DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

The problem of un-commanded descent by Mini reproduced

new2mavic

Well-Known Member
Joined
Dec 5, 2019
Messages
551
Reactions
366
Location
Hong Kong
There has been numerous reports on un-commanded descent for Mini and from the log files posted by owners, there are some common symptoms:
  • The craft suddenly descended by itself with no pilot input
  • “Maximum power load reached” warning showed up
  • Some motors have the revolution speed maxed out, often occurred on those at the back
  • Occurrence is inconsistent, cannot be reproduced at will
I have been suspecting that the problem is due to bad props not being able to generate sufficient thrust. As a matter of fact, a member has reported that the problem went away after prop replacement.

To verify whether bad props can indeed cause such behavior, I made two pairs of bad props yesterday by clamping the props between metal plates and baking it in the kitchen oven at 100 degree Celsius for 40 minutes. Props would have the pitch reduced after the process, They were mounted to the rear motors of my Mini.

Original prop ( top ) and bad prop ( bottom ) compared. The difference in pitch angle is obvious :

good and bad props.jpg


Motor speeds when hovering indoor before prop replacement. All four motors were spinning at about the same speed :

1581754624710.png


This is after prop replacement. The rear motors were spinning at their maximum speed. Something like this has been seen in some of the flight log files posted by members reporting the un-commanded descent problem. During take off, the Mini could no longer went straight up but drifted backward by about 1.5 meters before settling down. It remained very stable throughout hovering and landing though.

1581754666515.png


Then I took it outside. The wind was mild to none on the ground but likely stronger higher up. With the Mini in sport mode and at an altitude of about 18 meters, I repetitively flew it into the wind with full forward elevator and see what would happen. A number of behaviors were observed :

1) No major problem except that the tilting angle was small and the forward speed was low

1581754781600.png


2) The forward speed dropped after a short while and then the Mini descended by itself. All sticks were then returned to neutral immediately but the craft continued to descend before stopping by itself. Quite unexpectedly, it automatically ascended back to the original altitude afterwards.

1581754810215.png


3) Same as (2) but it went so low that I chickened out and applied full throttle to save it from crashing. The descent was arrested successfully but the craft would not automatically go back to the original altitude like what it did in (2)

1581754839584.png


4) Same as (3) but it hit the ground, bounced back and then descended again . Fortunately there was no damage

1581754894380.png

Screen capture of this case :




The .TXT and .DAT files are attached.

So what do you guys think ? Does this explain some of the reported incidents of un-commanded descent by the Mini ? Should Mini owners check the health being of the props by checking the speed of the motors when the craft is hovering in no-wind condition ? I haven’t tried it in stronger winds but I would imagine that the problem can only be worse.
 

Attachments

  • 2020-02-14_16-20-58_FLY080.DAT
    5.4 MB · Views: 6
  • DJIFlightRecord_2020-02-14_[16-21-14].txt
    1,021 KB · Views: 6
Last edited:
Interesting but all that really tells me nothing . Now
@sar104 @BudWalker @Meta4 can say if any what
have going there. You lost me with the first graph. ?
 
We know for fact that the uncommanded descents are due to the propellers deformation , we do not yet know what is the cause of deformation in the first place. My theory is that the arms of the mini flex during flight ( because they are hollow and filled with foamy material) and that leads to the propeller deformation over time.
 
  • Like
Reactions: Campin123
I think that this establishes that most, if not all, of the reported behaviors (specifically failure to achieve specified pitch angle, uncommanded descent, very high rear motor speeds) can be caused by deformed props. It does not definitively confirm that bad props are the only possible cause, but it is very suggestive.

It leads to a couple of other obvious questions. Firstly - all the reports seem to indicate problems with just the rear props. Does that mean that only the rear props are getting deformed, or do deformed props not affect performance if they are on the front motors? Did you try the baked props on the front, or only on the rear? Secondly - if it is just the rear props becoming deformed - is that an issue with how they are folded/stored or is that happening in flight due to higher loads, or similar?
 
There has been numerous reports on un-commanded descent for Mini and from the log files posted by owners, there are some common symptoms:
  • The craft suddenly descended by itself with no pilot input
  • “Maximum power load reached” warning showed up
  • Some motors have the revolution speed maxed out, often occurred on those at the back
  • Occurrence is inconsistent, cannot be reproduced at will
I have been suspecting that the problem is due to bad props not being able to generate sufficient thrust. As a matter of fact, a member has reported that the problem went away after prop replacement.

To verify whether bad props can indeed cause such behavior, I made two pairs of bad props yesterday by clamping the props between metal plates and baking it in the kitchen oven at 100 degree Celsius for 40 minutes. Props would have the pitch reduced after the process, They were mounted to the rear motors of my Mini.

Original prop ( top ) and bad prop ( bottom ) compared. The difference in pitch angle is obvious :

View attachment 94002


Motor speeds when hovering indoor before prop replacement. All four motors were spinning at about the same speed :

View attachment 93994


This is after prop replacement. The rear motors were spinning at their maximum speed. Something like this has been seen in some of the flight log files posted by members reporting the un-commanded descent problem. During take off, the Mini could no longer went straight up but drifted backward by about 1.5 meters before settling down. It remained very stable throughout hovering and landing though.

View attachment 93995


Then I took it outside. The wind was mild to none on the ground but likely stronger higher up. With the Mini in sport mode and at an altitude of about 18 meters, I repetitively flew it into the wind with full forward elevator and see what would happen. A number of behaviors were observed :

1) No major problem except that the tilting angle was small and the forward speed was low

View attachment 93996


2) The forward speed dropped after a short while and then the Mini descended by itself. All sticks were then returned to neutral immediately but the craft continued to descend before stopping by itself. Quite unexpectedly, it automatically ascended back to the original altitude afterwards.

View attachment 93997


3) Same as (2) but it went so low that I chickened out and applied full throttle to save it from crashing. The descent was arrested successfully but the craft would not automatically go back to the original altitude like what it did in (2)

View attachment 93999


4) Same as (3) but it hit the ground, bounced back and then descended again . Fortunately there was no damage

View attachment 94000

Screen capture of this case :




The .TXT and .DAT files are attached.

So what do you guys think ? Does this explain some of the reported incidents of un-commanded descent by the Mini ? Should Mini owners check the health being of the props by checking the speed of the motors when the craft is hovering in no-wind condition ? I haven’t tried it in stronger winds but I would imagine that the problem can only be worse.
I'm beginning to suspect that it's more than just deformed props. I get the impression these events happen towards the end of a long flight. Looking at the rightBack:commanded and rightBack:Speed shows that the efficiency of the rightBack motor/ESC drops over the course of the flight.
1581776099074.png
And here is the RBackSpeed/RBackCommanded shown using Excel. Maybe I should bite the bullet and start using Igor Pro
1581776329392.png

Here is a the same data for my MM where I was trying to get it do a umcommanded descent. It doesn't show an efficiency degradation.
1581776265105.png

I'm looking at ESC:temp and motor:Volts to see if they could be a cause. So far, it doesn't seem so. I'm also looking at V_out, which I think must be the output of the ESC, vs motorVolts.
 
We know for fact that the uncommanded descents are due to the propellers deformation , we do not yet know what is the cause of deformation in the first place. My theory is that the arms of the mini flex during flight ( because they are hollow and filled with foamy material) and that leads to the propeller deformation over time.
I must have missed that. It's still very much a mystery, at least to me.
 
I'm beginning to suspect that it's more than just deformed props. I get the impression these events happen towards the end of a long flight. Looking at the rightBack:commanded and rightBack:Speed shows that the efficiency of the rightBack motor/ESC drops over the course of the flight.
View attachment 94013
And here is the RBackSpeed/RBackCommanded shown using Excel. Maybe I should bite the bullet and start using Igor Pro
View attachment 94016

Here is a the same data for my MM where I was trying to get it do a umcommanded descent. It doesn't show an efficiency degradation.
View attachment 94015

I'm looking at ESC:temp and motor:Volts to see if they could be a cause. So far, it doesn't seem so. I'm also looking at V_out, which I think must be the output of the ESC, vs motorVolts.

Interesting observation on the demand/speed ratio. The big obvious difference in the first data set is that the motor speeds are over 50% higher. It would not be surprising if the motors are simply not designed to run that fast for extended periods.

Maybe I should bite the bullet and start using Igor Pro.

Just do it.
 
  • Like
Reactions: Simmo
I must have missed that. It's still very much a mystery, at least to me.
We know so far that exchanging the propellers for new ones solves the problem , motor rpm goes back to normal , we don't know for how long though.
 
I did some experiments with my MM where I was tying to cause the uncommanded descent behavior. In this flight (FLY024) there was actually a small altitude increase after releasing full down throttle. There are several more of these in that flight.
1581778316700.png

There were several runs with full elevator forward and full throttle down followed by abruptly releasing the throttle. FLY024.DAT is attached.
 

Attachments

  • 2020-02-11_15-47-48_FLY024.DAT
    7.3 MB · Views: 4
  • Like
Reactions: brett8883
I just had an uncommanded descent / motor can't reach full speed warning (I think that's what it said) this morning. Thankfully it descended from about 10' and landed on a thick area of clover, no damage. When I retrieved the drone I noticed that one of the rear props had snapped off. That doesn't explain underspeed but it explains the descent. The sound I heard was overspeed not underspeed. I probably read it wrong. I was a bit distracted by my drone crashing without help from me.
 
@sar104 @BudWalker @Meta4 , a few weeks ago I was flying at the local park and doing the last flight to kill the battery before heading home. It was gonna be a short flight but some thing strange happened and I had to walk 1/4 miles to picked up the MM. I landed it in the field near by due to the strong wind, but at about 2-3 minutes into the flight the MM descended on its own.
I hope I download the flight TXT correctly. They're all the same ,but I don't know which one you guys can use.
 

Attachments

  • DJIFlightRecord_2020-01-26_[14-14-23][1]log-force landing.zip
    573.1 KB · Views: 5
  • DJI MINI in CSV-force landing.zip
    142.3 KB · Views: 5
  • DJIFlightRecord_2020-01-26_[14-14-23][1]csv-verbose.zip
    254.1 KB · Views: 6
I'm beginning to suspect that it's more than just deformed props. I get the impression these events happen towards the end of a long flight. Looking at the rightBack:commanded and rightBack:Speed shows that the efficiency of the rightBack motor/ESC drops over the course of the flight.
View attachment 94013
And here is the RBackSpeed/RBackCommanded shown using Excel. Maybe I should bite the bullet and start using Igor Pro
View attachment 94016

Here is a the same data for my MM where I was trying to get it do a umcommanded descent. It doesn't show an efficiency degradation.
View attachment 94015

I'm looking at ESC:temp and motor:Volts to see if they could be a cause. So far, it doesn't seem so. I'm also looking at V_out, which I think must be the output of the ESC, vs motorVolts.

Good observation.

At the beginning I was flying in P mode and I could not get the Mini to go into uncommanded descent. Then I switched to sport mode and found it increasingly easy to reproduce the problem. Looking back, it seems that mode change was not the reason but the drop in motor power over time.

Every big dip in height at full forward elevator input is an uncommanded descent :
1581800965046.png
 
Man on man I hope DJI is paying attention. They’ve gotten a lot of information from the field by now. Time to come up with a solution. At least a plan of some kind to help us out. I’m sure a lot of minis are just sitting on shelves right now with no problem. But those of us who fly nearly every day need help, soon.
 
  • Like
Reactions: MAvic_South_Oz
We know for fact that the uncommanded descents are due to the propellers deformation , we do not yet know what is the cause of deformation in the first place. My theory is that the arms of the mini flex during flight ( because they are hollow and filled with foamy material) and that leads to the propeller deformation over time.
Not saying it is the only cause, but I have seen so many YouTube videos where people are carrying their MM's in a pocket or just throwing them in a backpack along with about 100 other loose items etc. No prop guards nothing. It makes me cringe every time. I am actually amazed when they pull them out and they still fly, but you have to wonder how many people do that and end up with weirdly compressed, bent, distorted props etc. In my opinion, a proper, well designed case is a necessity. So is making certain that the props are not pressing against anything.
 
Last edited:
We know for fact that the uncommanded descents are due to the propellers deformation , we do not yet know what is the cause of deformation in the first place. My theory is that the arms of the mini flex during flight ( because they are hollow and filled with foamy material) and that leads to the propeller deformation over time.
Sorry but I doubt at this "fact". I had the uncommanded descending just once with my MM. And it happened during flying home in P mode and making a short break with hoovering at an altitude of about 10m. When I detected the altitude loss I was able to compensate with more throttle and complete my return flight. There was only calm wind and no warnings about wind or power limits reached. After landing I replaced the battery and immediately took off for a further flight for verifying what just had happened. I was not able to reproduce the incident and since there I never experienced something similar again, although I didn't replace anything.

If your theory would be correct the problem would have shown up again for sure, especially as meanwhile I had flights in much stronger wind.
 
Sorry but I doubt at this "fact". I had the uncommanded descending just once with my MM. And it happened during flying home in P mode and making a short break with hoovering at an altitude of about 10m. When I detected the altitude loss I was able to compensate with more throttle and complete my return flight. There was only calm wind and no warnings about wind or power limits reached. After landing I replaced the battery and immediately took off for a further flight for verifying what just had happened. I was not able to reproduce the incident and since there I never experienced something similar again, although I didn't replace anything.

If your theory would be correct the problem would have shown up again for sure, especially as meanwhile I had flights in much stronger wind.

It's also possible that there is more than a single factor at play here. The OP in this thread seems to have shown quite conclusively that deformed props can cause this problem. Additionally, several logs posted here also show unreasonably high rear motor speeds that would pitch the aircraft forwards if the props were providing normal lift. But your example does suggest a different mechanism might be responsible in that case.
 
  • Like
Reactions: Peio64270
Sorry but I doubt at this "fact". I had the uncommanded descending just once with my MM. And it happened during flying home in P mode and making a short break with hoovering at an altitude of about 10m. When I detected the altitude loss I was able to compensate with more throttle and complete my return flight. There was only calm wind and no warnings about wind or power limits reached. After landing I replaced the battery and immediately took off for a further flight for verifying what just had happened. I was not able to reproduce the incident and since there I never experienced something similar again, although I didn't replace anything.

If your theory would be correct the problem would have shown up again for sure, especially as meanwhile I had flights in much stronger wind.
Or it might be another kind of uncommanded descent, the ones that we discuss here usually don't respond to more throttle well .
 
Last edited:
I dont have the MM yet but i do read this stuff to understand what the data says.
I believe in the prop issue. Makes total sense. But it seems clear that multiple factors can cause this behavior and a major difference is the ability to steer against the altitude loss. Which in the case of bad props is not helping.
I read about the barometric sensor , is there log data for that too?
 
  • Like
Reactions: m80116
My uncommanded descent happened after 380 flights, I spent the next 2 weeks trying to see if the error would happen again but nothing. If I had changed the propellers I would have said that fixed the issue but I did nothing which also apparently fixed the issue. Sent it back to DJI for assessment, they tested it and said it works fine and were about to send it back to me until they actually checked the flight data. The official response from DJI is 'We have the result of the flight data analysis, the motor is abnormal during the flight, we will replace the Mavic Mini drone for you...' Then I got the quote for a Power ESC Module and repair service fee for $86 in the email saying its a free service. So hope that detailed analysis clears up the issue for everyone
 
  • Like
Reactions: Peio64270
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
131,123
Messages
1,560,076
Members
160,099
Latest member
tflys78