Please help ! Mavic Air makes a dangerous spontaneous maneuvers

[djiuser_MLPTTHpRA3yw]

Hi everybody !
My Mavic Air makes any unexpected maneuvers on the flight ! He does this from time to time for several months. It just starts to slip somewhere and then it doesn't respond to the RC remote in my hand ! Only RTH button has so far been rescued. This RTH button has so far been the only resort. Today, however, it went too long and could have lost the drone . I've tried to find the error and already calibrated IMU many times and calibrated the compass but nothing has changed. In display no faults are shown before and during the flight. Finding a complicated defect is not always the case ! ( Sorry I'm just writing through google translation )
Add files to this defective flight here today and video where to see that i joysticks don't move when it happens .

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

I am very concerned and hope that the solution is still there I'd like to add the information you don't see in this video : RTH flight was right corner side in front not front side ?!

 

[Starz]

What app are you using?

 

[sar104]

IMU0, which was the active IMU, is faulty. Both the pitch and roll data go bad when you centered the elevator from full forwards at 675 seconds:





That looks like a hardware problem.

 

[Meta4]

I am very concerned and hope that the solution is still there.

A quick look at your flight data shows you flying south for 1.5 km at 123 metres and initiating RTH.
Your RTH was hands-off with no joystick inputs.
It looks like there was a headwind for the return flight and once things steadied, speed was approx 7.5 m/s.

There are some irregularities in the data soon after initiating RTH at 11:19.8.
Although the drone was already well above the set RTH height of 70 metres, it began to climb and descend irregularly until about 11:47.
The altitude varies between 95-130 metres, first climbing then descending and climbing again.
The speed and heading also vary strangely despite no joystick input.
From about 11:50 the speed, altitude and heading stabilise to the expected values and the drone comes straight home normally.

Wind could not possibly cause what the data seems to show.

Some of the recorded data is not believable and cannot be real.
The most obvious example is the instantaneous speed jump from 0.3 m/s to 11.1 m/s at 11:21.7. (red arrow in the graph)
That isn't possible and suggests that at least some of the other data irregularities is also false.
A GPS glitch might explain off speed and position information but the altitude data comes from the barometer and wouldn't be affected by any GPS problems.
The IMU malfunction mentioned in the previous post sounds like a likely explanation.

 

[hewo]

he has got loss of control incident.
usual course of examine is chk(software) firmware if corrupt by "rewriting err refreshing" and affirm success for the devices five each (drone, rc, battery, cellphone, windowsPc)

mine had this strange Tasmanian spinning devil upon landing. it lands by sticks intended but once the sonic sensors said grounded the propellers did their finale before turning off. it was this finale after touchdown that is called loss of control incident. the composite skids elevating the piezo sensor faces off the ground do show concrete marring owed to that dance.
i duplicated that landing incident to make certain of discrepant condition.
yes, chk obvious physical defects of devices.
but rewriting clears the issue.
what this means is the devices are not retaining their soft.
somehow the flash memory is getting changed or the writing never was truly successful as stated successful.
it is a bummer.
do we have to reinvigorate the cellphone operating system and applications or do they stay fully intact as when factory issued them?

 

[sar104]

A quick look at your flight data shows you flying south for 1.5 km at 123 metres and initiating RTH.
Your RTH was hands-off with no joystick inputs.
It looks like there was a headwind for the return flight and once things steadied, speed was approx 7.5 m/s.

There are some irregularities in the data soon after initiating RTH at 11:19.8.
Although the drone was already well above the set RTH height of 70 metres, it began to climb and descend irregularly until about 11:47.
The altitude varies between 95-130 metres, first climbing then descending and climbing again.
The speed and heading also vary strangely despite no joystick input.
From about 11:50 the speed, altitude and heading stabilise to the expected values and the drone comes straight home normally.

Wind could not possibly cause what the data seems to show.

Some of the recorded data is not believable and cannot be real.
The most obvious example is the instantaneous speed jump from 0.3 m/s to 11.1 m/s at 11:21.7. (red arrow in the graph)
That isn't possible and suggests that at least some of the other data irregularities is also false.
A GPS glitch might explain off speed and position information but the altitude data comes from the barometer and wouldn't be affected by any GPS problems.
The IMU malfunction mentioned in the previous post sounds like a likely explanation.

Those speed data are from IMU1, and are spurious. I was wrong in my previous post - IMU1 was the active IMU and the IMU0 attitude excursions were real - hence the "not enough power" messages. If you compare the data from IMU0 you can see what the speed was really doing:




The IMU0 green trace is the real speed - the OSD (txt log) speed (blue) was coming from IMU1 (red). So it is IMU1 that is faulty.

 

[RayOZ]

First question that comes to mind, Are you using genuine DJI props?

 

[brett8883]

Those speed data are from IMU1, and are spurious. I was wrong in my previous post - IMU1 was the active IMU and the IMU0 attitude excursions were real - hence the "not enough power" messages. If you compare the data from IMU0 you can see what the speed was really doing:

View attachment 71159


The IMU0 green trace is the real speed - the OSD (txt log) speed (blue) was coming from IMU1 (red). So it is IMU1 that is faulty.

Could the OP try an IMU calibration or do you think it’s too far past that point? Seems worth a try but the fact it wasn’t asking for one is also worrisome. Reflash firmware before giving up?

 

[Kilrah]

Calibration fixes bad offset and scaling, not suprious garbage data. Thing needs to go back to DJI and get a warranty repair.

 

[djiuser_MLPTTHpRA3yw]

First question that comes to mind, Are you using genuine DJI props?

Hi, This Mavic Air is 6 months old and all parts are originals. Nothing changed.

 

[djiuser_MLPTTHpRA3yw]

Those speed data are from IMU1, and are spurious. I was wrong in my previous post - IMU1 was the active IMU and the IMU0 attitude excursions were real - hence the "not enough power" messages. If you compare the data from IMU0 you can see what the speed was really doing:

View attachment 71159


The IMU0 green trace is the real speed - the OSD (txt log) speed (blue) was coming from IMU1 (red). So it is IMU1 that is faulty.

Biggest thanks for the info.
But If there is IMU a mechanical error so how can he always balance himself after the RTH button ?

 

[sar104]

Biggest thanks for the info.
But If there is IMU a mechanical error so how can he always balance himself after the RTH button ?

The problem appeared to be triggered by the relatively large (real) pitch excursion when you released the elevator. It eventually settled down somewhat, which was unrelated to the RTH function. And while there continued to be errors from IMU1, they were not large enough to prevent the aircraft from, albeit unsteadily, returning under some level of control.

 

[Starz]

The problem appeared to be triggered by the relatively large (real) pitch excursion when you released the elevator. It eventually settled down somewhat, which was unrelated to the RTH function. And while there continued to be errors from IMU1, they were not large enough to prevent the aircraft from, albeit unsteadily, returning under some level of control.

You’re wizard abilities never cease to amaze me. Fun.

 

[sar104]

You’re wizard abilities never cease to amaze me. Fun.

Thanks. It's just numbers. With the right tools and a little persistence they reveal stuff.

 

[djiuser_MLPTTHpRA3yw]

Thanks. It's just numbers. With the right tools and a little persistence they reveal stuff.

Many thanks !
In addition to personal analysis skills, you will be skills to use right tools equipment and software. Very many unfortunately, are can not handle it nowadays.
Many thanks and best wishes for your success !
Thanks also to everyone who helped me with my Mavic Air problem !

 

[djiuser_MLPTTHpRA3yw]

Hello and thanks again !

DJI acknowledged the problem with the warranty.
DJI sent a new Mavic Air.
I haven't tried flying yet. I hope that this drone is fine .

 

[sar104]

Hello and thanks again !

DJI acknowledged the problem with the warranty.
DJI sent a new Mavic Air.
I haven't tried flying yet. I hope that this drone is fine .

Good - that was the correct outcome since the aircraft had a faulty IMU.

 

[John MacAulay]

Good outcome and great analysis, DJI did the right thing. Enjoy your new Mavic Air.