Mavic Mini suddenly moves without stick input

[AnGo]

Hello,

today I took a flight with my Mavic Mini and experienced an unexpected horizontal move where no such stick input was done. I tried already to analyze the logfile with phantomhelp.com and I also there see the behavior recorded. Here is the link:

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

The issue starts at 2m23s after launch and ends 2m49s when I realized the unexpected drift during I did a slow CW yaw rotation and pushed the forward stick to regain control - which fortunately worked.
I am using Mode 3, so left stick controls roll and pitch, right stick does yaw and throttle.

I first thought maybe too less satellites and it went to Atti-Mode, but there was neither a message on the app nor in the log. Also there was no significant wind that could have overloaded the engines - also there was no strong wind warning raised.
Maybe someone here has an explanation?

Thanks and regards

 

[DrizzyDrones]

post the dat file and our wizards will be able to explain exactly what happened

 

[slup]

Hello,

today I took a flight with my Mavic Mini and experienced an unexpected horizontal move where no such stick input was done. I tried already to analyze the logfile with phantomhelp.com and I also there see the behavior recorded. Here is the link:

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

The issue starts at 2m23s after launch and ends 2m49s when I realized the unexpected drift during I did a slow CW yaw rotation and pushed the forward stick to regain control - which fortunately worked.
I am using Mode 3, so left stick controls roll and pitch, right stick does yaw and throttle.

I first thought maybe too less satellites and it went to Atti-Mode, but there was neither a message on the app nor in the log. Also there was no significant wind that could have overloaded the engines - also there was no strong wind warning raised.
Maybe someone here has an explanation?

Thanks and regards

Yes indeed, you had a drift there in P-mode with navhealth 5 & 14-16 sats locked ... from where you released the pitch stick to stop (your left forward, dark yellow graph) at approx 144sec into the flight the AC drifted approx 2m in the wind direction (see further below regarding winds) before you started to slowly rotate in CW direction at 155sec (green graph). During the rotation the AC continued to drift & by the time until you stopped the rotation & started to apply the pitch stick again the AC had drifted further 16m, (the marker is placed where you start to rotate in the chart). Click on the charts below to make them larger.



The winds in that spot where the drift occurred ...



... and the distance it drifted along the GPS path.



Even though the winds appeared to be stronger than you thought the AC didn't do anything to fight it ... it's seen when looking at the pitch (purple graph) & roll (light yellow graph) angles during the drift ... the AC is mainly horizontal.

But if we instead look at the differences in GPS & IMU velocity a possible reason emerge ... if everything is healthy those velocities should be pretty much the same, meaning the velocity differences should be very close to zero. But the log gives another picture ...

Overall this doesn't look good ... but there where the drift starts the velocities in direction north & east starts to create a major deviation (starts where the marker is when the drift starts & ends approx when you have finished the rotation). Usually this indicate a disturbed compass at power on or IMU problems. In order to dig deeper the mobile device DAT log is needed, read up here on how to retrieve (see section 3.) --> Mavic Flight Log Retrieval and Analysis Guide , the one needed for this flight ends with FLY045.DAT.



Another way to depict the deviation between the GPS & the IMU is like below ... the red path is from the GPS & the green one is how the IMU thought the path was performed.

 

[boblui]

It seems that the craft simply did not know it was drifting with the wind ...

I am examining the veloctiy variables xSpeed, ySpeed, VelN and VelE . I no longer have the Mini but what I have observed on my M2P is that :

When GPS is available, the velocities are likely the result of fusing GPS / inertia sensor outputs

xSpeed = VelN = velocity component in the north direction in the earth's FOR ( Frame Of Reference )
ySpeed = VelE = velocity component in the east direction in the earth's FOR

When GPS is NOT available ( indoor ), the velocities should be calculated from inertia sensor only ( may be the bottom camera as well ? )

xSpeed = forward / backward velocity in the craft's FOR
ySpeed = leftward / rightward velocity in the craft's FOR
VelN and VelE figures are not available.

Before the drifting started, all four speeds were essentially zero but when the craft was drifting, xSpeed and ySpeed remained zero but VelN and VelE picked up.

It seems that the drifting was sensed by the GPS receiver but not the insertia sensors and the flight controller has chosen to believe the inertia sensors !

Then full elevator was applied and the inertia sensors agreed with the GPS again so the abnomality stopped. If the OP can post the .DAT file, may be more insight can be obtained.

 

[AnGo]

Thanks for your analysis done so far. Here is the dat file of the flight as requested.
Edit: Now the correct one

 

[slup]

Thanks for your analysis done so far. Here is the dat file of the flight as requested.

That's not the correct DAT log... the correct one ends with FLY045. DAT.

 

[AnGo]

That's not the correct DAT log... the correct one ends with FLY045. DAT.

Sorry. I updated the resource

 

[slup]

The DAT log doesn't seem to give a definitive answer to the drift ... but some odd thing's occurring only the hover (with a slight yaw rotation) can be seen though.

The period depicted in the chart below is from where the drift starts until you release the yaw stick ... putting up the velocity signals from the IMU, the GPS & the VIO (reading off movements with the VPS sensor on the belly) show a small deviation which could have made the IMU to stick with it's info regarding the velocity. It's clear that the IMU don't register any movement even though the GPS show movement in positive north & negative east ... but adding in the registered VIO movement they are showing negative north & positive east. Could it be this deviation that made the FC to go with the IMU values?



As soon as you apply horizontal movement inputs the 3 velocity groups start to agree reasonably again ...



This flight started up with a compass calibration prompt due to both days since the last & the distance from where the last was performed was to large ...

You did it 2 times ... the first failed with a to large error



The second did go through but it wasn't a good calibration ... a larger error still remained



My recommendation is that you do a new compass calibration... & possibly also an IMU calibration due to what is shown in the DAT log event stream just after power on.

 

[AnGo]

That's true, it requested 2 times a compass calibration. First time failed after I did the 2 steps displayed in the app. Rotate along 2 axis. Is there a trick how to ensure it is a good calibration? I already had places where I got 5 times calibration failed before it was accepted, always doing the same steps

 

[AboriginalBrotherhood1162]

Yes indeed, you had a drift there in P-mode with navhealth 5 & 14-16 sats locked ... from where you released the pitch stick to stop (your left forward, dark yellow graph) at approx 144sec into the flight the AC drifted approx 2m in the wind direction (see further below regarding winds) before you started to slowly rotate in CW direction at 155sec (green graph). During the rotation the AC continued to drift & by the time until you stopped the rotation & started to apply the pitch stick again the AC had drifted further 16m, (the marker is placed where you start to rotate in the chart). Click on the charts below to make them larger.

View attachment 110414

The winds in that spot where the drift occurred ...

View attachment 110416

... and the distance it drifted along the GPS path.

View attachment 110417

Even though the winds appeared to be stronger than you thought the AC didn't do anything to fight it ... it's seen when looking at the pitch (purple graph) & roll (light yellow graph) angles during the drift ... the AC is mainly horizontal.

But if we instead look at the differences in GPS & IMU velocity a possible reason emerge ... if everything is healthy those velocities should be pretty much the same, meaning the velocity differences should be very close to zero. But the log gives another picture ...

Overall this doesn't look good ... but there where the drift starts the velocities in direction north & east starts to create a major deviation (starts where the marker is when the drift starts & ends approx when you have finished the rotation). Usually this indicate a disturbed compass at power on or IMU problems. In order to dig deeper the mobile device DAT log is needed, read up here on how to retrieve (see section 3.) --> Mavic Flight Log Retrieval and Analysis Guide , the one needed for this flight ends with FLY045.DAT.

View attachment 110418

Another way to depict the deviation between the GPS & the IMU is like below ... the red path is from the GPS & the green one is how the IMU thought the path was performed.

View attachment 110419

Would it be okay if I run iOs Android beta version litchi app. To check data or see something new.thankyou.....

 

[AboriginalBrotherhood1162]

You had not activated litchi iOS android beta!?.Or simply put follow me on tracking with....I put litchi as default a not a glitch in developer mode.....

 

[slup]

... Is there a trick how to ensure it is a good calibration?...

Keep the AC away from magnetic environments & objects (like belt buckles, watches, rings, bracelets), turn it smooth & in a constant low speed. Usually I stand up & myself move around in a lap with the AC in my hand with a fully stretched out arm ... this instead of rotating the AC alone which forces me to change the grip. Afterwards I check the DAT so the calibration was a good one.

You had not activated litchi iOS android beta!?..

Don't really understand what you're saying ... Litchi hadn't anything to do with this case. If you're wondering if Litchi create flight logs also ... then yes, below copied from this thread --> Mavic Flight Log Retrieval and Analysis Guide (under section 2.)

Mobile device third-party control app log files

Third-party control apps, such as Litchi, create their own custom log files in various formats using the DJI SDK, generally in CSV format. None that I have seen are as comprehensive as the DJI TXT log, let alone the DJI DAT log.

Since Litchi is probably the most common third-party app, it's worth noting that, under iOS and Android, Litchi creates its own CSV log and a standard DJI TXT log in a separate directory. No DAT file though.

 

[AnGo]

You had not activated litchi iOS android beta!?

Negative. Currently Litchi is not able yet to be used with Mavic Mini from all what I know.
I used DJI Fly App v1.1.6 and the current official firmware on the MM.

 

[AnGo]

Afterwards I check the DAT

How can you do that in the field?

a larger error still remained

What would be then a good value here?

 

[slup]

How can you do that in the field?


What would be then a good value here?

Can't do it in the field ... with the method I told you It's usually a successfully calibration & the error in the DAT log is 0.0 when I look. Rarely calibrate the compass on my MA1 but if the AC have a noticeable crabbing with only forward thrust I do it.

 

[AnGo]

Thank you very much for your valuable insights on the cause of this issue and how it can hopefully be prevented for the future. I will recalibrate both GPS and IMU with the suggested method and retry the same flight when there is an opportunity.

 

[hiflyer201]

. Rarely calibrate the compass on my MA1 but if the AC have a noticeable crabbing with only forward thrust I do it.

I'm requested to do a compass calibration or a controller calibration all the time with my Mavic 1. Most of the time it's accepted and I fly but I have the feeling that all is not as good as it should be and often it will request a controller calibration on the next flight. Thank you for helping me understand that not all calibrations are as good as they could be.

 

[AboriginalBrotherhood1162]

Negative. Currently Litchi is not able yet to be used with Mavic Mini from all what I know.
I used DJI Fly App v1.1.6 and the current official firmware on the MM.

Yes litchi functions with the beta test program, but you have to join the beta Cowboy club and discover the bugs and be one of the litchi mavic mini pioneers. There's a site online to get the Litchi app without money,but I can't remember the site at the moment

I'm requested to do a compass calibration or a controller calibration all the time with my Mavic 1. Most of the time it's accepted and I fly but I have the feeling that all is not as good as it should be and often it will request a controller calibration on the next flight. Thank you for helping me understand that not all calibrations are as good as they could be.

I prefer to do my compass calibration in the desert when the solar activity monitor/SAM, app shows the geomagnetic fields and solar x-rays are quiet.

 

[boblui]

I prefer to do my compass calibration in the desert when the solar activity monitor/SAM, app shows the geomagnetic fields and solar x-rays are quiet.

You just need a magnetic field of constant strength and direction to do it. The earth's magnetic field is used just for convenience. It's nothing more than making sure that the magnetometers of the three axes will produce the same output when put under the same magnetic field strength.

 

[AboriginalBrotherhood1162]

You just need a magnetic field of constant strength and direction to do it. It's nothing more than making sure that the magnetometers of the three axes will produce the same output when put under the magnetic field strength.

Exactly what I mentioned to the Australian chap yesterday.....