Mavic Mini Flyaway! Help ?

[sar104]

The immediate cause of the incident was probably the 30° difference between the compass and the FC's Yaw value at launch. The geoDeclination at the launch site is -3° so the Yaw/magYaw difference should have been 3°.

The .DAT begins 100 secs before the flight starts. At -100 secs the Yaw/magYaw difference was about 72°. At -93 secs it appears that the MM was picked up, turned and then placed about .7 meters lower. This allowed the Yaw/magYaw separation to be partially reconciled to the 30° seen at launch.
View attachment 87503
Why was the Yaw/magYaw separation 72° at the start of the .DAT? It can be seen from the clock value (tick#) that the MM had been powered up for roughly 24 secs prior to the start of the .DAT. I suspect that the MM was powered up in a geomagnetically distorted location resulting in both Yaw and magYaw being set to the same incorrect value. The MM was then moved to a less distorted location which resulted in the Yaw/magYaw separation seen when the .DAT started recording.

I still think something else is going on here. Look at the magnetic modulus:



It's all over the place as the aircraft takes off.

And comparing the total gyro Z with the unwrapped magnetic and IMU yaw values shows that the IMU yaw is going very wrong:



Any thoughts?

 

[skiptv]

The worse thing DJI ever did, the Mavic Mini.... Just when laws, regulations and drone crashes were "thinning the herd" we now have an entire new batch of "pilots" lol to work through.

 

[BudWalker]

I still think something else is going on here. Look at the magnetic modulus:

View attachment 87510

It's all over the place as the aircraft takes off.

And comparing the total gyro Z with the unwrapped magnetic and IMU yaw values shows that the IMU yaw is going very wrong:

View attachment 87511

Any thoughts?

But, the flight only lasted 17 secs or so. During that interval the magMod was OK. Not great, but OK. There was a dip when it hit the roof at 11.584 secs and then the sharp excursions at 17 secs when it hit the ground.


@Smgy94 can you tell us anything about the 24 secs before the .DAT started recording. Where it was powered on and was it moved before it was moved a second time seen at the beginning of the .DAT.

 

[sar104]

But, the flight only lasted 17 secs or so. During that interval the magMod was OK. Not great, but OK. There was a dip when it hit the roof at 11.584 secs and then the sharp excursions at 17 secs when it hit the ground.
View attachment 87535

@Smgy94 can you tell us anything about the 24 secs before the .DAT started recording. Where it was powered on and was it moved before it was moved a second time seen at the beginning of the .DAT.

Maybe so. I'm just not used to seeing that much variation during flight, and with a yaw error of only. 30° I'd have expected minor toilet bowl symptoms at most. Perhaps the Mini is more sensitive.

 

[BudWalker]

Maybe so. I'm just not used to seeing that much variation during flight, and with a yaw error of only. 30° I'd have expected minor toilet bowl symptoms at most. Perhaps the Mini is more sensitive.

Yeah 30° isn't that much. My flying buddy had a fly away incident with his P3 where there was 30° Yaw/magYaw separation.

 

[Deleted member 111775]

The worse thing DJI ever did, the Mavic Mini.... Just when laws, regulations and drone crashes were "thinning the herd" we now have an entire new batch of "pilots" lol to work through.

Nice Bandana

 

[ThorDar]

I see from the GPS coordinates that it looks like you launched from between two houses. If that's the case, my guess is that even though you showed 11 satellites, the signals were inaccurate. There's something called multipath errors that occurs when a GPS receiver is too close to obstacles. Some of the signals it will be receiving will be reflected off the nearby obstacles thereby causing incorrect calculations. To do accurate GPS calculations, the drone needs a clear view of the sky 360 degrees and as close down to the horizon as possible, ie, in an open field, especially on launch, so there's a direct line to to as many satellites as possible.

Also GPS satellites are not stationary. They circumnavigate the globe so the number and position of satellites in view changes from moment to moment. This is why you may get away with launching from a bad spot one time but not the next.

Once up in the air and clear of obstacles on the ground, this is much less of an issue.

 

[GerdS]

In addition to GPS the MM has its optical ground sensors to avoid jumping around with GPS signal reflection. And if these would not work it would be extremely dangerous flying a MM inhouse, where GPS reception is partly possible, too.

 

[lorenztanjoco]

how to upload the mavic mini flight log? i would like to share mine... i lost my mini, cause i cant control it.

 

[lorenztanjoco]

Hey Guys,

Flight log:

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

I pressed auto-take off on my Mavic Mini

Then right after it immediately flew straight forward into a house (it moved by itself)

It then bounce off the house and flew at 39mp/h the opposite direction and crashed into a field.

There was only mild winds, and I had satellite connection. I'm confused as to why this happened ?

I retrieved the Mavic Mini and it has some big scrapes on it, but it is in tact.

The gimbal seems slightly off though, even after calibration. The horizon does not look level.

I purchased my Mavic Mini within the past 15 days and I have contacted DJI in regards to getting a replacement under DOA manufacturer fault.

I have DJI Care Refresh but I do not want to claim through it as I believe this was a defective units fault and not pilot error.

Could anybody analyse my flight and see what they think caused this?

Thanks so much!

S

how did you upload your flight logs sir? i also lost my maic mini

 

[Deleted member 111775]

*UPDATE*

DJI are sending a brand new Mavic Mini!

 

[new2mavic]

I see from the GPS coordinates that it looks like you launched from between two houses. If that's the case, my guess is that even though you showed 11 satellites, the signals were inaccurate. There's something called multipath errors that occurs when a GPS receiver is too close to obstacles. Some of the signals it will be receiving will be reflected off the nearby obstacles thereby causing incorrect calculations. To do accurate GPS calculations, the drone needs a clear view of the sky 360 degrees and as close down to the horizon as possible, ie, in an open field, especially on launch, so there's a direct line to to as many satellites as possible.

Also GPS satellites are not stationary. They circumnavigate the globe so the number and position of satellites in view changes from moment to moment. This is why you may get away with launching from a bad spot one time but not the next.

Once up in the air and clear of obstacles on the ground, this is much less of an issue.

View attachment 87720

I think this is the most sensible explanation. I have experienced exactly the same phenomenon on my Mavic 2 Pro about a month ago. I was hovering the drone between tall buildings. I didn't pay attention to the GPS signal but the drone could enter GPS mode. It was OK at the beginning but then the dronze suddenly flew backward steadily with no stick input. I applied reverse control immediately and landed it successfully. Then I launched it again and the same thing happened but it flew sideway this time. Subsequently I moved the drone to another location and the problem disappeared. In fact it never happened again since then so I am convinced that it's not the drone's fault.

I noticed that the OP did apply throttle control and the drone responded so I believe the dronze was still under control. Had the OP applied appropriate elevator / aileron controls the crash might have been avoided.

 

[sar104]

I think this is the most sensible explanation. I have experienced exactly the same phenomenon on my Mavic 2 Pro about a month ago. I was hovering the drone between tall buildings. I didn't pay attention to the GPS signal but the drone could enter GPS mode. It was OK at the beginning but then the dronze suddenly flew backward steadily with no stick input. I applied reverse control immediately and landed it successfully. Then I launched it again and the same thing happened but it flew sideway this time. Subsequently I moved the drone to another location and the problem disappeared. In fact it never happened again since then so I am convinced that it's not the drone's fault.

I noticed that the OP did apply throttle control and the drone responded so I believe the dronze was still under control. Had the OP applied appropriate elevator / aileron controls the crash might have been avoided.

That's much more likely to have been caused by magnetic interference at the launch location than by GPS issues.

 

[new2mavic]

That's much more likely to have been caused by magnetic interference at the launch location than by GPS issues.

Can you elaborate more on that pls ? multipath can cause drift in GPS location although magnetic interference can also confuse the drone. Why is the latter the more likely cause ?

 

[sar104]

Can you elaborate more on that pls ? multipath can cause drift in GPS location although magnetic interference can also confuse the drone. Why is the latter the more likely cause ?

Yes - to clarify - I was not suggesting that the multipath problem doesn't potentially exist. My comment was an observation based on looking at the flight logs from large numbers of such events - they are almost always due to magnetic interference leading to an incorrectly initialized IMU yaw value. That causes uncontrolled flight when the FC applies thrust in the wrong direction to attempt to hold position or fly in a particular direction.