Mavic Mini uncommanded descent tests

[FoxhallGH]

I don't own a Mavic Mini, but I'm wondering if the underside of the drone ends up 'hot' after a flight, same as my Mavic Pro does? If the drone is still warm when packed into the Fly More case, and the rear blades are squashed onto the foam block under the drone, then not only are they going to be deformed, but they are going to be heated in that deformed state and will slowly cool while under pressure. An ideal recipe to ensure that the prop blade retains the distortion.
The tests done above are definitely confirming that the motor rpm is sufficient and there when needed. That reinforces for me, the thinking that the overall pitch of the rear prop's have been reduced (by the Fly More case squeeze!), and the problem is coming about when the prop rpm is so high, that they go into 'stall' condition and can no longer lift the aircraft. They will be effective (though not as efficient as a correct shaped prop), up until the point of stall. When the stall occurs, you can increase the rpm all you like, but you'll get no lift from the props!

 

[FoxhallGH]

Many thanks for carrying out the research to confirm some of the theories that have been around. Whilst I've not experienced the issue myself (only had about 15 flights since new, due to weather and only being able to fly in/above my garden), it's something I'd been monitoring on the forums even before I bought my Mini.

Would it be fair to say that the issue is less likely to occur at slower speeds? If so, people who spend most of their time in C or P mode in light winds may potentially never see the issue?

It might be more correct to assume that you are not going to see this condition until the MM gets into a 'problematic' situation where it's going to need a lot of lift from the rear prop-set. In other words, the bad thing about this issue is that you don't realise you have the problem until you need to get your MM out of trouble!

 

[ALan_m10]

I don't own a Mavic Mini, but I'm wondering if the underside of the drone ends up 'hot' after a flight, same as my Mavic Pro does?

I have noticed that the underside of mine was 'warm' on occasion, you may be on to something here!!

 

[sli-woody]

I just opened up a new pair of mavic mini props and they definitely feel "stiffer" than the ones I have been using on the drone for the past month. (no crashes)
The ones I have been using for approx 1 month feel like a well used playing card. They bend upwards and downwards with the lightest of pressure. Surely they shouldn't fatigue that quickly.
DJi sort your $+++ out and release some updated props that are stiffer!

 

[GerdS]

Until DJI will provide a fix I can just reommend:

- Don't fully trust your MM and always count with an uncommanded descending without proper reaction on stick input
- Don't fly over regions with high risk of loss, e.g. water, forest or impassable areas

 

[sar104]

Many thanks for carrying out the research to confirm some of the theories that have been around. Whilst I've not experienced the issue myself (only had about 15 flights since new, due to weather and only being able to fly in/above my garden), it's something I'd been monitoring on the forums even before I bought my Mini.

Would it be fair to say that the issue is less likely to occur at slower speeds? If so, people who spend most of their time in C or P mode in light winds may potentially never see the issue?

It almost certainly depends on how deformed the props have become. If they are bad enough the problem may appear in any mode. One early indicator of the issue is that on takeoff the rear lags the front in becoming airborne when the initially commanded rear motor speed is inadequate to lift the back of the aircraft. You can see the resulting positive pitch blip at the start of both the flights that I posted above. I'd watch for that as a test of the rear props.

 

[RichR]

Thanks - I think that's a good thing to add to the pre-flight checklist. If the rear is sluggish to lift, set back down again and examine rear props, replace if necessary.

 

[RayOZ]

So, is the uncommanded descend still happening if you release the elevator? Some of the other threads, releasing the sticks still result in descent. I wonder if flying backwards reduces the motor speed, to reduce the spoiling and maybe bring the props out of stall. Manually fly home in reverse?

 

[sar104]

So, is the uncommanded descend still happening if you release the elevator? Some of the other threads, releasing the sticks still result in descent. I wonder if flying backwards reduces the motor speed, to reduce the spoiling and maybe bring the props out of stall. Manually fly home in reverse?

Yes - if you look at the graphs above you will see that it occurs whether or not the elevator is released.

 

[Ice_2k]

wow, thanks a lot for experimenting with this and putting your Mini at risk!

 

[Prismatic]

It will be very interesting to see if an uncommanded descent can be arrested with an immediate application of reverse elevator and maximum throttle. If so, some Mini pilots will escape disaster, but that maneuver would be risky in many situations. First, one would have to react before too much altitude was lost. Second, you'd need to know instinctively or visually that there were no obstructions to the rear.

And if it worked, you'd want to shepherd that little bird back home for some propeller R&R (Removal & Replacement) before flying again.

 

[LivinLarge]

Nice work. DJI needs to address this problem.

 

[Herbie]

Nice work. DJI needs to address this problem.

New to drones and reading all about these uncommanded descends and crashes. What affect does density altitude have on drones?
When flying aeroplanes one always take into account height above sea level, how hot it is, which time of the day is best for flying etc.
Just asking.

 

[Ice_2k]

Nice work. DJI needs to address this problem.

Maybe they did, let’s wait for sar104’s tests with the new firmware

 

[LivinLarge]

Maybe they did, let’s wait for sar104’s tests with the new firmware

If the issue is deformed props, I don't see how a firmware change can fix it. A proper solution would be to give owners new, stiffer props and warn them about how to safely store the Mini to avoid distortion. DJI has not had a history of admitting their problems, so I would be surprised if they do in this case.

 

[Ice_2k]

If the issue is deformed props, I don't see how a firmware change can fix it. A proper solution would be to give owners new, stiffer props and warn them about how to safely store the Mini to avoid distortion. DJI has not had a history of admitting their problems, so I would be surprised if they do in this case.

Well, there are multiple things firmware can do. It can warn you when the props need changing (which it would appear it does) and it could prioritize not descending over trying to maintain speed. I’m hoping sar104’s tests will show the new firmware does both and that it does them well.

 

[sar104]

Okay - here are the same tests run under 0500. The actual flight control differences appear to be slight - all I see is a small increase in motor maximum rpm from around 16 krpm to 17 krpm. The FW may also be a little less aggressive in trying to hold forward pitch, which somewhat mitigates the onset of uncommanded descent, but doesn't eliminate it.

However, with representatively flattened rear props the forward speed in sport mode is still low (7 - 8 m/s) due to inability to hold more than around 15° pitch, while backward speed still achieves the specified 13 m/s. And uncommanded descent still happens - it just has to be pushed a bit harder to get there. You can see the three larger blips in vertical speed in the final third of the flight - in all three cases the aircraft either hit the ground or came close enough for ground effect to arrest the descent.

You will likely need to click on this graph to expand it enough to be clear.



So the only substantial difference that I see with the new firmware is the motor error warning. This test flight resulted in warnings on both rear motors:

97.4 MOTOR SPEED ERROR. LAND OR RETURN TO HOME PROMPTLY. AFTER POWERING OFF THE AIRCRAFT, REPLACE THE PROPELLER ON THE BEEPING ESC. IF THE ISSUE PERSISTS, CONTACT DJI SUPPORT (CODE: 302)

The warning doesn't actually change the behavior in terms of pitch or speed achieved.

In conclusion - this FW update is just a bandaid for the combination of poor propellers and how easy it is to deform them in storage. In my opinion there is no way to fix this in firmware without completely changing the aircraft specifications, and even then it is going to be marginal. The aircraft needs better/stiffer props, even if that puts it slightly over the 250 g limit. Realistically, no one is going to weigh the aircraft in real world use, so aftermarket props are the obvious solution even if DJI can't supply it with heavier props.

 

[Ice_2k]

Okay - here are the same tests run under 0500. The actual flight control differences appear to be slight - all I see is a small increase in motor maximum rpm from around 16 krpm to 17 krpm. The FW may also be a little less aggressive in trying to hold forward pitch, which somewhat mitigates the onset of uncommanded descent, but doesn't eliminate it.

However, with representatively flattened rear props the forward speed in sport mode is still low (7 - 8 m/s) due to inability to hold more than around 15° pitch, while backward speed still achieves the specified 13 m/s. And uncommanded descent still happens - it just has to be pushed a bit harder to get there. You can see the three larger blips in vertical speed in the final third of the flight - in all three cases the aircraft either hit the ground or came close enough for ground effect to arrest the descent.

You will likely need to click on this graph to expand it enough to be clear.

View attachment 99781

So the only substantial difference that I see with the new firmware is the motor error warning. This test flight resulted in warnings on both rear motors:

97.4 MOTOR SPEED ERROR. LAND OR RETURN TO HOME PROMPTLY. AFTER POWERING OFF THE AIRCRAFT, REPLACE THE PROPELLER ON THE BEEPING ESC. IF THE ISSUE PERSISTS, CONTACT DJI SUPPORT (CODE: 302)

The warning doesn't actually change the behavior in terms of pitch or speed achieved.

In conclusion - this FW update is just a bandaid for the combination of poor propellers and how easy it is to deform them in storage. In my opinion there is no way to fix this in firmware without completely changing the aircraft specifications, and even then it is going to be marginal. The aircraft needs better/stiffer props, even if that puts it slightly over the 250 g limit. Realistically, no one is going to weigh the aircraft in real world use, so aftermarket props are the obvious solution even if DJI can't supply it with heavier props.

Did you get a chance to test if going back on the right stick and forward on the left one stops the descent?

 

[sar104]

Did you get a chance to test if going back on the right stick and forward on the left one stops the descent?

You can see in the graph that I tried that at 597 seconds, and then repeated the same maneuver with uncommanded descent at 695 seconds. There was no distinguishable difference in vertical response.

 

[lee82gx]

You can see in the graph that I tried that at 597 seconds, and then repeated the same maneuver with uncommanded descent at 695 seconds. There was no distinguishable difference in vertical response.

sar104, would you say that as long as the flight data shows 18krpm (rear) vs 10krpm (front), the actual flight problem will not have been addressed? Depending on the pitch of the axis of motor front and rear, shouldn't be so much different in speed to achieve flat thrust needed to hover the aircraft right?
I remember that on my Pro 1, the RPM difference in hover / vertically stable was at most 500rpm different or in ratio, about 10%