Normal mode, I was trying to get video of the landscaping work we've done.  The drone suddenly went right and crashed into the ground.  DJI tells me it was pilot error.
I'm not sure I could crash it like this if I wanted to.  Wouldn't the sensors on the bottom keep it from slamming into the ground?https://www.icloud.com/iclouddrive/09EK_oIswHm8NKaZkXx_u4e9w#Drone_Crash

Hello !

Your drone before the crash has a strange shake ..... are you sure that the propellers were well fixed?

Max90payneFPV Posted at 7-1 08:25
Hello !

Your drone before the crash has a strange shake ..... are you sure that the propellers were well fixed?

None of the props came off during the crash.

It's possible the .DAT log file that got recorded by the Fly App will contain data explaining the incident. To see how to retrieve that .DAT look here.
https://fpvdronepilots.com/threads/dat-available.4225/#post-35016

If you post it via Dropbox or similar I'll take a look

The drone is shaking badly just before the crash although there is no wind. I think this is hardware related. Maybe bent props or smth. The flight logs will help for sure, please share.

BudWalker Posted at 7-1 09:40
It's possible the .DAT log file that got recorded by the Fly App will contain data explaining the incident. To see how to retrieve that .DAT look here.
https://fpvdronepilots.com/threads/dat-available.4225/#post-35016

https://www.icloud.com/iclouddrive/0j0tkf4Ny_IWWN1fnuSuKBNbQ

I'm not an expert drone pilot, but I don't just slam it into the ground either.

Here's what DJI said after analyzing the logs and the drone:
Stick Command Error under P Mode
FLY60
Incident Date: 6-17-2021
1.      The aircraft worked under GPS mode, and responded to the pilot’s command well;
2.      user flew the drone side way and crashed into obstacle
3.      Flight Time T=174 s, Relative Height H=5.8 m,  unit recorded external impact.

According to the analysis, the incident was not caused by any product malfunction.

BudWalker Posted at 7-1 09:40
It's possible the .DAT log file that got recorded by the Fly App will contain data explaining the incident. To see how to retrieve that .DAT look here.
https://fpvdronepilots.com/threads/dat-available.4225/#post-35016

Log File - https://www.icloud.com/iclouddrive/0j0tkf4Ny_IWWN1fnuSuKBNbQ

Here is the log file...I think.  Never had to retrieve one before.

timinconway Posted at 7-1 11:15
Log File - https://www.icloud.com/iclouddrive/0j0tkf4Ny_IWWN1fnuSuKBNbQ

Here is the log file...I think.  Never had to retrieve one before.

That's the .txt log file. Can you retrieve the .DAT in the FlightRecords/MCDatFlightRecords directory?

BudWalker Posted at 7-1 13:13
That's the .txt log file. Can you retrieve the .DAT in the FlightRecords/MCDatFlightRecords directory?

https://www.icloud.com/iclouddrive/0E8LCx1DFLkt5WjS-izhvRALw#FLY056-060-20210617164749

Try this one.  I assume that's the date/time code at the end of the file name.

timinconway Posted at 7-1 13:32
https://www.icloud.com/iclouddrive/0E8LCx1DFLkt5WjS-izhvRALw#FLY056-060-20210617164749

Try this one.  I assume that's the date/time code at the end of the file name.

Yes, that's the right one. I'm still looking at it. I should have more to say tomorrow morning west coast time. It's not pilot error.

Interesting incident. Bottom line is that it looks like the Flight Controller mistakenly computed the location to be longitude = 0, latitude = 0. The FC then tried to compensate and flew the wrong way. Before I give some more details could you confirm a few things. The beginning of the video was at point A? Point B is the end of the video where you picked up the FPV and walked it back to the launch point?

BudWalker Posted at 7-2 06:05
Interesting incident. Bottom line is that it looks like the Flight Controller mistakenly computed the location to be longitude = 0, latitude = 0. The FC then tried to compensate and flew the wrong way. Before I give some more details could you confirm a few things. The beginning of the video was at point A? Point B is the end of the video where you picked up the FPV and walked it back to the launch point?


[Image]

You are correct on Points A & B.

timinconway Posted at 7-2 06:54
You are correct on Points A & B.

The Flight Controller computes the location estimate by fusing several  different data. This includes GPS, accelerometer, gyro, magnetometer and  possibly other data. The FC also computes the confidence of that  estimate.

In this incident at time 100.897 secs the FC reduced the confidence from  5 (best) to 0 (no confidence).  Normally when the confidence falls to 0  the FC will switch to ATTI mode and not attempt to correct for any  positional errors. In this case the FC switched to ATTI briefly (0.5  secs) and then back to GPS_ATTI mode. But the FC then started computing  the location to be [0°, 0°] which is south from it's actual position.  Since it was in GPS_ATTI mode the FC was then commanding a northerly  direction in an attempt to correct the incorrect  [0°, 0°] location.

The blue shows the confidence falling to 0 and the longitude (green  plot) going to 0° from -92.5°. The pink background depicts ATTI mode.

It's largely irrelevant why the confidence fell to 0 and the location  estimate went to a default value of  [0°, 0°] - the FC should have  switched to and remained in ATTI where it would not attempt to correct  position.

Its probably that one of the props came loose

Or there might have been some signal interference

BudWalker Posted at 7-2 08:07
The Flight Controller computes the location estimate by fusing several  different data. This includes GPS, accelerometer, gyro, magnetometer and  possibly other data. The FC also computes the confidence of that  estimate.

In this incident at time 100.897 secs the FC reduced the confidence from  5 (best) to 0 (no confidence).  Normally when the confidence falls to 0  the FC will switch to ATTI mode and not attempt to correct for any  positional errors. In this case the FC switched to ATTI briefly (0.5  secs) and then back to GPS_ATTI mode. But the FC then started computing  the location to be [0°, 0°] which is south from it's actual position.  Since it was in GPS_ATTI mode the FC was then commanding a northerly  direction in an attempt to correct the incorrect  [0°, 0°] location.
[Image]

DJI wants $114 to repair it. I’ll probably never convince them it was not my fault.

Which could have meant the drone responded slowly or more vigorously than it should have

BudWalker Posted at 7-2 08:07
The Flight Controller computes the location estimate by fusing several  different data. This includes GPS, accelerometer, gyro, magnetometer and  possibly other data. The FC also computes the confidence of that  estimate.

In this incident at time 100.897 secs the FC reduced the confidence from  5 (best) to 0 (no confidence).  Normally when the confidence falls to 0  the FC will switch to ATTI mode and not attempt to correct for any  positional errors. In this case the FC switched to ATTI briefly (0.5  secs) and then back to GPS_ATTI mode. But the FC then started computing  the location to be [0°, 0°] which is south from it's actual position.  Since it was in GPS_ATTI mode the FC was then commanding a northerly  direction in an attempt to correct the incorrect  [0°, 0°] location.

I understand why it would have sent it in a northern direction but any idea why it would also send it to the ground?

Do the flight logs show the position of the control sticks during all this?

Alvara74 Posted at 7-2 08:35
Its probably that one of the props came loose

The props were intact when I recovered it.  They were bent but they were solidly attached.

Let me know if you find a fix

timinconway Posted at 7-2 09:05
The props were intact when I recovered it.  They were bent but they were solidly attached.[view_image]

Are you certain that there was no damage to the props before the flight?

It could have been that you may have skimmed past something mid flight that would have slowly damaged a prop

timinconway Posted at 7-2 09:05
The props were intact when I recovered it.  They were bent but they were solidly attached.[view_image]

And that looks brutal...

Can you check your flight log or the area where you were flying to see if your drone came anywhere close to a strong object which might have impacted a prop?

Alvara74 Posted at 7-2 09:35
Can you check your flight log or the area where you were flying to see if your drone came anywhere close to a strong object which might have impacted a prop?

There is absolutely no data to suggest that there was a prop strike. Also, the video has no indication that there was a prop strike.

timinconway Posted at 7-2 09:04
I understand why it would have sent it in a northern direction but any idea why it would also send it to the ground?

Do the flight logs show the position of the control sticks during all this?

The upper chart shows the control inputs, the lower shows the response. There had been some CW roll input until 100.733 secs. After that the roll and pitch input were 0.0. As can be seen there continued to be CW rol and pitch up response after the roll input was removed. The extreme roll and pitch after the dotted line was the impact.




Not sure why the FPV lost altitude. The ultrasonic height data and vision system both indicate < 0.5 m altitude was lost. How much did it actually lose?

BudWalker Posted at 7-2 13:07
The upper chart shows the control inputs, the lower shows the response. There had been some CW roll input until 100.733 secs. After that the roll and pitch input were 0.0. As can be seen there continued to be CW rol and pitch up response after the roll input was removed. The extreme roll and pitch after the dotted line was the impact.

I would estimate it was at 3m altitude when this happened.

BudWalker Posted at 7-2 12:33
There is absolutely no data to suggest that there was a prop strike. Also, the video has no indication that there was a prop strike.

Did you not even get a warning on your screen?

Usually there's a warning if there is an error in a part of the drone

Alvara74 Posted at 7-2 13:55
Did you not even get a warning on your screen?

I suspect you are referrring to messages like
Propulsion system error.
Motor overloaded Motor is Blocked.
Not Enough Force/ESC Error.

Those occurred after the FPV had hit the ground with the motors still running.

BudWalker Posted at 7-2 14:12
I suspect you are referrring to messages like
Propulsion system error.
Motor overloaded Motor is Blocked.

Then I don't know what might have happened

Sorry for not being much help

>  The drone suddenly went right and crashed into the ground.
>  Wouldn't the sensors on the bottom keep it from slamming into the ground?
It didn't crash into the ground but into a tree that was behind it, that's why collision detection sensors couldn't help. Crash seems to happen about a meter above ground. This can be seen clearly if you inspect video frame by frame.

> DJI tells me it was pilot error.
This should be as simple as looking at flight log and comparing stick position record to drone movement at the time. If the flight log does not show your sticks being in the position to cause this movement, then it clearly can't be a pilot error, and it would be utterly ridiculus to claim that it was.

I suggest using https://airdata.com/ to process your flight log and look at stick position prior to the crash. That's black on white evidence of it being or not being pilot error.

Shadetail Posted at 7-3 09:03
>  The drone suddenly went right and crashed into the ground.
>  Wouldn't the sensors on the bottom keep it from slamming into the ground?
It didn't crash into the ground but into a tree that was behind it, that's why collision detection sensors couldn't help. Crash seems to happen about a meter above ground. This can be seen clearly if you inspect video frame by frame.

I used Airdata and it shows that as soon as the drone started flying off to the right, I let go of the sticks and it continued to crash.  Both sticks were in the neutral position.  DJI refuses (so far) to admit this.  Tech Support told me..."I have to go with what our engineers concluded from the flight data."

That's ridiculous of them to claim that. I don't know enough about law to say if lawsuit would make any sense financially, but I'm sure that you'd win if you were to pursue it given that there's clearest possible evidence in your favor. At the very least you ought to try to bring this to media attention, make a short youtube video about it, or at least a blog post, share it on reddit, twitter, tweet DJI directly, somehow get their community managers or PR people to notice, ask reddit for advice on how to most effectively do this in all relevant subreddits. People do tend to take out their pitchforks and rally behind the victim en-masse in such cases, and while it's impossible to say how much of such cases go unnoticed, I've often seen this done successfully and big companies apologizing publically as a result and refunding users.

timinconway Posted at 7-3 09:44
I used Airdata and it shows that as soon as the drone started flying off to the right, I let go of the sticks and it continued to crash.  Both sticks were in the neutral position.  DJI refuses (so far) to admit this.  Tech Support told me..."I have to go with what our engineers concluded from the flight data."[view_image]

Having analyzed hundreds of incidents I sometimes get to see what happens when DJI can't or won't look at the evidence.  An easier path to victory may be to start over with DJI tech support. Start a new chat, email or telephone conversation and maybe you will get a better response. There seems to be no unified approach in DJI's response in cases like this. Sometimes it's just some flight analyst wanna-be who takes a very superficial look. Other times it can be somebody who wants to take a closer look and do the right thing.


An attempt at a PR campaign or legal coercion to convince DJI to do the right thing won't be successful.