Dji mini 2 se won't land, then crashes

[DJI man]

Hi, I was flying my my mini 2 se when in suddenly wouldn't land. After a while I changed to "normal" mode and I could get it down. When it was about 5 m above ground I was flying it towards me when it suddenly speeded up and crashed in the wall. Anyone who can interpret the flight log, and could see if it was a malfunction?

DJI crash log - Filemail

Click here to view and download these shared files from Filemail.com

www.filemail.com

Regards

 

[Yorkshire_Pud]

Phantomhelp doesn't seem to like the "Data table" of the unzipped version of your log, there is also mention of something relating to "JSON". The scrolling data beneath the map of the replay is empty.
I don't know if this is a PH error or something in the unzipped version of your log itself.
Therefore, can you upload the original version of the .txt flight log to

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

and post the URL here please.

That said, at a glance, the CSV downloaded from PH looks normal.

However, the indicated behaviour during the actual RTH seems odd to me.
The RTH height was set at 30m/98.4ft and the drone reached that around 161 seconds, it then started towards home. You did not give the drone any throttle during the RTH yet, during the trip home, the drone continued to climb, quite steadily, reaching a height of 231ft.

During the flight home the VPS system shows chaotic values, all under 4ft, which I suspect is impossible. But this might be the reason for the climb during the flight home.
I suspect that the VPS value during more or less the whole flight are questionable.

During the descent the descent speed was very slow, under 40cm/s.

Overall, I suspect there is a problem with the VPS sensors, they might be broken or their lenses may need cleaning. Whether or not this explains the failure to land I do not know but when my mini 1 & 2 reject the landing site I think simply holding the throttle closed normally forces the drone to land but I might have to touch any warning showing in the app and accept any risk.
Having said that' the process is so routine I don't really pay attention to what is required and I do not recollect having a problem getting the drone to land.

As for the crash, I suspect you simply commanded the drone to descend until it was low enough to be below the tree tops at that location and it just clipped a tree.
I made a mistake there.

What would I have done in your situation? Probably caught the drone and then flipped it upside down.

Was the drone lost ? If not was it damaged ?

 

[DJI man]

Thank you, I uploaded the file here: DJI Flight Log Viewer - PhantomHelp.com

The drone was a couple of hundred meters away, above the treeline. I then used RTH and it came towards me. I could then see it above me beside our house. It took a while to get it down, but when it finally did, I was turning it towards me. Then it speeded up, and crashed in the wall right beside me. It did not hit anything else than the wall.

Several of the propels broke and some of the body. The issue now is that I sent it to DJI (based on guarantee since I bought it in July), and they say there was no malfunction. I disagree. Is there any point arguing with them, if any of you guys see it differently?

 

[Meta4]

Thank you, I uploaded the file here: DJI Flight Log Viewer - PhantomHelp.com

I've had a quick look at your flight data and it ppears that the reason for the drone's speed before the collision was that you were pulling all the way back on the right stick until one second before the crash.
For the last second, you pushed the right stick fully forward but the drone was flying at 8.3 metres/sec and the speed only reduced to 4.6 metres before the collision.

I'll look at the data in more detail a little later to see what else is in there.

 

[DJI man]

Aha, amazing that you can see that in the log. Thanks!

When I think about it, the reason could be that I had to go from "Cine" to "Normal" to get it to land. When I then pulled the right stick back, it got higher speed than I'm used to.

8,3 m/s is normal to reach within a couple of sec?

And do I have a case here, considering it would't land before i put it in "Normal" mode?

 

[Yorkshire_Pud]

Thank you, I uploaded the file here: DJI Flight Log Viewer - PhantomHelp.com

Ahh do you get an error message from that page ? I do in both firefox and chrome.


If you upload the log to Airdata

Airdata UAV - Flight Data Analysis for Drones

Airdata UAV - Flight Data Analysis for Drones

app.airdata.com

then click on notifications and "HD flight player" you can get a replay of the flight including joystick commands.

Then it speeded up, and crashed in the wall right beside me.

Ahh I made a mistake there, I have no idea why.

Looking at the log again, from approximately 393 seconds to 397.8 seconds you applied a combination of aileron and elevator that caused the drone to reach a speed of 8.3m/s. From 398.3 seconds to 399.1 seconds you applied full forward elevator to counteract the reverse speed but suspect the drone didn't have time to brake and avoid the wall. As such that looks like your fault, sorry.

I still think think there was a problem with the VPS sensors but don't know if it is possible to determine whether or not it was an actual problem with the sensors or something on the lenses.
The former might be grounds for a warranty claim but I don't think the latter is.
However I do note that the VPS height gives more realistic, and perhaps possible, values from 397 seconds onwards.

The attached shows the plots of various things produced by CsvView

CsvView Downloads

after it has processed the PH CSV, hopefully it is understandable. If you click on the thumbnail it will expand.

 

[Yorkshire_Pud]

"Cine" to "Normal" to get it to land

That shouldn't affect the drone's ability to land.

 

[DJI man]

My memory is slowly coming back to me- there was a bit of fog that day. A google-search makes me realize that the fog could be the reason for the drone not to land. I guess this is all on me.. Thank you both for helping me out here, much appreciated!

 

[Yorkshire_Pud]

My memory is slowly coming back to me- there was a bit of fog that day. A google-search makes me realize that the fog could be the reason for the drone not to land. I guess this is all on me.. Thank you both for helping me out here, much appreciated!

Ahh condensation on the lenses could have been the problem but before you give up all hope wait and see what the Meta4 et al thinks of the VPS data.

 

[Felix le Chat]

Hi, I was flying my my mini 2 se when in suddenly wouldn't land. After a while I changed to "normal" mode and I could get it down. When it was about 5 m above ground I was flying it towards me when it suddenly speeded up and crashed in the wall. Anyone who can interpret the flight log, and could see if it was a malfunction?

DJI crash log - Filemail

Click here to view and download these shared files from Filemail.com

www.filemail.com

Regards

Question. Do you have any third party clip-on accessories fitted to the belly of the drone?

 

[Yorkshire_Pud]

there was a bit of fog that day

Giving this a bit more thought, fogged sensors ARE thought to be able to cause slow descents but I don't see why they should prevent a landing.

Again, you want more knowledgeable folks than me to address that issue but IF my concern, that fogged sensors SHOULD NOT prevent a landing, is correct then it might be worth arguing that the inability to land was the cause of the rest of the flight and that if you had been able to land then you would not have subsequently crashed the drone.
Of course I suppose that DJI could argue that the drone shouldn't have been flown in or through fog.

It looks as if you didn't shoot video during the flight, a video might have been useful as evidence to show whether or not the drone was flown through or perhaps near fog.

The above said, if my concern is thought valid then the place to argue this is, perhaps, the DJI forum where the mods will sometimes re-refer a rejected claim if sensible supporting arguements and evidence is/are supplied.

 

[DJI man]

Question. Do you have any third party clip-on accessories fitted to the belly of the drone?

No.

 

[DJI man]

The above said, if my concern is thought valid then the place to argue this is, perhaps, the DJI forum where the mods will sometimes re-refer a rejected claim if sensible supporting arguements and evidence is/are supplied.

Good advice. No risk trying!

 

[msinger]

My memory is slowly coming back to me- there was a bit of fog that day

The VPS Height in your flight log shows there was something beneath the drone for most of the flight -- likely fog since you had no 3rd party accessories installed beneath the drone. That would have made the drone auto ascend to prevent from crashing into what it thought was an obstacle below.

 

[Yorkshire_Pud]

The VPS Height in your flight log shows there was something beneath the drone for most of the flight -- likely fog since you had no 3rd party accessories installed beneath the drone. That would have made the drone auto ascend to prevent from crashing into what it thought was an obstacle below.

View attachment 177979

Nice advert for the forth coming version of FlightReader.

Three questions
1) Do you see any evidence that this was a malfunction of the VPS system ?
2) Do you recollect any instances of fogged sensors preventing landing ?
3) Other than the drone flying clear of fog or the wind having blasted the lenses clear of condensation, do you have any explanation for why the VPS readings become more realistic in the last 3 seconds of the flight ?

 

[msinger]

Do you see any evidence that this was a malfunction of the VPS system ?

The OP confirmed he was flying through fog, so that's consistent with the data reported by the VPS system.

Do you recollect any instances of fogged sensors preventing landing ?

Is that a thing? Wouldn't the camera have been fogged up too?

Other than the drone flying clear of fog or the wind having blasted the lenses clear of condensation, do you have any explanation for why the VPS readings become more realistic in the last 3 seconds of the flight ?

Fog is very inconsistent. I'd expect the readings to change as the drone traveled through the fog.

There were likely areas of the flight (especially at lower altitudes) where no fog was present. That's likely why the VPS readings are higher in areas where the drone was closer to the ground.

 

[Yorkshire_Pud]

The OP confirmed he was flying through fog,

Not being awkward but did he? I think he said there was fog but I dont see him confirming he flew through it. I think I was the person who said "through fog".
And even if the drone did fly through fog I am still curious if, since you are better at this stuff than I am, you can spot any evidence of malfuctioning VPS ?

Is that a thing? Wouldn't the camera have been fogged up too?

I don't understand the relevance of that to the question I asked. Again I am bowing to your greater log reading ablility and possibly mental-library of incidents.

 

[msinger]

Not being awkward but did he?

See post #8 above.

 

[msinger]

And even if the drone did fly through fog I am still curious if, since you are better at this stuff than I am, you can spot any evidence of malfuctioning VPS ?

Possible evidence might be a scenario where the VPS is showing readings while the drone is high in the sky on a clear day. Furthermore, VPS sensors rarely malfunction, so it would be quite the oddity if that was the case here.

I don't understand the relevance of that to the question I asked. Again I am bowing to your greater log reading ablility and possibly mental-library of incidents.

I'm asking why the camera wasn't also fogged up. The OP didn't note he wasn't able to see the live video from the drone.

 

[DJI man]

Posted in dji forum, we'll see. And note, there was not a lot a fog this day, I could see the drone fine, especially when it finally landed.

DJI mini 2 SE

DJI mini 2 SE

forum.dji.com