Mavic Mini loosing control mid-air -> Second time

[rafare]

This is a follow-up of this topic.

Mavic mini loosing control mid-air?

Hey guys, First of all, there was no loss here, but I really wanted to figure it out what happened. 1- Contacted "ATC" of the airport that is less than 500m from my take off, they just told me to stay parallel to the runway, but theres almost no movement as this is a very small private...

mavicpilots.com

But this is other incident, hence why the new topic.

Time of incident:
7m4.7s - 7m6.4s

Can you guys see anything wrong with my MM in the logs? It has been doing this the day I got it, I would say each 2 or 3 flights.
In this flight the wind was moderate, I could feel the drone not happy flying straight into the wind, it was about 20 kph gusts according to UAV Forecast.

The only thing I can say that both flights had in common is the compass calibration prompt:
"3- Before take off it prompted me to calibrate compass, The first calibration was unsucessful, for the second I walked away from the spot I was a few feet just to make sure it wasn't some type of metal underneath the grass, and it calibrated sucessfully "

I also noticed in this specific flight that the mini was not easily maintaing altitude, it was going down considerabily each time I let off the throttle. I know MM do not have much altitute stability but in this specific incident appeared to be less stable.

Logs:

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

 

[boblui]

Time of incident:
7m4.7s - 7m6.4s

Can you guys see anything wrong with my MM in the logs? It has been doing this the day I got it .....

Can you elaborate what this is ? As far as I can see from the log, the craft responded normally to all control inputs

 

[rafare]

Can you elaborate what this is ? As far as I can see from the log, the craft responded normally to all control inputs

Hey, Im sorry, forgot to put this in this topic.

But basically the drone lost its orientation, the camera shaked in all the directions, just like the last topic. Just like last time, I thought was the gimbal doing its thing, but looking at the logs you can see the MM moving off its intended flight path.
you can see in the logs just before that it was yawing to the left with no stick inputs. I did not noticed that when flying, could be an compass error. I remember it flying straight just before this incident.
I recommend taking a look at the topic I mentioned. Its much more detailed and it’s exactly the same thing that happened here.

 

[boblui]

...you can see in the logs just before that it was yawing to the left with no stick inputs...

I am afraid I cannot see it. You have moved the rudder stick a few times around the mentioned time period and the craft responded normally :

 

[slup]

I am afraid I cannot see it. You have moved the rudder stick a few times around the mentioned time period and the craft responded normally :

View attachment 112245

It's this uncommanded ... or actually opposite commanded yaw movement he means, it goes negative 40 degrees but then back & adds more positive rotation as commanded earlier.

DAT log ends with FLY073.DAT

 

[rafare]

It's this uncommanded ... or actually opposite commanded yaw movement he means, it goes negative 40 degrees but then back & adds more positive rotation as commanded earlier.

Theres the DAT file. I'll also add to the first post.

Can you find anything in the logs just after 424s showing the drone move around? Looking at the camera feed I thought it was going to fall off the sky because the the camera was looking to every direction but the one I was actually going (I know the gimbal disarms sometimes). Looking at the flight path on google earth it drifted to the side a little, but nothing absurd. but very very identical to the topic I mentioned.

 

[slup]

Theres the DAT file. I'll also add to the first post.

Can you find anything in the logs just after 424s showing the drone move around? Looking at the camera feed I thought it was going to fall off the sky because the the camera was looking to every direction but the one I was actually going (I know the gimbal disarms sometimes). Looking at the flight path on google earth it drifted to the side a little, but nothing absurd. but very very identical to the topic I mentioned.

Mmm ... don't really think that this is so mysterious actually.

If we first check off if all sensors is sensing the slight uncommanded 43 degree CCW turn. This firstly by comparing the Gyro (green) with the unwrapped imuYaw (red) ... they agree. Then comparing the imuYaw (180/180 y-axis) (blue) with the magYaw (dark green) ... they agree. The turn is real, all sensor values agrees. And we see by the yaw stick command (brown) that the movement is uncommanded, it starts before the small input that comes at 425sec. & the yawing goes the opposite direction than the later stick command.



It's only motor revs together with props grabbing the air that create torque that make the AC yaw ... this by having 2 CW & 2 CCW torquing motors/props, every thing in balance = no yaw. Below a pic that explain ... torque direction is opposite turning direction.



Looking into how the motor revs behave during the uncommanded yaw ... the black graph is the left rear & it's mainly maxed out.



Is this what the flight controller want? Yepp (as it want to keep generating lift) ... the left rear (light blue) is commanded to 100%



So how about the rest of the flight ... here first the revs, then the commanding. Easy to see that something is going on with the left rear ... it's well above all the others the majority of the time ... the graph is from a longer period of Sport mode bashing.

Revs (left rear black)



Command (left rear light blue)



So ... the yawing is most probably due to a deformed left rear prop not being able to generate sufficient grip in the air ... bringing down the torque that prop generate. This means unbalance between the CW & the CCW corners leading to a weaker CW total torque ... meaning yaw CCW.

So any height instability then ..? Well it looks to behave reasonably well so far, but flying a bit more might very well have created a uncommanded descent incident also. Below a comparison between your throttle inputs (dark green) with corresponding vertical speed change (blue). With full throttle for ascending it should create a large negative vertical speed value ... and that happens here during the same long Sport mode session.



My advice is to change to a fresh set of new props (especially the one in left rear) & try again.

 

[rafare]

I'm sure your analysis is much more in depth than if I had contacted DJI. Thanks very much.

The conclusion is this is different than the last time, where one of the rear motors decelerated, causing the uncommanded yaw.

I'll change the props and maybe see if I can see the deformation in the prop you mentioned.

Thanks a lot.

 

[rafare]

@slup Just to give you feedback. when Inspecting the props I could see the right rear was a little bent near the tip. Could not see anything wrong at the left rear. But I changed both props to make sure.

 

[slup]

@slup Just to give you feedback. when Inspecting the props I could see the right rear was a little bent near the tip. Could not see anything wrong at the left rear. But I changed both props to make sure.

The right rear performs as it should according to the log ... the usual deformation causing thing's like this is more of a flattening which cause the leading edge of the prop to have a to low angle of attack. The known deformation of the Mini's props can be very hard to see & sometimes completely new props are already flattened from start. The latest AC FW 1.05 sense this to fast rotating & through up a motor error advising the pilot to change.

 

[rafare]

the usual deformation causing thing's like this is more of a flattening which cause the leading edge of the prop to have a to low angle of attack

This was exactly the case with the right rear. As you said there was no indication when flying, but the prop is clearly bent, if looked from the side you could see the leading edge near the tip of the prop it was poiting upwards. Hard to explain but it was clearly damaged. Hope that solves this issue about the mini "drifting"

Just a offtopic question about props: Every prop package comes 4 props, can I change just one pair of props ( one motor ), or I need to change all 4 props from 2 different motors, using all the props that came in the package?

 

[boblui]

...., can I change just one pair of props ( one motor ), or I need to change all 4 props from 2 different motors, using all the props that came in the package?

I would suggest to check the motor speeds when hovering. Ideally, they should be exactly the same. This was my Mini when hovering in a windless environment. All props were in good shapes :



The revolution speed of the motors at the back was significantly higher than those of the front motors. This is not good because it means the motors at the back will reach their speed limit sooner so the overall power of the craft is reduced.

I have been suspecting that the higher revolution speed is caused by the pitch of the props reduced for some reasons. To verify this I swapped the front props with those at the back and checked the motor speeds when hovering indoor. This is the result. Now the front motors spin faster :



Then I swapped one blade of the RFront motor with one of those of the LBack motor. The blade swapped was just randomly selected. The speed of the four motors became very close when hovering indoor :