First flight crash

[Enrico BBB]

Ok, waiting for the guys then.

Meanwhile I'll buy a new one cause I can't wait til February...

It lasts two month just to have their point of view. .. then working time to replace pieces, shipping back.... I'll get old

 

[Fly Dawg]

Ok, waiting for the guys then.

This appears to be an IMU issue. There were velocity and positioning errors as the aicraft began to speed up. SAR104 can give you a more detailed response on the IMU issue, however he will most likely want the device .dat file as well. The file you will need is FLY007.DAT. To locate this, see this link: Retrieve a V3 DAT From Mobile Device

Below are the errors:

 

[Enrico BBB]

This appears to be an IMU issue. There were velocity and positioning errors as the aicraft began to speed up. SAR104 can give you a more detailed response on the IMU issue, however he will most likely want the device .dat file as well. The file you will need is FLY007.DAT. To locate this, see this link: Retrieve a V3 DAT From Mobile Device

Below are the errors:

View attachment 87897

Thank you
What could be the reason of IMU issues?

 

[Fly Dawg]

What could be the reason of IMU issues?

That is why the .dat file will be needed to further investigate.

 

[Prismatic]

... What could be the reason of IMU issues?

If memory serves me--having watched threads like this for some time--IMU issues are almost always hardware failures, not user/pilot errors. But you'll still want to post the DAT file if you can, because I've already said more than I know. ?

 

[Enrico BBB]

is this the file we need?

 

[slup]

?
Keeping my fingers crossed that they find some clear indication of HW failure, easy business with DJI after that ... will be interesting to hear what they find.

 

[Fly Dawg]

is this the file we need?

Thats is the correct file.

 

[Enrico BBB]

?
Keeping my fingers crossed that they find some clear indication of HW failure, easy business with DJI after that ... will be interesting to hear what they find.

All that I can say is that drone suddenly got crazy and I couldn't do anything.... 4 seconds from life to death.... Is a really bad sensation when you see your drone flying like crazy and crashing while the controller become completely unuseful ..
I thank you all for your precious help and I will keep you updated when I receive an answer by Dji... Dealer told me that this could last some months.... Meanwhile I will buy a second Mavic Mini hoping in a better luck

Thank you all for helping me ???

 

[BudWalker]

pardon my typing, it's one-handed.

from the .txt, the video, and, now the .DAT it's almost certainly the case that the incident was caused by a compromised yaw at launch. the MM was actually facing SW yet the yaw value was NE. in turn, the compromised yaw was caused by incorrect magnetometer data at the launch site.

i've been working on the theory that the geomagnetic distortion was caused by surrounding buildings - not the small area surrounding the launch site. there a couple of discrepancies that cast some doubt but i still like this theory.

the magMod value (i.e. the total magnetic field strength) has some correlation with height. this would be expected if the surrounding buildings were distorting the geomagnetic field


there is also something weird about the magY data. at time 28.5 secs there is heading fluctuation seen in the magYaw data that's caused by a fluctuation in the magY data. however, it's not seen in the gyroZ data or the video.


following that the MM rotates CCW as seen by the gyroZ data and the video. but the magYaw rotates CW.

maybe @sar104 will have more to say when he returns

I have never seen a car completely not responding at any input, gaining speed by itself.

Now, a drone, due to 1.000reasons can get out of control BUT, it was in C mode so it can't reach 40mph as it happened ok?

It's not allowed to change directions and to gain speed. If something go wrong it has to turn on a safety mode which do not allow it to raise speed. It has to stand where it is and to land slowly, even if in the ocean, in a forest or wherever. If it goes like a crazy rocket could kill someone... It's not a pilot matter... I could have done anything to stop it. How could be a crash on someone's face at 40mph? Dji has a security problem. How can a user forecast this kind of things? How could I imagine that if I forget a setting I could run a rocket in the sky? Please ...

i get it. if they can send a man to the moon.....

but, there are technical trade offs that prevent knowing when the AC is in a fly away.

but, IMHO DJI is derelict because the most effective, easiest, robust method to prevent a fly away like yours doesn't appear in any DJI documentation. that method is to check the red triangle heading indicator in the map display to see if it matches the actual AC orientation. if you had known about this you would've seen that these didn't agree.

 

[Fly Dawg]

@BudWalker For what it's worth. At the point of the magyaw fluctuation, there was a slight change in Roll. ~5 degrees. It looks like the Roll occurred before the magyaw fluctuation but they do co-incide and they are very close. As I said, it probably doesn't mean anything, but I dont see anything else that obviously co-incides.

 

[Prismatic]

... So do you think I won the Certificate of Idiot 2019 and I won't be covered by warranty?

I can’t address your warranty question, but I can say with confidence—having stayed current with this forum all year—that you’re not even in the top 5 contenders for that award! ?

 

[Enrico BBB]

pardon my typing, it's one-handed.

from the .txt, the video, and, now the .DAT it's almost certainly the case that the incident was caused by a compromised yaw at launch. the MM was actually facing SW yet the yaw value was NE. in turn, the compromised yaw was caused by incorrect magnetometer data at the launch site.

i've been working on the theory that the geomagnetic distortion was caused by surrounding buildings - not the small area surrounding the launch site. there a couple of discrepancies that cast some doubt but i still like this theory.

the magMod value (i.e. the total magnetic field strength) has some correlation with height. this would be expected if the surrounding buildings were distorting the geomagnetic field
View attachment 87960

there is also something weird about the magY data. at time 28.5 secs there is heading fluctuation seen in the magYaw data that's caused by a fluctuation in the magY data. however, it's not seen in the gyroZ data or the video.
View attachment 87959

following that the MM rotates CCW as seen by the gyroZ data and the video. but the magYaw rotates CW.

maybe @sar104 will have more to say when he returns

i get it. if they can send a man to the moon.....

but, there are technical trade offs that prevent knowing when the AC is in a fly away.

but, IMHO DJI is derelict because the most effective, easiest, robust method to prevent a fly away like yours doesn't appear in any DJI documentation. that method is to check the red triangle heading indicator in the map display to see if it matches the actual AC orientation. if you had known about this you would've seen that these didn't agree.

Thank you so much for this detailed analista... Very good job ...
I have a last question cause I am not so technical.... It's an internal fault or my personal fault? Thank you again for yup

I can’t address your warranty question, but I can say with confidence—having stayed current with this forum all year—that you’re not even in the top 5 contenders for that award! ?

So, drone crashed and no award? What a bad life ???

 

[bigswifty]

but, IMHO DJI is derelict because the most effective, easiest, robust method to prevent a fly away like yours doesn't appear in any DJI documentation. that method is to check the red triangle heading indicator in the map display to see if it matches the actual AC orientation. if you had known about this you would've seen that these didn't agree.

For many users, including me for a long while, the red triangle *never* matched the AC orientation. And the drone flew and behaved otherwise normally.

 

[DanMan32]

I agree in checking orientation, but in relationship to what?

If one went by the map or attitude indicator, considering top of map or indicator was related to the direction you're facing, that could be inaccurate as the mobile compass may likely be inaccurate. My Androids usually are until I do the compass shake.

Google Maps usually are correctly oriented especially when you're moving because it can use GPS delta.

 

[slup]

For many users, including me for a long while, the red triangle *never* matched the AC orientation...

I agree in checking orientation, but in relationship to what? ...

The drone "triangle" in relation to the objects in the map ...



I'm above just using the picture from the AirData.com analyze of the flight log in this thread for demo purposes as it have map objects & a drone triangle...

In all these 3 scenarios the drone is pointing to a light colored car & are parallel to that darker play ground ... if this is similar for my eye in reality as a pilot I know that the drone aligns correctly. Now why 3 different pics with different angles? This can occur if your mobile device doesn't have the heading correct & needs to be calibrated by need to do the "compass shake" mentioned. But the message in all 3 cases are clear anyhow.

 

[BudWalker]

@BudWalker For what it's worth. At the point of the magyaw fluctuation, there was a slight change in Roll. ~5 degrees. It looks like the Roll occurred before the magyaw fluctuation but they do co-incide and they are very close. As I said, it probably doesn't mean anything, but I dont see anything else that obviously co-incides.

View attachment 87961

Good call.
That shouldn't happen. Roll and pitch are used to correct the mag data - absent any actual rotation magYaw should've remained constant.

Looking into this i realized
1) that the magY data passed through 0 during the fluctuation. and,
2) the fact that the MM was facing east so the geomagnetic field would have been aligned with the Y axis, and
3) the MM was very close to vertical.

These taken together means that either
A) the geomagnetic field inclination was very close to vertical - very different from the 62° in that region. This distortion would likely be the result of the surrounding buildings. At launch the declination was 80° according to the mag data. or
B) the compass wasn't calibrated correctly. in particular the magY bias value was incorrect, or
C) i'm sure @sar104 can come up with other scenarios

I'm going with the theory that the launch area, not just the launch point was geomagnetically distorted.

 

[BudWalker]

For many users, including me for a long while, the red triangle *never* matched the AC orientation. And the drone flew and behaved otherwise normally.

Maybe you're referring to the "radar" display? I should've been a bit more specific. i'm referring to the heading indicator in the map display, not the "radar" display


The radar display can also be used but only if the tablet compass is calibrated correctly. Also, the tablet has to be aligned with the AC.

EDIT; just saw @DanMan32 made this point.

 

[BudWalker]

Thank you so much for this detailed analista... Very good job ...
I have a last question cause I am not so technical.... It's an internal fault or my personal fault? Thank you again for yup

Sorry, i'm going to pass on determining fault. The actual cause of the incident is still in question. As i said, the real answer is for DJI to document the heading check i described.

 

[Prismatic]

I'm going with the theory that the launch area, not just the launch point was geomagnetically distorted.

That would suggest that no flight launched from that location would ever be safe.

I've always thought that magnetic distortion was quite localized--I mean centimeter range distances, a meter or two at most.

If your drone demands compass calibration, they say, you should first try moving your launch point a short distance. Hand launching, I've heard, is always safe from disruptive magnetic fields (providing you're not wearing or carrying something magnetic).

Is that not true? If the surrounding structures were the source of the disrupted geomagnetism, hand launching there would offer no advantage.