Mavic 2 Zoom crashed and DJI says drone has no log

[BlueKale]

My Mavic 2 Zoom crashed a few weeks ago. It was flying a parallel sweep pattern (“Norm Mission”) at around 325’ AGL using Map Pilot and had completed the first flight of the day without incident. I installed a fresh battery and resumed the mission. I confirmed on the controller display that the drone was in waypoint mode. About 1 minute 12 seconds into the second flight, while the drone was flying straight and with no control input to the mission resume point, I stopped receiving telemetry from the drone. An observer noted that the drone fell out of the sky, pitching and rolling on the way down as if uncontrolled.

The accident occurred in a rural area with no people or structures for at least half a mile. The drone was less than 250’ from me and I had a clear line of sight. I do not suspect any sort of signal interference.

Looking at the Map Pilot log file, it appears that the telemetry stopped abruptly around when the drone started falling. I see no major red flags in the log, although the drone did increase its pitch by about 6 degrees in the 3 seconds preceding the crash, suggesting that perhaps it was fighting a wind gust.

I was able to find the drone about 10 minutes later. The drone suffered damage to the landing gear, battery and all 4 propellers, and possibly a scratch to the camera lens. I sent it to DJI as a warranty claim, hoping they can analyze the flight log and tell me what happened. I have DJI Care Refresh and am willing to pay for the repair if I was at fault, but I’d like to confirm whether the crash was the result of a product defect.

Prior to sending the drone to DJI, I used DJI Assistant 2 to download the logs from the aircraft. I was able to obtain a 1.4 GB .DAT file whose timestamp matched the time of the crash. Of course I can’t do anything with it given DJI’s decision to encrypt Mavic 2 onboard flight logs and failure to provide tools to extract data from them.

DJI’s response to my warranty request was:

According to the analysis, no crash was found in the available flight records.

I’m a little perplexed, given that I was able to download a large .DAT file with a corresponding timestamp from the drone, that DJI can’t find anything. I’ve been composing a message back to them asking for more details on what they did find, and offering to provide them a copy of the .DAT file that I have in case for some reason they can’t download it from the drone.

The Map Pilot log file can be downloaded from:

https://app.airdata.com/original?flight=d9105535fdcac3104325d157be9845e6

The AirData analysis is at:

https://app.airdata.com/flight/d9105535fdcac3104325d157be9845e6/

I’d be happy to supply the .DAT file if anyone has figured out a way to extract data from them….

I welcome any thoughts on the next step I should take with DJI. I’d like them to analyze the .DAT file and tell me what may have happened. Having the flight logs encrypted and being unable to get DJI to tell me what in them them following a catastrophic aircraft failure doesn’t inspire confidence.

 

[mrreddog]

Man that sucks, I wonder if DJI is just trying to pull a fast one, not knowing you backed up the .dat files??
Wish we could get the encryption key

 

[sar104]

I doubt that they mean they can't find the DAT file - it's more likely that the log just stops, containing little or no record of the crash. How carefully they looked at it is another matter. Onboard DAT files record data at a much higher sample rate than the mobile device logs, which log at 10 s/s.

In the case of power loss due to a battery disconnect (one cause of this kind of behavior), the 50 s/s battery voltage data in the onboard DAT file typically show a characteristic voltage collapse over a timescale of a few tens of milliseconds. Alternatively, if it is an FC process crash, which seems to be the other known cause, the logs just stop, but often preceded by indications of other errors. It's never been clear to me whether the first-level DJI support staff who look at these logs are familiar with these kinds of subtleties.

 

[mrreddog]

But I would imagine the sheer number of Dat files they review, whether it be Spark, Mavic, M2, Inspire, ETC. you would think they would develop patterns with the data and the complaint and have a vague idea of the problem.... But I suppose their in this business to make money. No Data, No Replacement.... Who knows...

 

[sar104]

But I would imagine the sheer number of Dat files they review, whether it be Spark, Mavic, M2, Inspire, ETC. you would think they would develop patterns with the data and the complaint and have a vague idea of the problem.... But I suppose their in this business to make money. No Data, No Replacement.... Who knows...

Agreed, but from numerous cases where they have missed things or misinterpreted the data, and then subsequently got it right after review, it's clear that at least some of their analysis is rather cursory.