Mavic Air crash due to IMU error

[KalaKai]

I had IMU and compass error when I turned her on, but cleared both of those prior to flight.

Flying started OK, but very quickly ended in crash after IMU error and lost control. I would very much appreciate someone looking at the logs and letting me know what went wrong. She's up in a tree and I haven't found a way to get her down yet. argh!

link to flight log viewer

is this the correct way to provide this info, or is it better to upload the txt file?

 

[slup]

I had IMU and compass error when I turned her on, but cleared both of those prior to flight.

Flying started OK, but very quickly ended in crash after IMU error and lost control. I would very much appreciate someone looking at the logs and letting me know what went wrong. She's up in a tree and I haven't found a way to get her down yet. argh!

link to flight log viewer

is this the correct way to provide this info, or is it better to upload the txt file?

Looks like you started from a boat ... I have my suspicions about the reason but need the mobile device .DAT log to diagnose it properly.

How to retrieve the mobile device .DAT log from the device you flew with --> https://mavicpilots.com/threads/mavic-flight-log-retrieval-and-analysis-guide.78627/#post-889224 (read under section 3.). The .DAT log for this flight ends with FLY093.DAT ... retrieve it & attach it to a new post in this thread.

 

[KalaKai]

... retrieve it & attach it to a new post in this thread.

I'm getting an error uploading that file is too large...only a minute flight, file size 26,000 kb...any ideas of the problem

ETA; while flight was only1 minute, the file probably includes all data from time unit turned on until batteries died...guess this is why file size is so large...tried to zip the file, but still to large

 

[slup]

I'm getting an error uploading that file is too large...only a minute flight, file size 26,000 kb...any ideas of the problem

ETA; while flight was only1 minute, the file probably includes all data from time unit turned on until batteries died...guess this is why file size is so large...tried to zip the file, but still to large

Impossible ... a MOBILE device .DAT isn't that large, think you've retrieved the Aircraft DAT. Confirm that it's the right DAT & then use the "Attach"-button under the new post you try to include it in.

 

[KalaKai]

Looks like you started from a boat ... I have my suspicions about the reason but need the mobile device .DAT log to diagnose it properly.

How to retrieve the mobile device .DAT log from the device you flew with --> Mavic Flight Log Retrieval and Analysis Guide (read under section 3.). The .DAT log for this flight ends with FLY093.DAT ... retrieve it & attach it to a new post in this thread.

Impossible ... a MOBILE device .DAT isn't that large, think you've retrieved the Aircraft DAT. Confirm that it's the right DAT & then use the "Attach"-button under the new post you try to include it in.

The file is called 2020-07-27 17-31-56 FLY093.DAT and is located in the FlightRecords/MCDatFlightRecords folder. all of the other files in that folder are between 1 & 5,000 kb. This one is 26,000 kb. Am I in the wrong place?

 

[slup]

If you are in the mobile phone, that's the correct place ... try upload it to some cloud storage like dropbox & share a link, really curious to see what makes that DAT so large.

 

[boblui]

The file is called 2020-07-27 17-31-56 FLY093.DAT and is located in the FlightRecords/MCDatFlightRecords folder. all of the other files in that folder are between 1 & 5,000 kb. This one is 26,000 kb. Am I in the wrong place?

The flight was made on 2020-07-28 , motor start time was 17:34:29. The .DAT file you have mention was created on the day before ( July 27 ) so it's not the correct file.

In the .TXT file, the data field OSD.flycVersion = 31 so the correct log file should end with "FLY031" I think. May be the others can correct me if I am wrong

 

[BudWalker]

The flight was made on 2020-07-28 , motor start time was 17:34:29. The .DAT file you have mention was created on the day before ( July 27 ) so it's not the correct file.

In the .TXT file, the data field OSD.flycVersion = 31 so the correct log file should end with "FLY031" I think. May be the others can correct me if I am wrong

HOME.dataRecorderFileIndex is the column that has the .DAT file number. In this case that's 93 so
......FLY093.DAT is the correct .DAT. Don't know why the 2020-07-28 vs 2020-07-27 discrepancy. Maybe a typo. Maybe related to why the .DAT is so large....

 

[KalaKai]

couldn't upload, so link below is to public shared onedrive folder....

public folder with dat file

If you are in the mobile phone, that's the correct place ... try upload it to some cloud storage like dropbox & share a link, really curious to see what makes that DAT so large.

one drive public folder with file

 

[KalaKai]

The flight was made on 2020-07-28 , motor start time was 17:34:29. The .DAT file you have mention was created on the day before ( July 27 ) so it's not the correct file.

In the .TXT file, the data field OSD.flycVersion = 31 so the correct log file should end with "FLY031" I think. May be the others can correct me if I am wrong

Sorry, my mistake...typo...filename is 28, not 27.

File uploaded to dat file

 

[boblui]

The .DAT file is so big because the drone remained powered on and the app was not closed until about 80 minutes after the crash. Amazingly, CsvView has no problem in processing the file. Kudo to Bud !

This is the flight :


From the pattern of the yaw curve, IMU1 was used throughout the entire flight. Not sure if switching has occurred in the last couple of seconds but it probably wouldn't have made any difference. At 56.5 sec ( after the motors are powered up ), the system issued the warning "IMU heading error. Please try flying forward and backward or land as soon as possible to calibrate the compass" . This was seen in the flight log. Crash occurred 6 seconds later.

It can be seen from the chart below that the yaw reading of IMU0 and IMU1 are very different.



Upon comparing with the magnetic field direction, It turns out that IMU1's compass was faulty. IMU0's compass is off as well but the error stayed at about 20 degrees which should not cause anything catastrophic.



It is also noticed that the motion sensors of IMU1 were faulty as well because the velocities calculated from the sensor data was very different from those indicated by GPS :



In short, IMU1 was almost totally faulty and unfortunately it was selected by the flight controller for reasons not apparent to me. Flying with a bad IMU is like a drunk person trying to walk a straight line. Losing control is the expected result.

This happened to my Mini when the compass was off by 180 degrees. I did not know that until I read the flight log afterwards :

 

[KalaKai]

In short, IMU1 was almost totally faulty and unfortunately it was selected by the flight controller for reasons not apparent to me. Flying with a bad IMU is like a drunk person trying to walk a straight line. Losing control is the expected result.

Wow. Your ability to read and interpret these logs is quite impressive! thanks so much!

I don't think this is in the flight log as it may have been before the motor started, but I did receive error 'IMU Redundancy Switch' when I turned the drone on. I see this message when I watch the flight reply video.

The drone is still about 40 or so' up in a tree and will probably stay there until blown down by the next storm. It's so frustrating as I could see her, but couldn't get to her! I'm traveling, so when she finally does decide to come down, I won't even be there to give her a proper burial

Are these types of failures common? Could something I have done contributed to the failure? Moisture, salt air environment, other?

 

[boblui]

Are these types of failures common? Could something I have done contributed to the failure? Moisture, salt air environment, other?

I would say it's not rare. Crashes due to IMU problems have been reported in this forum from time to time. Things like crafts ascending by itself, veering off the course and colliding into trees, etc have happened.

As the IMU is composed of sealed chips, I don't think they are affected by moisture. Here is a video on the construction of IMU and how it works in case you are interested :

 

[KalaKai]

The .DAT file is so big because the drone remained powered on and the app was not closed until about 80 minutes after the crash. Amazingly, CsvView has no problem in processing the file. Kudo to Bud !

This is the flight :
View attachment 109278

From the pattern of the yaw curve, IMU1 was used throughout the entire flight. Not sure if switching has occurred in the last couple of seconds but it probably wouldn't have made any difference. At 56.5 sec ( after the motors are powered up ), the system issued the warning "IMU heading error. Please try flying forward and backward or land as soon as possible to calibrate the compass" . This was seen in the flight log. Crash occurred 6 seconds later.

It can be seen from the chart below that the yaw reading of IMU0 and IMU1 are very different.

View attachment 109279

Upon comparing with the magnetic field direction, It turns out that IMU1's compass was faulty. IMU0's compass is off as well but the error stayed at about 20 degrees which should not cause anything catastrophic.

View attachment 109280

It is also noticed that the motion sensors of IMU1 were faulty as well because the velocities calculated from the sensor data was very different from those indicated by GPS :

View attachment 109281

In short, IMU1 was almost totally faulty and unfortunately it was selected by the flight controller for reasons not apparent to me. Flying with a bad IMU is like a drunk person trying to walk a straight line. Losing control is the expected result.

This happened to my Mini when the compass was off by 180 degrees. I did not know that until I read the flight log afterwards :

I retrieved the drone! That's the good news. Bad news is that I assume that whatever caused it to crash was not resolved.

Is there a way to run a diagnostics to determine if all hardware is still working OK? I guess she runs some sort of diagnostics each time I start her, but is there an program that does something more indepth I can get?

I can do some low level flights and see what happens, but would like to know before I send her on long missions again.

 

[Rishu]

Hi
Yesterday my mavic air also crashed due to imu. Take off was smooth. It was hovering on 17 meters . I got a error "main data controller error " and after getting this error the drone crashed .
Now my mavic air gimble and front left arm is broken
I have changed the front left arm
And i changed the gimble also but after changing gimble no camera feed shown . So i removed the gimble .
Now waiting for new imu board and gimble to come .