My Mavic Air falled from the SKY! Can anyone explain why?

[sar104]

The throttle stick was indeed held down at the end of the recorded flight; the AC descended from a height of 48.6m to 25.2 and then the log ends. Hard to say it landed normally at 25m above ground.

View attachment 90756

The problem was partly the usual one in these conditions - the VPS height sensor was confused by the fog. At the end of the flight as it was descending the VPS height went to zero and the FC concluded that it was on the ground (OSD_groundOrSky = "Ground") and shut down the motors (OSD_isMotorUp = "False").



That shouldn't have happened though, because the IMU was still reporting a vertical descent speed of 2.5 m/s. It appears to have been complicated by "motor blocked" flags and "not enough force" flags:



Unwise to fly in those conditions, but that was an aircraft fault.

 

[Deleted member 94047]

@sar104 - wouldn't the flyCState change to "Landing" or "ForceLanding" if that was the case? Or am I missing something as usual?

 

[slup]

@sar104 ... how to explain the sudden un-commanded Yaw in the very end (isMotorUpFalse comes later)?

 

[sar104]

@sar104 ... how to explain the sudden un-commanded Yaw in the very end (isMotorUpFalse comes later)?

View attachment 90768

There wasn't much yaw at the end, and the pitch and roll excursions are more interesting.



My first guess would be that the props were icing up.

 

[sar104]

@sar104 - wouldn't the flyCState change to "Landing" or "ForceLanding" if that was the case? Or am I missing something as usual?

Not if it unexpectedly finds itself on the ground - i.e. no descent recorded for several seconds.

 

[sar104]

It would be interesting to take a look at the mobile device DAT file for this flight.

 

[slup]

Sorry do nag you @sar104 ... but just want to know what I'm missing & why your Igor graphs look so different then those in CsvView.
The biggest movements I see (or wrongly interpret it to be) are the Yaw compered with the Pitch & Roll.



Then also a question regarding your previous post (just for me to bump up my knowledge level )

Question: So by still having a vertical movement together with a Barometer altitude > 0 should mean that VPS altitude are of lower priority ... is it therefore you conclude that it's a AC fault?

...and shut down the motors (OSD_isMotorUp = "False").

...That shouldn't have happened though, because the IMU was still reporting a vertical descent speed of 2.5 m/s. It appears to have been complicated by "motor blocked" flags and "not enough force" flags:

...that was an aircraft fault.

 

[Skywatcher2001]

The flight record (as far as I understand) shows three issues, two of them seem to be related:
1) You get a warning that the batterytemperature at start is to low and you should warm the battery. This at first seems to be a not so critical information but...
2) During the flight you have several severe battery problems. i.m.o. they are related to the warning at the start: Your battery was to cold.
and then the final and fatal error to the end of the log:
3) There are three warnings of an obstructed motor. Though not directly related to the battery temperature warning, this warning gives a good indication why you crashed: Most likely is, what was stated before by the other comments, that you had ice build up on your props and possibly in the motor. This, combined with a battery that had some issues only moment before, might have led to the crash.

Conclusion: The hardest conclusion in tihs is, that it was a clear pilot error. You ignored the warning and if the only damage after all indeed is two damaged props, you are very lucky. The idea of drying the drone after this incident is good, as the ice melts, water comes and even if it is not much, it might cause further damage.

I could not have said it any better. Spot on . . .

 

[GregE]

I’ve been caught in icing conditions several times flying a UH 1. Each time reminds you that it is better to be on the ground wishing you were flying than the other way around

 

[hiflyer201]

Thank you for your reply. I know it is not the optimum condition to fly but I think it should be smart enough to warn like "don't take off" or after take off "low tempeture, emergency landing" in cold weather conditions etc.

The thing I wonder actually, is flight records tell the exact reason?

The drone fly's because of lift from the props. The prop is an airfoil. It produces lift. Ice which can form rapidly in the conditions that you took off in destroys lift. It's as simple as that. Extreme heat and cold can cause numerous other problems.

 

[sar104]

Sorry do nag you @sar104 ... but just want to know what I'm missing & why your Igor graphs look so different then those in CsvView.
The biggest movements I see (or wrongly interpret it to be) are the Yaw compered with the Pitch & Roll.

View attachment 90772

That's because you are looking at wrapped yaw, modulo 360. So if the yaw changes from -179° to 179° that looks huge on the graph until you take into account that 180° and -180° are the same yaw, and it's actually only a 2° change. That's why I always unwrap the yaw signal for these kinds of analyses.

Then also a question regarding your previous post (just for me to bump up my knowledge level )
Question: So by still having a vertical movement together with a Barometer altitude > 0 should mean that VPS altitude are of lower priority ... is it therefore you conclude that it's a AC fault?

The FC should not rely just on the VPS data, and generally doesn't. I've seen lots of cases where full down throttle with a VPS height of less than 3 meters leads to the aircraft entering forced landing mode, which is appropriate, but this is the first time I've seen the FC shut down the motors while still measuring a positive descent rate. As I mentioned, it may be related to the other error flags, but it still seems like a fault. That's why we need to see the DAT file, which will explicitly give the motor stop reason.

 

[slup]

That's because you are looking at wrapped yaw, modulo 360. So if the yaw changes from -179° to 179° that looks huge on the graph until you take into account that 180° and -180° are the same yaw, and it's actually only a 2° change. That's why I always unwrap the yaw signal for these kinds of analyses.



The FC should not rely just on the VPS data, and generally doesn't. I've seen lots of cases where full down throttle with a VPS height of less than 3 meters leads to the aircraft entering forced landing mode, which is appropriate, but this is the first time I've seen the FC shut down the motors while still measuring a positive descent rate. As I mentioned, it may be related to the other error flags, but it still seems like a fault. That's why we need to see the DAT file, which will explicitly give the motor stop reason.

Great explanations @sar104 ... much appreciated, that with the Yaw angle were obvious when you said it
Imagine that a long experience around how DJI's strategies usually are is very valuable when it comes to draw the
right conclusions for a event.

Now let's hope we can get our hands on the DAT file ... I'm not so much for speculations, nothing beats pure facts

 

[m80116]

@sar104 what do you think it would happen with a Mavic Mini: drone above heavy cloud of fog.

In that case what would you do if you HAD to fly anyway ?

 

[new2mavic]

@sar104 what do you think it would happen with a Mavic Mini: drone above heavy cloud of fog.

In that case what would you do if you HAD to fly anyway ?

Just don't hold the throttle stick fully down for a prolonged period of time. Release it momentarily every 2 seconds or so.

 

[Pietros]

I’ve been caught in icing conditions several times flying a UH 1. Each time reminds you that it is better to be on the ground wishing you were flying than the other way around

Nothing else in the sky sounds like a UH1.

 

[sar104]

@sar104 what do you think it would happen with a Mavic Mini: drone above heavy cloud of fog.

In that case what would you do if you HAD to fly anyway ?

Fog fools the VPS, as noted above. Flying above the clouds/fog is out of VLOS and therefore generally illegal. Why would you have to do that?

 

[m80116]

Because of... drones. If I have a cloud of fog above me I might need to poke a hole and get out of it, see what's in the distance. It's the entire point of having a drone. How many times we've seen Mavics doing things that aren't supposed to do like going out while it snows.

Besides CSC, User Manual p.39. I've set my [Fly app] Emergency propeller stop in Safety > Advanced Safety settings to only in case of emergency. Should you configure it otherwise I do believe you might risk a prop halt during descent if VPS falsely detects ground.

 

[sar104]

Because of... drones. If I have a cloud of fog above me I might need to poke a hole and get out of it, see what's in the distance. It's the entire point of having a drone. How many times we've seen Mavics doing things that aren't supposed to do like going out while it snows.

Besides CSC, User Manual p.39. I've set my [Fly app] Emergency propeller stop in Safety > Advanced Safety settings to only in case of emergency. Should you configure it otherwise I do believe you might risk a prop halt during descent if VPS falsely detects ground.

I'm sorry - I can't help you further if your attitude is that the law is there to be broken just because you feel like it.

 

[GregE]

NOt sure what the specific drone requirements are for Italy but ICAO standardizes aviation rules for all countries.

 

[Olav van Gerven]

Because of... drones. If I have a cloud of fog above me I might need to poke a hole and get out of it, see what's in the distance. It's the entire point of having a drone. How many times we've seen Mavics doing things that aren't supposed to do like going out while it snows.

I remember being young, many years ago, as I did something stupid. I was asked what on earth made me think I could do it and I answered "Because the others do it too...". I was convinced that was a valid reason, until the policeman at the other side of the table asked me "Do you really think that because other people do it, something is automaticly right? Are you able to think yourself and evaluate what the consequences of your actions might be? Let me help you, have you thought of..." and then he listed a lot of things that could had happened to me, as it had happened to "the others" I refered to.

My question to you now is: Do you really think that the fact that other people put their property, themself and other in danger by their actions is less important then the fact that they want to have some "cool video"?