Magnetic Interference just cause the drone to go into ATTI mode? and drift off? I can't see why and have never read how a Magnetic Interference error would cause a drone to just drop out of the sky. That seems more like a dreaded FC crash or battery disconnect. "It fell into the sea" implies that it lost power.

I agree with statements about moving to a new location and calibration of the compass.

If the OP received no errors on take off, then the IMU's should have been OK.

Also remember that the Mavic 2 Pro will ask you for a compass calibration (showing magnetic interference as the error) if you have moved 30 miles or more from your last flight. (At least mine does). The first time this happened, I was shocked by the magnetic interference error as I was on sandstone rock and knew there was nothing magnetic in the area.

Flight log will hopefully show something, and to OP sorry to hear about your loss.

Paul C
 
  • Like
Reactions: Phantomrain.org
Not all of the M2's request calibration every 30 days/30 mile distance.

If you get magnetic interference, don't calibrate. Instead, move away from the interference. Even with a calibration request, it might actually be an interference issue instead.

You may be able to take off where then the error goes away, but then the "damage" has been done where different sensors disagree.

I've said that if there is a compass/IMU disagreement, AC should be able to use GPS to figure it out. It would require the AC lateral movement to get delta GPS coordinates though, and that flight pattern could encounter an obstacle.

There have been reported incidents of the AC getting magnetized, causing chronic compass issues. This can be caused by being stored near a speaker for example.
Degaussing has corrected such situations.
 
I agree not all ask for it. But many do. My point being a lot of pilots believe they have true magnetic interference when it’s just the 30 miles or more issue. Both of my M2 pros ask. They also will show this message about compass issues id it’s been more than 30 days since last compass calibration. Both are current on firmware.

Paul C
 
  • Like
Reactions: crlmjohnson
I was flying my drone today (Mavic 2 Pro) around the coast to film the sunrise and after 5min of flight the drone just started to be uncontrollable for no apparent reason.
When I lost control, the battery was at 80% and the drone had been hovering since 20sec because I was changing the video settings. After another 20sec of uncontrollable movement it fell into the sea close to rocks in a place with a lot of current so I couldn't dive to get it back.
I would like to understand what happened, I have about 15 hours of flight with this drone and I've never seen such a weird behaviour.

I'd really like to understand what happen and if I've done something wrong, It's a shame to lose a drone without understanding why...
All the discussion about compass issues is irrelevant because this incident was not at all compass related.
It also had nothing to do with another drone flying nearby.
Post #7 is pointing to the explanation.
At 5:20.2 it looks like you touched the autoland button and the drone started to descend.
Maybe you thought you were initiating RTH?
You did not cancel the autolanding and the drone descended to the water.
 
They may be irrelevant for the crash but there is still a lot of misunderstanding about the generic magnetic interference error message from the app and what makes that message appear and how severe it is.

Paul C
 
  • Like
Reactions: Phantomrain.org
They may be irrelevant for the crash but there is still a lot of misunderstanding about the generic magnetic interference error message from the app and what makes that message appear and how severe it is.
The OP asked for information regarding a specific issue and posted flight data which showed exactly what the issue was.
In response to that he got 6 posts from people that did not look at the flight data and went on to discuss something that had no impact on the incident.
They may as well have been discussing potential issues flying in strong winds, which can also have serious impacts but are completely irrelevant to this particular incident.

There's a lot of misunderstanding about many aspects of drone flying.
Some is even caused by irrelevant discussion in forums.
 
And there was a lot of miss leading info listed about magnetic interference issues. It’s still possible he had the message due to the fact he moved over 30 miles or it had been 30 days. I realize the drone was auto landed.

No I did not look at the flight data. I answered a question he had in the post about magnetic interference.

Whatever.

Paul C
 
  • Like
Reactions: Van Shteffy
All the discussion about compass issues is irrelevant because this incident was not at all compass related.
It also had nothing to do with another drone flying nearby.
Post #7 is pointing to the explanation.
At 5:20.2 it looks like you touched the autoland button and the drone started to descend.
Maybe you thought you were initiating RTH?
You did not cancel the autolanding and the drone descended to the water.

Thank you all for your answers ! I’m amazed by the reactivity on this forum, I need to come here more often ?

So let me try to add as much info as possible:
- I didn’t restart the drone between the two calibrations but everything looked OK after the second calibration.
- I didn’t have a lot of wind during the flight, I couldn’t even feel a breeze.
- I saw that the auto landing turned on but the drone went out of control at 5:02 at 5:20 the drone was already uncontrollable and slowly falling in spiral. I probably panicked and hit the auto landing when trying to control the drone but nothing was working. You can see on the log that the drone is already having a very strange behaviour whiteout the auto landing activated and without any instruction from the remote control.

Do you think I have a chance to make the DJI warranty work ? Or is it an operator mistake? I don’t understand what I did wrong.
 
I saw that the auto landing turned on but the drone went out of control at 5:02 at 5:20 the drone was already uncontrollable and slowly falling in spiral. I probably panicked and hit the auto landing when trying to control the drone but nothing was working. You can see on the log that the drone is already having a very strange behaviour whiteout the auto landing activated and without any instruction from the remote control.

Do you think I have a chance to make the DJI warranty work ? Or is it an operator mistake? I don’t understand what I did wrong.
OK ... there's something unusual there before the autolanding begins.
From 4:44.9 you left the drone hovering at an altitude of 42 feet above launch height.
At 5:02.8 there is a slight clockwise rotation of the drone.
Although you did not move the joysticks again until 5:05.4, the drone started descending at 5:03.9.
When you gave it full left stick to stop the descent, it was already down to 30 feet and continued descending although you gave it more left stick up.

You also gave some rudder input and right stick which both seem to have worked properly.
This is very mysterious that the drone started descending on its own and applying throttle had no effect.
It appears that the drone was going in regardless of the autolanding.
Getting it back over the nearest land would have been the only chance to prevent the loss.

Whatever the cause was, it's not immediately obvious.
This case looks like a good mystery for @sar104.
He enjoys a good challenge
Do you think I have a chance to make the DJI warranty work ? Or is it an operator mistake? I don’t understand what I did wrong.
There's no indication of user error causing the initial uncommanded descent.
Warranty claim?
Not having the drone to submit can make that difficult but there have been cases where the data was enough to convince DJI to replace even without the drone.

Start by synching your flight data (click the cloud icon on your flight records page in the app) to upload the data to DJI.
Then go to DJI's online help and open a chat to get things started with them.
 
Possibly a warranty claim. I would hazard a guess it might be interference from the other MP's remote. I am usually vary of other drones around. Once I was flying my MP with someone that was flying their MPP. I had my goggles on, and suddenly loss connection. After I asked my friend to land their MPP and turn off their remote, my remote reestablished connection with my MP, and it was RTH.
 
  • Like
Reactions: Isaino
I'm going to disagree with the diagnosis here - this was not an inadvertent autolanding event. The IMU lost control well before that happened, at 303 seconds:

Delta_V.png

The red and green traces represent fusion disagreements in velocity, and should be close to zero as they are for most of the flight. Looking at pitch/roll/yaw in relation to stick commands and height:

comp.png

Note the uncommanded, and out-of-range roll excursion at 303 seconds, with no stick inputs at that time. That corresponds exactly (as expected) with the excessive roll error messages in the log. The aircraft immediately began to lose altitude, and then goes into a descending spin, reaching over 10 m/s descent speed. Autolanding kicked in at 320 seconds, but was not app or RC commanded.

This looks like a warranty issue to me, but the mobile device DAT file (FLY036.DAT) may give significant extra information to support that:

Mobile device DAT file: http://www.datfile.net/DatCon/retrieveV3DatfromTablet.htm
 
I'm going to disagree with the diagnosis here - this was not an inadvertent autolanding event. The IMU lost control well before that happened, at 303 seconds:

View attachment 81447

The red and green traces represent fusion disagreements in velocity, and should be close to zero as they are for most of the flight. Looking at pitch/roll/yaw in relation to stick commands and height:

View attachment 81448

Note the uncommanded, and out-of-range roll excursion at 303 seconds, with no stick inputs at that time. That corresponds exactly (as expected) with the excessive roll error messages in the log. The aircraft immediately began to lose altitude, and then goes into a descending spin, reaching over 10 m/s descent speed. Autolanding kicked in at 320 seconds, but was not app or RC commanded.

This looks like a warranty issue to me, but the mobile device DAT file (FLY036.DAT) may give significant extra information to support that:

Mobile device DAT file: http://www.datfile.net/DatCon/retrieveV3DatfromTablet.htm
Looks to me like a not so typical loss of propulsion incident - either leftFront or rightBack. At 315 secs it started CW rotations but it wasn't until around 321 secs that it looks like it may have started tumbling. Full throttle between 314 secs and 321 secs had a small effect on the altitude.
 
  • Like
Reactions: Isaino and sar104
I'm going to disagree with the diagnosis here - this was not an inadvertent autolanding event. The IMU lost control well before that happened, at 303 seconds:

View attachment 81447

The red and green traces represent fusion disagreements in velocity, and should be close to zero as they are for most of the flight. Looking at pitch/roll/yaw in relation to stick commands and height:

View attachment 81448

Note the uncommanded, and out-of-range roll excursion at 303 seconds, with no stick inputs at that time. That corresponds exactly (as expected) with the excessive roll error messages in the log. The aircraft immediately began to lose altitude, and then goes into a descending spin, reaching over 10 m/s descent speed. Autolanding kicked in at 320 seconds, but was not app or RC commanded.

This looks like a warranty issue to me, but the mobile device DAT file (FLY036.DAT) may give significant extra information to support that:

Mobile device DAT file: http://www.datfile.net/DatCon/retrieveV3DatfromTablet.htm
HI Sar. I read all these crash / failure results and try to remember them all and relate it to my M2P do you think the IMU lost control because it needed calibration or do you think it was something else, I love reading your post's because of all the information in them, Thank you John
 
  • Like
Reactions: Viking buck
HI Sar. I read all these crash / failure results and try to remember them all and relate it to my M2P do you think the IMU lost control because it needed calibration or do you think it was something else, I love reading your post's because of all the information in them, Thank you John

I doubt it was actually an IMU fault - I suspect that @BudWalker is correct that it was a propulsion problem. The DAT file will most likely identify the exact issue.
 
Thank you for your deep analysis @Meta4 & @sar104.

I've attached the DAT file hoping this will bring more information.

@BudWalker thank you for your help, do you know where the propulsion loss could come from ?

I was thinking maybe it could come from the propeller but they were all in perfect condition and I did a 20min flight the day before the crash without facing any issue.

I will try to open a case on the DJI website to see if it can go under warranty, thank you all for your help.
 

Attachments

  • 2019-09-09_05-53-31_FLY036.DAT
    3.7 MB · Views: 20
  • Like
Reactions: Van Shteffy
Hello Everyone,

I was flying my drone today (Mavic 2 Pro) around the coast to film the sunrise and after 5min of flight the drone just started to be uncontrollable for no apparent reason.

When I lost control, the battery was at 80% and the drone had been hovering since 20sec because I was changing the video settings. After another 20sec of uncontrollable movement it fell into the sea close to rocks in a place with a lot of current so I couldn't dive to get it back.

I would like to understand what happened, I have about 15 hours of flight with this drone and I've never seen such a weird behaviour.

Here is the log of the flight: DJI Flight Log Viewer - PhantomHelp.com

Before the take off I had a compas error due to magnetic interference so I did the calibration process. I was next to a big cross (about 5m high) probably made of metal so the calibration failed at the first attempt. I tried to do it again a few meters away of the cross and it worked, no more warning messages so I thought I was good for take off.
Another guy was flying his Mavic Pro 1st Gen at the same time. Could it be some disturbance caused by the signal of the other remote control? I didn't have any warning and the drone's behaviour during the first 5min of flight was ok.
It was very early in the morning and I didn't see any birds around, I watched the video cash of the flight and I can't see any birds so I don't think that something hit the drone but it might be the reason of the crash ?


I'd really like to understand what happen and if I've done something wrong, It's a shame to lose a drone without understanding why...

I would be very grateful for any help !
Isaino, sorry you lost your bird there. It saddens me even more that DJI still havent fixed this problem, after probably 100's now of similar crashes reported on the forums, starting way back in previous Phantom generations! In 2017 I lost my P3pro in almost exact fashion. The app reported IMU and interference issues, then in the logfile it went out of GPS to ATTI mode, and spiralled into the ground at a reported 100ft/s. (I was filming an offroad race 200km from home if I didnt bring my MP1 along as well it would have been a woeful wasted day.) Geographically totally opposite - 1000km inland, very dry country (Botswana), no other drones around etc. IMHO this really is an inherent, latent firmware issue in a part of the AC firmware (navigation?) suite that hasnt been touched over the model generations. I would suggest some very direct questions to DJI Support, and even a warranty claim for what is conceivably a known problem.
 
Last edited:
  • Like
Reactions: awsan5
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,583
Messages
1,554,087
Members
159,586
Latest member
maniac2000