Mavic mini up and down with no operator command

[marcosvc]

There are already many comments on this topic in this forum, and this is yet another documented contribution.
I saw that DJI replied to a client that the cause is the wind. I find that answer weak. I believe that DJI will fix this bug in some update.
Here is the description of what is in the video:
Video time:
15s went down to 9.9m he went up to 13m, 11m, 10m without the left stick command.
1.13s 11m, 13m, 12m without command
1.42s 13m, 16m, 17m without command (was against the wind)
2.15s stick down to 9.4m
2.23s started to rise, and I commanded the stick down. Up to 20m
2.44s started to descend without command until 7.7m so I put the stick up and he obeyed.
Obs. there was no communication delay, I saw the drone all the time and the responses were immediate.
Both the climbs without command and the descents were smooth, and as I was aware, I adjusted. After that it didn't happen again.

 

[Deleted member 114080]

Some more info is available in your Phone under Root/DJI/dji.go.v5/FlightRecord. Find the .txt fike of the right flight. Make an account on airdata.com upload the file and see if you can get more info what happened. There are also members here with more sophisticated analysing software such as @sar104, just upload the tekst file here and ask to take a look.

 

[sirbadger]

More info about how to retrieve your logs

Mavic Flight Log Retrieval and Analysis Guide

With the introduction of the Mavic Mini there has been a noticeable uptick in the number of posts requesting help with lost or crashed aircraft. While there is plenty of help and advice available on this forum, and detailed guidance and options on several other websites, I thought that perhaps a...

mavicpilots.com

 

[Thomas B]

Please post your logs.

 

[marcosvc]

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

 

[marcosvc]

 

[marcosvc]

more logs

 

[Meta4]

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

The times you have listed in post #1 do not match the times in the flight data.
I's going to be very difficult to match the data to the actions you are concerned about.

 

[sar104]

That is rather strange - there are several clear uncommanded height excursions during that flight.



Any chance you can find the mobile device DAT file ending FLY076.DAT?

 

[Mavic_Mini_Pilot]

I really really hope that DJI will fix this problem. I'm afraid that it could actually be something entirely hardware related, hence that's why only some Minis are exhibiting this behavior! My Mavic Mini has never lost nor gained altitude all by itself. I am watching out for that, but currently up till this point my Mini seems to be behaving really really good. I'll be watching this thread very carefully, hopefully it wasn't really actually the wind causing your Mini to go up and down a little bit, since your changes of elevation are not that big. Well they're big, but some other people get many many meters descents ans ascents.

cool cool

 

[sar104]

Okay - I found the direct cause of the problem in this case - it's kind of apparent in the graph above. Note that the descents are accompanied by appropriate values of the IMU vertical velocity (OSD_zSpeed), but that the periods of uncommanded climb are not. Differentiating the height with respect to time, then smoothing similarly to to the OSD algorithm and displaying it positive down shows the following - that derivative should match the OSD_zSpeed.



For the uncommanded climbs the IMU is not detecting the vertical motion, even though the OSD_height parameter is detecting the change in height, presumably via the barometer. It's not consistent though - sometimes it is detecting the climbs. Anyway - there is undetected climb occurring and then the FC subsequently corrects to bring the aircraft back down - the uncommanded descent. We really need to see the raw sensor data in the DAT file, but this is clearly a hardware or firmware problem.

 

[Mavic_Mini_Pilot]

Okay - I found the direct cause of the problem in this case - it's kind of apparent in the graph above. Note that the descents are accompanied by appropriate values of the IMU vertical velocity (OSD_zSpeed), but that the periods of uncommanded climb are not. Differentiating the height with respect to time, then smoothing similarly to to the OSD algorithm and displaying it positive down shows the following - that derivative should match the OSD_zSpeed.

View attachment 92957

For the uncommanded climbs the IMU is not detecting the vertical motion, even though the OSD_height parameter is detecting the change in height, presumably via the barometer. It's not consistent though - sometimes it is detecting the climbs. Anyway - there is undetected climb occurring and then the FC subsequently corrects to bring the aircraft back down - the uncommanded descent. We really need to see the raw sensor data in the DAT file, but this is clearly a hardware or firmware problem.

Differentiating height with respect to time, some sweet awesome Calculus hey. Wow you are really really smart hey, I mean like super smart! I hope that DJI fixes this problem. I'm willing to bet that it is a Hardware problem, and that it may be fixable via new firmware, but most probably not. That part which is detecting the non-real ascends and descents needs to be changed out for a new one that works properly.

So its safe to say that if you haven't had this issue you probably will not get it - and if you have had it then it will not go away easily..

I'm so happy that mine is OK, I hope that DJI really does something about this, either new drones, free service or something..

cool cool

 

[sar104]

Differentiating height with respect to time, some sweet awesome Calculus hey. Wow you are really really smart hey, I mean like super smart! I hope that DJI fixes this problem. I'm willing to bet that it is a Hardware problem, and that it may be fixable via new firmware, but most probably not. That part which is detecting the non-real ascends and descents needs to be changed out for a new one that works properly.

So its safe to say that if you haven't had this issue you probably will not get it - and if you have had it then it will not go away easily..

I'm so happy that mine is OK, I hope that DJI really does something about this, either new drones, free service or something..

cool cool

But, just to clarify, this doesn't look to me at this stage to be related to the multiple reports of uncommanded descent where the aircraft simply descends uncontrollably to the ground.

 

[marcosvc]

Sar104. Follow attached the file you asked.
Thanks for your help

 

[marcosvc]

The times you have listed in post #1 do not match the times in the flight data.
I's going to be very difficult to match the data to the actions you are concerned about.

Here are the drone time (instead of my video)
8:38 went down to 9.9m he went up to 13m, 11m, 10m without the left stick command.
9:36s 11m, 13m, 12m without command
10:05s 13m, 16m, 17m without command (was against the wind)
10:37s stick down to 9.4m
10:45s started to rise, and I commanded the stick down. Up to 20m
11:07s started to descend without command until 7.7m so I put the stick up and he obeyed.

 

[sar104]

Sar104. Follow attached the file you asked.
Thanks for your help

That's not the flight - just a 30 seconds power up. There should be another, larger DAT file with the same suffix, unless it's fallen victim to the Fly app DAT file deletion issue.

 

[flyframes]

I had the same issue with mu brand new Mavic Mini. First flight.... after had this issue i search online for this problem. Removed firmware 0400 to 0300 and issue dissapeared. I think evertthing is in faulty firmware.

 

[marcosvc]

I had the same issue with mu brand new Mavic Mini. First flight.... after had this issue i search online for this problem. Removed firmware 0400 to 0300 and issue dissapeared. I think evertthing is in faulty firmware.

I had this problem only once. I have 45 flights with the MM, perfect ones. I even made a long range test with 4km (total 8km) with no variances. I hope to find this one time bug.

 

[marcosvc]

That's not the flight - just a 30 seconds power up. There should be another, larger DAT file with the same suffix, unless it's fallen victim to the Fly app DAT file deletion issue.

I did not find any other fly076.dat file. Would help if I download the log from the aircraft ?

 

[sar104]

I did not find any other fly076.dat file. Would help if I download the log from the aircraft ?

No - unfortunately only DJI can decrypt those.