Mavic 2 Zoom - RTH, then she's down.

[Montana_Flyer]

Not sure what happened, but I was flying my M2Z over a huge (about a square mile) pasture across from my house as I have done many times before, and had just switched from sport to position mode to make a smooth ride home, then the battery got to 25% and indicated that it will RTH in 10 seconds. During the countdown I repositioned myself preparing for the RTH adventure, then the video went a little shaky, followed by a video freeze, then the the announcement "Landing". I obviously got concerned, so I quickly pressed the pause button on my smart controller to stop any autonomous activities, but no joy, everything seemed to be disconnected and unresponsive, and no video feed. After about a minute, I realized that I would have to go on foot and cross the cow field (about 3500 ft) and try to locate the drone.I kept the controller on and followed the GPS to the last known location. When I got about 300 ft away, the signal came back, showing it was at about 5900 ft elevation, and about 50 ft closer to home than the last known location. I finally spotted it nose down in the soft turf (just missed a cow pattie), front end demolished, but still powered on.

I have been flying my original zoom for over 2 years with no issues, but one day in late February I took it out of the case to cycle the batteries (I don't get much of a chance to fly in the winter in Montana) but got a 200 error and could not get it fly ready. I opened a case with DJI refresh+ (last year of eligibility, and only about a month before it expires). They did the replacement option ($149) and sent me a shiny replacement. Lo and behold, it was DOA with a gimble overload error. After another month of arguing, they honored another replacement, this time without the replacement fee, but now beyond refresh deadline. I have had if for a little more than a month, flying it a few times, and now this happens. Now that I am outside of any refresh options, I am hoping that since it was just put into my possession that DJI will see this as a defect, replacing it again at no fee.

Not really sure what happened, but the flight log from my smart controller showed RTH initiated, preascending to RTH height (186 ft), adjusting heading, then LANDING, all with 25% battery and only 3500 ft from home, no wind, beautiful sunset evening, clear LOS the whole time. The video ended during the RTH climb, and did not reveal any of the events that killed my zoom. Log showed 16-17 satellites at the time of the mishap, but the log also showed that it switched from ATTI-GPS to ATTI during that last milliseconds of the flight before I lost it. I could not retrieve the log from the drone, so no data there. Here are the significant entries in the controller log as the events turned (full log attached):

Time	Flight Mode	GPS	IMU Altitude	VPS Altitude	Speed	Home Distance	Battery	Battery Voltage	Cell 1	Cell 2	Cell 3	Cell 4	Cell Deviation	Message

11m 5.6s

Go Home

17satellites

3.6ft

N/A

22.1mph

3,571.8ft

30%

14.653V

3.658V

3.669V

3.674V

3.652V

0.022V

Low power. Returning home.

11m 15.6s

Go Home

17satellites

3.0ft

33.5ft

0.2mph

3,553.5ft

29%

14.561V

3.636V

3.648V

3.65V

3.627V

0.023V

RTH: Preascending

11m 25.1s

Go Home

17satellites

64.0ft

N/A

2.9mph

3,544.0ft

28%

14.51V

3.621V

3.633V

3.643V

3.613V

0.03V

RTH: Heading alignment

11m 25.7s

Go Home

17satellites

64.3ft

N/A

2.0mph

3,545.3ft

28%

14.508V

3.621V

3.633V

3.64V

3.614V

0.026V

RTH: Ascending to RTH altitude

11m 37.3s

Auto Landing

17satellites

167.3ft

N/A

0.7mph

3,547.3ft

27%

14.403V

3.596V

3.607V

3.615V

3.585V

0.03V

Obstacle Avoidance will be disabled in landing.

11m 38.6s

Auto Landing

17satellites

174.5ft

N/A

0.2mph

3,546.9ft

27%

14.4V

3.596V

3.609V

3.611V

3.584V

0.027V

Weak signal. Avoid blocking...

11m 55.1s

P-GPS

14satellites

97.8ft

N/A

19.8mph

3,241.7ft

25%

14.329V

3.584V

3.588V

3.598V

3.559V

0.039V

The remaining battery is only enough for RTH. Return home now.

11m 55.4s

ATTI

16satellites

80.4ft

N/A

16.6mph

3,241.0ft

25%

14.338V

3.585V

3.604V

3.593V

3.556V

0.048V

Other (8)

11m 56.3s

ATTI

16satellites

18.0ft

N/A

5.9mph

3,242.2ft

25%

14.499V

3.62V

3.638V

3.631V

3.61V

0.028V

Weak GPS signal. Aircraft is in Attitude mode and hovering may be unstable. Fly with caution.; Other (8)

That was the last significant log entry on my smart controller prior to losing it, until I found it and it logging started again. The following entry was as I carried it home (dodging cows and patties the whole way):

25m 28.3s

Auto Landing

17satellites

3,943.6ft

N/A

10.6mph

3,236.0ft

25%

14.499V

3.62V

3.638V

3.631V

3.61V

0.028V

Maximum Flight Altitude Reached; Downlink Restored (after 13m 32s)

Has anyone else had issues shortly after RTH initiated? I Also have a M2P that I am now afraid to fly until I get an explanation from DJI. Good news is that State Farm is ready to reimburse me, but I am holding out for DJI's assessment and conclusion. I think that they just sent me another dud from the refresh pool and they should replace it (3rd time a charm?).

Thoughts anyone?

</SS>

 

[Meta4]

Your attached .txt file isn't able to be read with any of the usual programs

 

[Montana_Flyer]

Apologies, that was the CSV version I use for my database. Here is the original...

 

[Meta4]

Not really sure what happened, but the flight log from my smart controller showed RTH initiated, preascending to RTH height (186 ft), adjusting heading, then LANDING, all with 25% battery and only 3500 ft from home, no wind, beautiful sunset evening, clear LOS the whole time.

Here's what your flight data looks like: DJI Flight Log Viewer - PhantomHelp.com

Log showed 16-17 satellites at the time of the mishap, but the log also showed that it switched from ATTI-GPS to ATTI during that last milliseconds of the flight before I lost it.

Your flight data shows you cancelling RTH/Autolanding at 11:43.9 and resuming control.
At 11:52.4 your drone was rocked by something and rolled upside down, started tumbling and losing height.
It lost signal and crashed into the ground just after 11:56.3.
The message indicating the drone had lost GPS and was in atti mode, was because the drone was upside down, spinning and tumbling.
All other flight data after the crash impact is garbage due to the imact affecting the IMU sensors.

Has anyone else had issues shortly after RTH initiated? I Also have a M2P that I am now afraid to fly until I get an explanation from DJI.

Your incident has nothing to do with RTH, which operated normally and was cancelled well before the incident.

 

[Montana_Flyer]

Let me get this straight, 11:37.2 it is going home @ 167.3 it goes from "Go Home" to "auto-landing" in 1/10 of a second @ 167 ft., then @ 11:52.4 it was "rocked by something", still over 160 ft in the sky where there is nothing there in the path over 20ft tall (a single bush that I was well clear of). Shot out of the sky???

Also, what cancelled RTH? If it was cancelled by the RC wouldn't it just hover instead of instantly going into a hyper autoland?

11m 37.2s

Go Home

17satellites

166.3ft

N/A

0.7mph

3,547.3ft

27%

14.403V

3.596V

3.607V

3.615V

3.585V

0.03V

11m 37.3s

Auto Landing

17satellites

167.3ft

N/A

0.7mph

3,547.3ft

27%

14.403V

3.596V

3.607V

3.615V

3.585V

0.03V

11m 52.3s

P-GPS

17satellites

168.6ft

N/A

28.2mph

3,327.1ft

25%

14.14V

3.532V

3.546V

3.549V

3.513V

0.036V

11m 52.4s

P-GPS

17satellites

169.0ft

N/A

27.8mph

3,323.0ft

25%

14.14V

3.532V

3.546V

3.549V

3.513V

0.036V

11m 52.5s

P-GPS

17satellites

168.6ft

N/A

27.7mph

3,319.0ft

25%

14.14V

3.532V

3.546V

3.549V

3.513V

0.036V

Are you seeing that info in the verbose data? I don't see anything different between 11:52.3 and 11:52.5 in the brief version on PhantomHelp that indicates being rocked by something.

I appreciate your help, but need to understand what happened.

 

[Meta4]

I'll look some more when I get home in an hour

 

[c64-dev]

Meta4 is absolutely right. I think I know what caused the crash, but the whole incident seems to be a bit more complicated than first appearances.

First of all, the flight ground did not seem to be level. At some point where the IMU altitude indicates 0 ft, the VPS system kicks in and reads at about 30-31 ft AGL. This could very well be that the ground at that area was lower that the Home point alt, thus the disparity.

After the drone enters RTH at 11:05.6 it starts slowing down autonomously changing trajectory to avoid a tree on its path (11:07 mark).

It goes into Ascend to RTH mode but this is where things get weird.
The drone starts going up normally reaching an indicated altitude of 166.3 ft at 11:37.2 but right at that moment the drone goes into autolanding mode while still ascending. The drone keeps claiming to be autolanding but the altitude shows that it is still ascending, all the way to 172.6 ft at 11:43.6, where the pilot cancels the RTH and takes over manual control of the aircraft.

However, the next bit is even more interesting though. With the aircraft oriented properly, the pilot very nicely goes full stick forward flying back home. The drone flies perfectly level (as indicated by the virtual horizon on the log viewer at bottom right) until the 11:52.4 mark where you can clearly see the virtual horizon becoming badly skewed, indicating a hard tilt of the aircraft to the left.

Keeping an eye on the virtual horizon, look at what happens next: the drone tils completely to the left going completely sideways (11:52.5) and then actually tumbles over (11:52.7), and from that point starts tumbling over and over repeatedly (I counted over 10 tumbles) until it grinds to a halt at 11:56.

It is clear that the drone collided with something badly enough to cause it to tumble so violently over and over like this, but what? I was looking at the map and then it hit me: at 11:52.4, exactly at the point of collision the drone goes over what appears to be a small hill. In my opinion, I think that despite the indicated altitude the drone couldn't clear the top of the hill and hit the ground full speed, which resulted in its tumbling down the side of the hill.


@Montana_Flyer, did you have the VPS sensors disabled while flying in P-GPS mode? Can you check on the Go app?

 

[Meta4]

Let me get this straight, 11:37.2 it is going home @ 167.3 it goes from "Go Home" to "auto-landing" in 1/10 of a second @ 167 ft., then @ 11:52.4 it was "rocked by something", still over 160 ft in the sky where there is nothing there in the path over 20ft tall (a single bush that I was well clear of). Shot out of the sky???

That's what the data shows.

Also, what cancelled RTH? If it was cancelled by the RC wouldn't it just hover instead of instantly going into a hyper autoland?

The reason for this isn't obvious.
Did you push the Land button by mistake?

I don't see anything different between 11:52.3 and 11:52.5 in the brief version on PhantomHelp that indicates being rocked by something.

You have to look deeper into the data.
At 11:52.4 the pitch, roll and yaw data shows the drone being rocked violently, as if it collided with something.
Here's some of the data from that time:

The line highlighted in yellow shows when the problem occurred.
Note the sudden drastic changes in the pitch and roll angles.
The numbers below that highlighted in green show tumbling.

If it wasn't a collision, it was loss of the front left motor or propeller that caused the drone to fall.
If it was a collision, it would have to have been a bird or a high powerline (and I don't see any powerlines in Google Earth imagery).

The ground you flew over was almost flat (there are no hills) and the drone crashed down from about 170 ft.

 

[MavicAir2Marc]

That's what the data shows.

The reason for this isn't obvious.
Did you push the Land button by mistake?

You have to look deeper into the data.
At 11:52.4 the pitch, roll and yaw data shows the drone being rocked violently, as if it collided with something.
Here's some of the data from that time:

The line highlighted in yellow shows when the problem occurred.
Note the sudden drastic changes in the pitch and roll angles.
The numbers below that highlighted in green show tumbling.

If it wasn't a collision, it was loss of the front left motor or propeller that caused the drone to fall.
If it was a collision, it would have to have been a bird or a high powerline (and I don't see any powerlines in Google Earth imagery).

The ground you flew over was almost flat (there are no hills) and the drone crashed down from about 170 ft.

Hey @Meta4 ,
If he would have lost a prop or a motor, wouldn't that show as a warning or error in the data?

 

[Meta4]

Hey @Meta4 ,
If he would have lost a prop or a motor, wouldn't that show as a warning or error in the data?

Not necessarily ... it often doesn't.

 

[Montana_Flyer]

Odd indeed. Thanks everyone for your valuable input. I have posted the video of the last minute or so of the flight(LastMinuteOfFlight 8k.mp4) to google drive just before the crash including the terrain (no, there were no hills or anything in the path). Odd thing is that the video ends without anything that I see as violent, it just stops. I can't remember if I stopped recording, which I sometimes do for RTH. The tilting is likely due to having the gimbal in FPV mode. The last frame shows the heading directly toward my house, and I overlaid the location of recovery and home point(which is direct center on the other edge of the pasture). It seems to have been still ascending at the time the video stopped.

I also have a picture of the drone after bringing it home, no feathers or anything suggesting a collision with anything biological. Also, there was no gunshot at the time, as I would have heard it (a frequent sound living here).

@Montana_Flyer[/USER], did you have the VPS sensors disabled while flying in P-GPS mode? Can you check on the Go app?

No, VPS sensors were enabled, excepting the downward ones so that I can hand catch it at landing if needed.

It is clear that the drone collided with something badly enough to cause it to tumble so violently over and over like this, but what? I was looking at the map and then it hit me: at 11:52.4, exactly at the point of collision the drone goes over what appears to be a small hill. In my opinion, I think that despite the indicated altitude the drone couldn't clear the top of the hill and hit the ground full speed, which resulted in its tumbling down the side of the hill.

There was no hill, just flat cow pasture. See the video.

After the drone enters RTH at 11:05.6 it starts slowing down autonomously changing trajectory to avoid a tree on its path (11:07 mark).

No tree in the path according to the video

So far, I think we have determined that:

the terrain was flat with no obstructions in the path or nearby​

I had clear LOS to aircraft during the incident​

Airdata historical weather shows slight wind, 65 degrees, clear skies​

I was able to cancel return to home as I had attempted​

there was a violent reaction to something at around 170 ft​

the aircraft tumbled as it descended to its final resting ground​

I have sent the drone off to DJI for their evaluation and synced my flight record from the controller, which should be interesting. I guess it is up to the judges @ DJI to make their final verdict. Good thing is that I have insurance coverage as well (State Farm), and I already contacted them and they are willing to settle if DJI comes up short.

Meta4 and Ilmarinen, thank you so much for your objective analysis, REALLY APPRECIATED!!

If anyone else has something to add or discover, please chime in!!

</SS>

 

[Barbara]

Very interesting read. From the photo I see the left front arm is missing. It doesn't look like a break, but rather a clean seperation at the joint. Did you find it? If so, was it at the crash site, or some distance away?

 

[Montana_Flyer]

Very interesting read. From the photo I see the left front arm is missing. It doesn't look like a break, but rather a clean seperation at the joint. Did you find it? If so, was it at the crash site, or some distance away?

I was unable to locate the missing arm, which I would think is highly unusual. I only looked in the immediate area thinking the ground caused the damage, but now that I know something happened @ 170 ft. so I should return to the crime scene and look in a larger area. I have never had a crash like this before, so this is all new to me. Learning the hard way is an effective way to learn (as I've always told my kids), and I have learned a lot. Verbose flight logs, keeping your controller on after the crash to help locate it, etc.

One impact this has had is my willingness (or lack of) to fly my more valuable drones. Since the crash I have limited myself to my Spark and Tello, leaving my Mini 2 and Mavic 2 Pro in the drone cabinet until I understand what happened.

 

[Barbara]

The plot thickens. Not finding the arm at the crash site probably indicates that it separated at altitude. Possibly a prop failure? Resulting in loss of control and excessive vibration. Serious vibration could explain the loss of the arm. Check downwind from the crash site.

Those look like aftermarket props and I've read some stories about questionable quality (balance issues). If it was balance, then perhaps that caused a vibration which over time resulted in the arm separating from the drone. Be interesting to see if you can find the arm whether or not the prop is intact.

Just speculating here, but might explain much.

 

[Montana_Flyer]

Those look like aftermarket props and I've read some stories about questionable quality (balance issues). If it was balance, then perhaps that caused a vibration which over time resulted in the arm separating from the drone. Be interesting to see if you can find the arm whether or not the prop is intact.

Just speculating here, but might explain much.

Yes, these are Master Airscrew low noise. Never had issues before, but I think that I'll switch back to stock just to be safe...

 

[Montana_Flyer]

After more than a month since opening this issue (3 months since this swap-fair began), here is the resulting response:

PROPULSION ISSUE
FLY040

1. Unit was in GPS mode and was responsive to RC commands.
2. At t=11:06s, relative height=1.2m, the goes to RTH mode due to low battery level.
3. At t=11:37s, relative height=52.4m, the unit goes to LANDING mode due to motor overload issue.
4. There was no sign of impact or obstructions at this point.
5. At t=11:52s, relative height=51.5m, the unit flips and fall.
Date of incident: 2021-05-10
Incident GPS coordinates: 46.3375434, -114.1631321
Activation date: 2021-04-09

According to the analysis, the incident was caused by non pilot error. Therefore, it is concluded to warranty.

Thanks all for the feedback. Now that I have my 3rd replacement from DJI, hope this one lasts and stays aloft. Still no explanation as to how the clean separation of the front left arm happened. Guess I'll never know.

</SS>