MAVIC MINI 2 CRASH AND BURN SOON AFTER FIRMWARE UPDATE - TOTAL LOSS - HELP!

[sar104]

Meta 4 Thanks for the reply - I took the liberty to quote a number of your replies below to confirm the data on this drone has you both confused and guessing with regard to the drones doomed flight. For this reason, I advised you that saved data is no more accurate than the processor itself which caused the crash . . .

In conclusion, you should always respect the thoughts of other pilots, expecially when your euvaluting data which seems to have you confused and uncertain.

Thanks - Mike
www.rotory.com


Meta 4 Comments:

That doesn't account for the altitude data being unbelievable

There's a motor error at 4:19 just a few seconds after the altitude data becomes unreliable.

A failed motor would be easily seen in the pitch, roll and yaw data if that was what happened, but the data doesn't support that.

There's a problem matching the drone's flight to your description of the incident.

But after that the altitude data doesn't make any sense.

I can only guess that there was a malfunction of the altitude sensor

No - @Meta4 is not confused, you simply don't understand what he is saying, and why he is saying it. See below.

You either experienced and motor/prop failure or just a good old glitch and lockup.

Here's a problem - you can't write technical gibberish like that and maintain any kind of image of someone who understands flight control algorithms. And none of your posts on this forum, nor anything on your late 80s-style website, suggest that you do.

Can any Mini 2 Experts tell me what happened?

The data show a progression of problems, starting at t = 256 s, when either the z-axis accelerometer or its data feed to the IMU/FC failed. That's becoming a bit of a pattern on the Mini 2 - this is the third or fourth I've seen reported on this forum. It's very valuable to have the entire video to compare to the telemetry, but note that time zero on the video is at t = 104 s on the flight data, and the flight data end at t = 276.4, 5 s before the video ends, when the aircraft motors shut down (OSD_isMotorOn ➛ FALSE) and the aircraft begins the left roll into an inverted attitude.



There are several indicators that it is the z-axis accelerometer. Firstly, the response of the x, y and z velocity data: since the aircraft tilt is small the most significant effect is on the z-axis velocity (blue trace), which immediately goes unphysically positive (downwards) with an acceleration that peaks at just over 3g. But the negative pitch and roll at that time, on a southerly heading, also result in positive excursions in the x (red) and y (green) velocities at the same time.

Secondly the relationship between the IMU fusion solution for height (OSD_height) and the time integral of the z-axis velocity. Note that OSD_height (solid black trace) rapidly decreases, but the negative integral of the z-axis velocity (dashed black trace) decreases even faster. That indicates that the barometric sensor was working correctly and offsetting the incorrect inertial height calculation.

The FC attempts to correct the non-existent rapid descent, and the aircraft climbs. The data indicate continued descent even under full power, which is what triggers the "not enough force" errors. The data that have little or no dependence on the z-axis accelerometer continue to report accurately until t = 270.8 s, when the roll calculation fails, which suggests x-axis rate gyro failure, but could be too many conflicting data leading to IMU fusion failure. The OSD_yaw data also start to deviate from reality at that time. The gimbal yaw data remain steady to the end of the data, but off by around 30°. It's apparent that the compass was out of calibration for that flight, although not enough to lead to uncontrolled flight.

The steady flight then continues for around 6 seconds until the FC shuts down the motors, the logged data end, while the video continues for another 5 seconds showing the aircraft inverting and falling to the ground.

Bottom line - this is flight control failure, almost certainly IMU accelerometer failure on the z-axis, although it is impossible to rule out IMU fusion failure. Either way it's a clear warranty issue if the aircraft is less than 12 months old.

 

[Mike Mas]

Hello All - Having worked with telemetry data for decades with military drones, while the data can be very accurate at times, it may also be very error ridden as well. This is especially true should any component on the machine short (stalled motor shorted board, etc.) which lowers processor voltage which may corrupt the recorded data.

Concerning other matters - when Myself and a few other voiced our thoughts on what we thought happened to the drone we were immediately confronted by Meta 4 that he already evaluated the cause and there was no reason for anyone to say otherwise. Regretfully, his comments clearly show a lack of understanding of the drones processors and their ability to provide data.

In any case, I hope the OP will be successful with DJI replacing the drone since the incident and video clearly show the pilot was not at fault.

Fly Safe Guys - Mike

 

[Meta4]

Hello All - Having worked with telemetry data for decades with military drones, while the data can be very accurate at times, it may also be very error ridden as well. This is especially true should any component on the machine short (stalled motor shorted board, etc.) which lowers processor voltage which may corrupt the recorded data.

The failure of one sensor doesn't invalidate the data from other independent sensors.
You've heard from three different sources that almost all of the data for this flight is reliable.
But somehow without reading the data, you know that everyone else is wrong?
If you had read the data, and you were right, you'd be able to point out where everyone else is wrong.

Concerning other matters - when Myself and a few other voiced our thoughts on what we thought happened to the drone we were immediately confronted by Meta 4 that he already evaluated the cause and there was no reason for anyone to say otherwise. Regretfully, his comments clearly show a lack of understanding of the drones processors and their ability to provide data.

You just don't know when to give up.
You haven't (probably can't) read the flight data.
If you did, you wouldn't keep on with this.
Your claims of vast experience just don't match the nonsense you've been posting.

 

[slup]

...In any case, I hope the OP will be successful with DJI...

You totally miss the intention with threads like this ... members post in this section for mainly 2 reasons:

1. To learn what was the cause & if it turns out to be pilot error, they want to know so they don't repeat it.
2. To get some backing and facts they can use in contact with DJI when doing a claim.

And a bonus for all other reading through the thread ...

3. Get knowledge so they don't do the same mistakes if it turns out to be pilot error ... or learn what can happen to these flying thing's even though using them by the book.

Unfounded guessing without confirming anything through the provided logs on hand, place that work on others that actually look into the logs ... & will not give the OP anything valuable.

We have seen guesses without pointers to neither the video or the log about:

Bad GPS & RC reception
ESC failure
Low battery levels
Prop/Motor failure
Good old glitch and lockup

All of these have been explained to be unlikely ... & a plausible cause have been presented, which fulfills the OP's 2 needs.

If you want to be anything near trustworthy ... put in more time to explain, out from the material the OP have provided (video+log), why your hypothesizes might line up... instead of waving with your web page and, your self claimed skill & experience.

 

[Mike Mas]

Slup thanks for your reply and important notes to help the OP!

It appears a shorted component, or shorted / stalled motor, might have lowered the processors voltage causing the crash and resulting erroneous data. Another possibility is; due to the voltage drop, a re-boot might have caused the final death dive. Making matters worse, the user recently upgraded the firmware which could also be an underlined cause.

Meta 4 thanks for your reply as well - Please re-visit my post outlining your comments (#38) on the data, as they clearly show confusion and guessing on your part regarding the data. For this reason, I suggest you avoid discouraging others from voicing their thoughts on the topic. Everyone's option is valuable in an open forum.

Hopefully, DJI will offer the owner an opportunity to replace the drone since it clearly shows the pilot was not at fault.

Fly Safe Guys - Mike

 

[sar104]

Slup thanks for your reply and important notes to help the OP!

It appears a shorted component, or shorted / stalled motor, might have lowered the processors voltage causing the crash and resulting erroneous data. Another possibility is; due to the voltage drop, a re-boot might have caused the final death dive. Making matters worse, the user recently upgraded the firmware which could also be an underlined cause.

No - just stop posting this nonsense. It does not remotely appear that a shorted component (whatever you imagine that might even mean) or a faulty motor caused this. In fact it is perfectly clear that wasn't the cause. There were no motor problems - the FC simply shut them down. The cell voltages were within spec for the entire flight. There was no reboot - the FC recorded the motor shutdown. The faulty data that caused this were recorded, as you would know if you could be bothered to look and/or understand what it means.

Meta 4 thanks for your reply as well - Please re-visit my post outlining your comments (#38) on the data, as they clearly show confusion and guessing on your part regarding the data. For this reason, I suggest you avoid discouraging others from voicing their thoughts on the topic. Everyone's option is valuable in an open forum.

Ah yes - the inevitable claim that your ignorant speculation is just a valid as detailed data forensics. I prefer a different approach - when I have no useful knowledge on a subject I control my urge to try to lecture on it and I keep quiet.

 

[Mike Mas]

Hello all and thanks for the replies - Meta 4 I went ahead and listed some addition comments you made which clearly show you were totally confused regarding the data and offered no help to the OP. When I and others "offered possible legitimate causes" you responded: "No need for guessing, the data has been analysed and the incident explained " this is comical, because your the only one guessing. Re-read you own comments they are self explanatory!

With that, let's move on and stay with the subject matter and keep the thread informative and helpful to others!

Best Regards - Mike

Meta 4 Comments:

I couldn't tell if they are genuine errors or a result of the IMU malfunction.

The GPS messages are irrelevant to the incident

That doesn't account for the altitude data being unbelievable

There's a motor error at 4:19 just a few seconds after the altitude data becomes unreliable.

A failed motor would be easily seen in the pitch, roll and yaw data if that was what happened, but the data doesn't support that.

That doesn't account for the altitude data being unbelievable

There's a problem matching the drone's flight to your description of the incident.

But after that the altitude data doesn't make any sense.

I can only guess that there was a malfunction of the altitude sensor

 

[clackey]

Hey all - cool it please. We can disagree but let’s keep it friendly.

 

[Meta4]

Hello all and thanks for the replies - Meta 4 I went ahead and listed some addition comments you made which clearly show you were totally confused regarding the data and offered no help to the OP. When I and others "offered possible legitimate causes"

The trouble Mike, is that nothing you've suggested was legitimate.
Ignoring everyone who explains that, doesn't change it.
No matter how many times it's said, it's still wrong.

 

[sar104]

The trouble Mike, is that nothing you've suggested was legitimate.
Ignoring everyone who explains that, doesn't change it.
No matter how many times it's said, it's still wrong.

I think that you are wasting your time. It looks to me that he is either doing a quite remarkable job of pretending that he hasn't seen my detailed assessment in post #41, or he has me on his ignore list and is avoiding the awkward evidence that way. Either way, at this point he has turned it into a trolling exhibition rather than a legitimate exchange of views.

 

[Yaros]

But you clearly aren't a guy that takes advice from anyone.

You don't need to be so rude!
Why don't you treat everyone with respect, doesn't matter if they are wrong, or right!

I'd suggest that you should respect the evaluation from flyers that can and do read the flight data.

And I'd suggest you stop talking this way.

If you want to help further ... don't.

Here everyone tries to help, I'm sure that the OP is fine with that, every opinion counts, no matter if it's right or wrong by your opinion. We can't say 100% what caused this incident, but we can deduct some causes from the logs and data. You don't need to insult other people and say that they don't know how to read the logs because I'm sure you wouldn't like that done to you, right?

 

[Meta4]

You don't need to insult other people and say that they don't know how to read the logs

Did you even read the thread?

The person you are talking about came in after two data analysts had explained the incident and without reading any flight data, suggested a guess which had already been explained as something that had not occurred.

because I'm sure you wouldn't like that done to you, right?

That's what he did .. six times.
Three data analysts took the time to explain in detail what happened and why his suggestion was not correct.
But he came back six times, each time ignoring the explanations that were given, saying the the people that actually read the data and explained it didn't know what they were talking about able to support his claim with anything from the data.

 

[dirkclod]

Ok guys @clackey said to cool it but see you guys just keep on.
I think all has been said that can be said so unless Clackey wants
it reopened it’s closed.

CLOSED​