Mavic Crashed into a River

[Selous Scout]

Hey All,

My Mavic crashed into a river a few weeks ago. I reviewed my flight through the DJI Go 4 app and it seems the Mavic descended on its own. I did not see any control inputs on my part that would have made it go down.

I opened a ticket with DJI. They responded today and said it was my fault. Could anyone review the log and tell me what they see? Thanks for your help.

 

[Robbyg]

Sorry, but from the data it seems that DJI is correct. It shows you had full down throttle when it crashed.



Rob

 

[Selous Scout]

Thanks for your help. That's interesting since the app doesn't show any down throttle. Only left turn.

 

[msinger]

I opened a ticket with DJI. They responded today and said it was my fault.

What did they say caused the crash?

That's interesting since the app doesn't show any down throttle.

You're correct -- that's not what your flight log shows.

At the end of your flight, you were flying at about 10 feet above the surface of the river. As you continued to slowly follow it, the ground elevation was increasing. You eventually flew your Mavic into the surface of the water.

You can view your flight log online here:
DJI Flight Log Viewer - PhantomHelp.com

The "VPS Altitude" column shows how close the water was below your Mavic.

 

[BudWalker]

What did they say caused the crash?


You're correct -- that's not what your flight log shows.

At the end of your flight, you were flying at about 10 feet above the surface of the river. As you continued to slowly follow it, the ground elevation was increasing. You eventually flew your Mavic into the surface of the water.

You can view your flight log online here:
DJI Flight Log Viewer - PhantomHelp.com

The "VPS Altitude" column shows how close the water was below your Mavic.

View attachment 15916

This is kinda puzzling though. In those last 3 secs the Mavic appears to have moved about 2 meters. If the surface of the water rises 2.8 meters in that distance that's some pretty healthy waterfalls. Actually I looked on Google Earth and the river surface and the surrounding terrain appears to be flat. There are trees on the banks though.

Also the IMU height appears to be dropping as well. I gotta go with the Mavic was actually dropping in this interval as opposed to the surface rising to meet it. Don't have a clue why that happened.

 

[Selous Scout]

Thanks for the responses everyone. Yes, very puzzling. Here is the response from DJI:

"For your claimed case CAS-702357-P5C3F4, we have already finished the analysis, and the result is as follows:
1. The aircraft was responsive to stick commands during the flight;
2. T=16-31, h=0m, d=99.3m, flight record ended when pilot pushed elevator forward; the aircraft descended at a speed of 0.2m/s
According to the analysis, the incident was not caused by any product malfunction factors. As such, we could not provide warranty service.
Should you have more questions, please feel free to let me know."

This just makes me scratch my head. My cached video and playback (and my memory) on the app don't jive with their analysis and the ones provided here by you guys. I can post up the video later.

Budwalker, correct there were no waterfalls. I actually compensated for the slope of the river as I was flying upstream by increasing altitude.

I'm hesitant to go in the river and search because it's cold and fast moving from snowmelt.

 

[msinger]

I looked on Google Earth and the river surface and the surrounding terrain appears to be flat.

Here are some of the elevations (in feet) according to Google Earth:

the IMU height appears to be dropping as well

The IMU data is odd. In the last 6.4 seconds of the log, the IMU altitude went from 2 feet --> 1.6 feet --> 4.3 feet --> 0.3 feet. During that same time frame, the VPS altitude was gradually decreasing from 11.8 feet --> 0 feet. For that reason, I was ignoring that data since it doesn't seem to be accurate.

 

[Robbyg]

msinger I am a bit confused about what I am seeing on Phantomhelp.

The Graph I linked shows that he held the down elevator and descended down from 50M to 1M or less off the waters surface. As you can see by the graph he continued flying for another 76 seconds until he hit into the water. I even see a small touch on the throttle to go even lower about 24 seconds before impact.

I see you are using the VPS height as a gauge but isn't that totally unreliable when flying over water?

I think the assumption by DJI is that the pilot flew into the water which is what I said and you have also said. My confusion is that you do not see the final touch on the throttle and I also think that his height above the water was a lot closer than 10ft. for those last 24 seconds.

Rob

 

[BudWalker]

Here are some of the elevations (in feet) according to Google Earth:

View attachment 15942



The IMU data is odd. In the last 6.4 seconds of the log, the IMU altitude went from 2 feet --> 1.6 feet --> 4.3 feet --> 0.3 feet. During that same time frame, the VPS altitude was gradually decreasing from 11.8 feet --> 0 feet. For that reason, I was ignoring that data since it doesn't seem to be accurate.

Those two meters look like this

Using your Google Earth image they look something close to this.

 

[Robbyg]

Ok when was Budwalker here
A lot seemed to have happened while I was typing up my post.

Ok never mind he was typing while I was typing.

 

[msinger]

The Graph I linked shows that he held the down elevator and descended down from 50M to 1M or less off the waters surface.

Right -- according to the IMU altitude data. However, it does not show the OP had full down throttle when the Mavic crashed.

I see you are using the VPS height as a gauge but isn't that totally unreliable when flying over water?

It can be totally unreliable over water. There really is no guarantee.

 

[msinger]

Those two meters look like this
View attachment 15943

I'm thinking the Mavic was sinking in the water for the last couple seconds of the log. If you look at the speed or distance from the home point, you can see it stopped making forward progress.

Perhaps it'll make more sense when the OP posts the video.

 

[BudWalker]

Thanks for the responses everyone. Yes, very puzzling. Here is the response from DJI:

"For your claimed case CAS-702357-P5C3F4, we have already finished the analysis, and the result is as follows:
1. The aircraft was responsive to stick commands during the flight;
2. T=16-31, h=0m, d=99.3m, flight record ended when pilot pushed elevator forward; the aircraft descended at a speed of 0.2m/s
According to the analysis, the incident was not caused by any product malfunction factors. As such, we could not provide warranty service.
Should you have more questions, please feel free to let me know."

This just makes me scratch my head. My cached video and playback (and my memory) on the app don't jive with their analysis and the ones provided here by you guys. I can post up the video later.

Budwalker, correct there were no waterfalls. I actually compensated for the slope of the river as I was flying upstream by increasing altitude.

I'm hesitant to go in the river and search because it's cold and fast moving from snowmelt.

I'm gonna speculate that the short (.67 secs) full elevator caused it to sink when the elevator was released

I know elevator should only affect speed but I've seen my Mavic sink when elevator is abruptly removed

 

[BudWalker]

I'm thinking the Mavic was sinking in the water for the last couple seconds of the log. If you look at the speed or distance from the home point, you can see it stopped making forward progress.

Perhaps it'll make more sense when the OP posts the video.

But, the distance travelled continues to increase in the last 3 seconds. Distance travelled is computed from successive gps updates I believe.


Also, distance from HP increase during that interval

 

[msinger]

But, the distance travelled continues to increase in the last 3 seconds

According to the GPS coordinates, the Mavic moved 3 feet during that time frame. I don't think that's an unreasonable outcome after an object hits the surface of moving water and sinks.

 

[Robbyg]

Wow who ever knew this would get so complicated! It's fun from the analytic's point of view but of course sucks for the pilot involved.

I think the bottom line was that he got way too close to the water and went in.

Rob

 

[Robbyg]

Right -- according to the IMU altitude data. However, it does not show the OP had full down throttle when the Mavic crashed.


It can be totally unreliable over water. There really is no guarantee.

Thats why I think he was a lot lower than 10ft. I am dying for him to upload the video so we can see what really happened.

Rob

 

[msinger]

Thats why I think he was a lot lower than 10ft.

Me too. I think the VPS altitude was off by 8-9 feet.

 

[Selous Scout]

I'm guessing I was ~6 feet up before the final decent with no forward motion.

 

[Robbyg]

Me too. I think the VPS altitude was off by 8-9 feet.

Skimming the water at 1-2 FT is pretty risky. I have never had the Jewels to try a maneuver like that, I know the altitude on the Mavic can vary too much for that.

Rob