Loss of Control in Flight - MAVIC PLATINUM PRO

[South-West-Skies]

@slup might like to look at this one.

Any help as to the cause would be greatly appreciated, one issue I noticed on the flight data was the rudder response delay being in orange.

What's with the gimbal control problem? At one point your forward acceleration tilt shows the propellers.

No idea, just suddenly happened.

 

[Meta4]

No concrete under the tarmac to the best of my knowledge, as you thinking of steel bars under the concrete?

The drone on take off was about 6 m from the concrete road surface.

Thankyou
I can see the problem, I can eliminate various guesses, but can't find an obvious cause.

 

[South-West-Skies]

Thankyou
I can see the problem, I can eliminate various guesses, but can't find an obvious cause.

Could this be wind related issue?

 

[FLDave]

@South-West-Skies What can you tell us about the wind during the flight. I've seen my Air 2s have some violent movement related to wind when it made some turns during a flight. The Air 2s fought to maintain position and rotation with tilts more than I would like to see. The drone was well within view with take-off site wind measured at about 8-15 mph; maximum in-flight wind was 17 mph. Your drone is large enough to be stable in wind.
Was the video with the propeller showing due to your acceleration or reaction to wind?
How close were you to the drone when all of this happened?
So, if it's not external factors or user control, then it would likely be hardware/software with drone or Fly app.

 

[Meta4]

Could this be wind related issue?

Again ... NO!!
A strong wind might blow the drone sideways or slow its progress if you flew into it.
You couldn't fly in a wind that strong without noticing.

But there's no kind of wind that could spin your drone around for 1-2 seconds and then stop.
Wind had nothing at all to do with this incident.
Nothing, nada, nix, zip, zilch.
It wasn't wind related.

 

[Meta4]

@South-West-Skies What can you tell us about the wind during the flight. I've seen my Air 2s have some violent movement related to wind when it made some turns during a flight. The Air 2s fought to maintain position and rotation with tilts more than I would like to see. The drone was well within view with take-off site wind measured at about 8-15 mph; maximum in-flight wind was 17 mph. Your drone is large enough to be stable in wind.
Was the video with the propeller showing due to your acceleration or reaction to wind?
How close were you to the drone when all of this happened?
So, if it's not external factors or user control, then it would likely be hardware/software with drone or Fly app.

Guessing is great for coming up with improbable things that didn't/couldn't happen.
The flight data shows what happened and none of those suggestions could cause that.

 

[slup]

@slup might like to look at this one.

Sorry, but have had a mishap with my computer so can only look at this on my phone or Tab ...

But out from the Airdata views & the attached video the problem seems to be 2 parted (maybe connected though). The first looks like a gimbal problem (have seen similar erratic movements before from especially MP & MPP's). The second is uncommanded CCW yaw movements during braking but also during the last decent.

Certain thing's can be ruled out IMHO though ...

-Wind ... no, the winds in the flight was in average 6,2m/s gusting up to 7,8m/s coming from north west & a measure point just where the first uncommanded yaw movement occurs says 6,7m/s. Furthermore a to strong wind wouldn't spin the drone, it would cause a drift down wind.

-A magnetic interfered power on spot creating a yaw error... no, if that was the case, wild consequences should have been seen pretty soon after take off ... not just isolated to a couple yaw turns at 50m height 3 minutes into the flight.

-RC or connection interference/delays ... no way, interference will at most make the live view glitch & RC delays will not cause uncommanded inputs.

To further look at this the DAT log is needed ... These are retrieved by the same method as the TXT logs. Under both iOS and Android they are in a subfolder, MCDatFlightRecords, in the folder that contains the TXT logs. The DAT log for this flight ends with FLY315.DAT. Retrieve it and attach it in a new post here, then maybe somebody else can look at it (@BudWalker ?) and maybe see what's going on ... the motor data would be interesting to have a look at, this as yaw movements is coming from a change in the generated rotational torque from the 4 corners ... is a change here commanded from the FC or not?

 

[South-West-Skies]

Sorry, but have had a mishap with my computer so can only look at this on my phone or Tab ...

But out from the Airdata views & the attached video the problem seems to be 2 parted (maybe connected though). The first looks like a gimbal problem (have seen similar erratic movements before from especially MP & MPP's). The second is uncommanded CCW yaw movements during braking but also during the last decent.

Certain thing's can be ruled out IMHO though ...

-Wind ... no, the winds in the flight was in average 6,2m/s gusting up to 7,8m/s coming from north west & a measure point just where the first uncommanded yaw movement occurs says 6,7m/s. Furthermore a to strong wind wouldn't spin the drone, it would cause a drift down wind.

-A magnetic interfered power on spot creating a yaw error... no, if that was the case, wild consequences should have been seen pretty soon after take off ... not just isolated to a couple yaw turns at 50m height 3 minutes into the flight.

-RC or connection interference/delays ... no way, interference will at most make the live view glitch & RC delays will not cause uncommanded inputs.

To further look at this the DAT log is needed ... These are retrieved by the same method as the TXT logs. Under both iOS and Android they are in a subfolder, MCDatFlightRecords, in the folder that contains the TXT logs. The DAT log for this flight ends with FLY315.DAT. Retrieve it and attach it in a new post here, then maybe somebody else can look at it (@BudWalker ?) and maybe see what's going on ... the motor data would be interesting to have a look at, this as yaw movements is coming from a change in the generated rotational torque from the 4 corners ... is a change here commanded from the FC or not?

I have attached a copy of the MCDatFlightRecords to this posting.

The initial change of direction is from the controller at return the drone to the initial take off and landing site.
11.09 seconds into the video the drone appears to judder
14.00 seconds the gimbal is a white out
14.07 rudder and upwards movement
There follows a number of movements and what appears to be a spin type movement in flight.

Then there is yaw and more movement and at that point we made a decision to make an emergency landing as we feared loss of control and the potential of loss of the drone.

 

[BudWalker]

Sorry, but have had a mishap with my computer so can only look at this on my phone or Tab ...

But out from the Airdata views & the attached video the problem seems to be 2 parted (maybe connected though). The first looks like a gimbal problem (have seen similar erratic movements before from especially MP & MPP's). The second is uncommanded CCW yaw movements during braking but also during the last decent.

Certain thing's can be ruled out IMHO though ...

-Wind ... no, the winds in the flight was in average 6,2m/s gusting up to 7,8m/s coming from north west & a measure point just where the first uncommanded yaw movement occurs says 6,7m/s. Furthermore a to strong wind wouldn't spin the drone, it would cause a drift down wind.

-A magnetic interfered power on spot creating a yaw error... no, if that was the case, wild consequences should have been seen pretty soon after take off ... not just isolated to a couple yaw turns at 50m height 3 minutes into the flight.

-RC or connection interference/delays ... no way, interference will at most make the live view glitch & RC delays will not cause uncommanded inputs.

To further look at this the DAT log is needed ... These are retrieved by the same method as the TXT logs. Under both iOS and Android they are in a subfolder, MCDatFlightRecords, in the folder that contains the TXT logs. The DAT log for this flight ends with FLY315.DAT. Retrieve it and attach it in a new post here, then maybe somebody else can look at it (@BudWalker ?) and maybe see what's going on ... the motor data would be interesting to have a look at, this as yaw movements is coming from a change in the generated rotational torque from the 4 corners ... is a change here commanded from the FC or not?

Didn't see much in the .DAT. But, looking at the .txt I noticed that the uncommanded yaw occurs immediately following an uncommanded Gimbal:Yaw in the opposite direction. I'm supposing the flight controller is trying to keep the gimbal pointed the same direction by performing an airframe CCW yaw to compensate for an uncommanded CW Gimbal:Yaw.

At the points indicated the Gimbal:Yaw (red) shows a CW rotation yet the rudder (not shown) is neutral. The response is for the FC to rotate the MP CCW to compensate. Similar behavior is also seen in pitch and roll.

Maybe the gimbal is loose?
@South-West-Skies

 

[South-West-Skies]

Didn't see much in the .DAT. But, looking at the .txt I noticed that the uncommanded yaw occurs immediately following an uncommanded Gimbal:Yaw in the opposite direction. I'm supposing the flight controller is trying to keep the gimbal pointed the same direction by performing an airframe CCW yaw to compensate for an uncommanded CW Gimbal:Yaw.
View attachment 150219
At the points indicated the Gimbal:Yaw (red) shows a CW rotation yet the rudder (not shown) is neutral. The response is for the FC to rotate the MP CCW to compensate. Similar behavior is also seen in pitch and roll.

Maybe the gimbal is loose?
@South-West-Skies

I have arranged to send the Mavic to be tested and repaired if required, be interesting to see the results.

 

[Don Testme]

Yesterday I performed a long range flight with my Mavic 3. On the way back with 60% battery. While crossing a main road it suddenly rotated and began falling. It’s momentum helped it make it across the street and crash onto a cement sidewalk. The battery popped out and was cracked severely. The camera seemed to take the brunt of the crash. It was in pieces and shattered. One leg was missing.

I drove to it’s last location. Luckily a pedestrian saw it fall from the sky. I asked him if he saw a crashed drone. Luckily he did and directed me to the exact position it crashed. If anyone wants to see the last flight log let me know. I’m curious to know why it just suddenly failed in mid flight. Question, I can still get the logs from just my remote right? I sent my destroyed drone in for care refresh replacement.

 

[South-West-Skies]

Update for all who were helpful, the MAVIC was returned to be repaired and inspected, turns out that the gimbal was at fault and will be replaced under warranty.

Interestingly the gimbal was a replacement after an earlier incident and was the first time I had flown the mavic since the repair carried out by a DJI repair agent.

Waiting for my mavic to find it's way back next week and hopefully back in the air soon, though might be a little cautious to start with.

Thanks all, the help has been greatly appreciated.

 

[fiberdude]

Update for all who were helpful, the MAVIC was returned to be repaired and inspected, turns out that the gimbal was at fault and will be replaced under warranty.

Interestingly the gimbal was a replacement after an earlier incident and was the first time I had flown the mavic since the repair carried out by a DJI repair agent.

Waiting for my mavic to find it's way back next week and hopefully back in the air soon, though might be a little cautious to start with.

Thanks all, the help has been greatly appreciated.

MP and MPP have the weakest gimble set.

When you have an exchange for crash MP/MPP and get a REFURBISHED copy of MP/MPP (DJI will NOT give you a brandnew, box-sealed MP/MPP in such a situation), there is a high chance that the gimble turns out to be faulty. It's much more likely now than 4-5 years ago as DJI is running of inventory for refurbished MP/MPP. In fact, in China, there was a short supply of refurb MP/MPP even in early 2019.

 

[South-West-Skies]

MP and MPP have the weakest gimble set.

When you have an exchange for crash MP/MPP and get a REFURBISHED copy of MP/MPP (DJI will NOT give you a brandnew, box-sealed MP/MPP in such a situation), there is a high chance that the gimble turns out to be faulty. It's much more likely now than 4-5 years ago as DJI is running of inventory for refurbished MP/MPP. In fact, in China, there was a short supply of refurb MP/MPP even in early 2019.

I was charged for a new replacement gimbal, I am not sure what you mean by MP and MPP?

Thanks.

 

[fiberdude]

I was charged for a new replacement gimbal, I am not sure what you mean by MP and MPP?

Thanks.

MP = Mavic Pro; MPP = Mavic Pro Platinum.

The two are pretty the same drone inside and out. In 2019, when DJI flagship store in the city of Shenzhen ran out of refurb MP, they gave me a free "upgrade" to a refurb MPP instead. Replacing a certain component in the drone will take quite some time and requires a return of the drone to factory; if one is in a hurry, as is the case usually, DJI will swap a refurb instead.