What do I do when my Mavic Mini does a fly-away?

[rickmurray1989]

Well, I am certain there was no response to any joystick inputs. I was receiving clear video for this entire process until it did its RTH which I suspect only happened because the wind dropped and it had breached its geofence - would Airdata tell me this fact?

I basically need a way to mitigate the SHTF scenario caused by the "special" way DJI has their software react to wind. Analysing the logs is not going to get me out of the windy-fly-away-software-bug scenario.

What?

 

[old man mavic]

@mavic_minis_fly_away ,firstly welcome to the forum
i have owned an original MM since the 30-10-2019 and ordered mine direct from DJI on the day they were released
as many have found, as a first attempt at a sub250g drone ,combined with a new App as well ,the early experiences with the drone were quite entertaining ,i even had to do a front arm stiffener fix to stop them twisting ,if the wind was buffeting the drone around
and as many of us discovered ,its capabilities in windy weather ,were at best a bit hit and miss,
but for a first attempt at a small lightweight platform ,compared to what was out there at the time ,it was quite a game changer
and to be fair within a year of release DJI gave us the MINI 2 which answered a lot of the criticism's of the MM and was a much better drone on many fronts
to be fair i have flown several hundred flights with mine and it still flies great for what it is
as you know it only has advanced WIFI transmission between the RC and drone ,and in certain scenarios it is very easy for connection to be lost even momentarily and this has caught many people out ,from reading your account ,for me because of your location,at the time of the flight ,and what tall structures may have been around ,i feel that the drone went into ATTI mode due to a drop in satellite numbers ,which would then cause it to be blown along by the wind this would have prevented RTH from working correctly ,but after a while it regained sufficient Sat lock and then was able to return home,this is just my take on your post ,please converse with @Meta4 ,who will be able to give you a detailed analysis of what happened

 

[Meta4]

Flight data has been analysed and shows that the drone didn't fly away and performed exactly as it was supposed to but operator disorientation caused confusion about what was happening during the flight.

 

[Yorkshire_Pud]

I fundamentally could not make the drone move back in my direction. Does anyone know if killing the motors is possible in this scenario because RTH does nothing?

Killing the motor is possible providing that the drone is in control range but you need to change the response to the CSC from "Emergency Only" to "Anytime", p32 of the manual. If set to "Anytime" the motor will stop when the CSC position has been held for around 2 seconds.

Due to the short delay period it is unwise to use this "Anytime" setting as the 'norm', people have lost their drones because they have had "Anytime" set and inadvertently used the CSC position whilst flying.
So....whilst in a 'panic' you would have to remember to switch the response from "Emergency Only" to "Anytime" and then execute the CSC. Assuming the drone was initially high enough the result will be a free fall, which will probably end up at around 14 to 16 m/s.

 

[mavic_minis_fly_away]

Killing the motor is possible providing that the drone is in control range but you need to change the response to the CSC from "Emergency Only" to "Anytime", p32 of the manual. If set to "Anytime" the motor will stop when the CSC position has been held for around 2 seconds.

Due to the short delay period it is unwise to use this "Anytime" setting as the 'norm', people have lost their drones because they have had "Anytime" set and inadvertently used the CSC position whilst flying.
So....whilst in a 'panic' you would have to remember to switch the response from "Emergency Only" to "Anytime" and then execute the CSC. Assuming the drone was initially high enough the result will be a free fall, which will probably end up at around 14 to 16 m/s.

It's a hard choice. I actually changed that setting as soon as I got home and discovered that it existed.

Still chatting with Meta4 privately about the log as I am certain the drone did not respond to my inputs.

 

[mavic_minis_fly_away]

@mavic_minis_fly_away ,firstly welcome to the forum
i have owned an original MM since the 30-10-2019 and ordered mine direct from DJI on the day they were released
as many have found, as a first attempt at a sub250g drone ,combined with a new App as well ,the early experiences with the drone were quite entertaining ,i even had to do a front arm stiffener fix to stop them twisting ,if the wind was buffeting the drone around
and as many of us discovered ,its capabilities in windy weather ,were at best a bit hit and miss,
but for a first attempt at a small lightweight platform ,compared to what was out there at the time ,it was quite a game changer
and to be fair within a year of release DJI gave us the MINI 2 which answered a lot of the criticism's of the MM and was a much better drone on many fronts
to be fair i have flown several hundred flights with mine and it still flies great for what it is
as you know it only has advanced WIFI transmission between the RC and drone ,and in certain scenarios it is very easy for connection to be lost even momentarily and this has caught many people out ,from reading your account ,for me because of your location,at the time of the flight ,and what tall structures may have been around ,i feel that the drone went into ATTI mode due to a drop in satellite numbers ,which would then cause it to be blown along by the wind this would have prevented RTH from working correctly ,but after a while it regained sufficient Sat lock and then was able to return home,this is just my take on your post ,please converse with @Meta4 ,who will be able to give you a detailed analysis of what happened

It's a useful theory to bear in mind but I was flying in the open amongst fields.

 

[Yorkshire_Pud]

It's a hard choice. I actually changed that setting as soon as I got home and discovered that it existed.

Still chatting with Meta4 privately about the log as I am certain the drone did not respond to my inputs.

I STRONGLY suggest you return the setting to the default of "Emergency Only".
It is far too easy to inadvertently use the CSC stick position if you are 'chucking the drone about the sky' and, with a delay period that short, you are going to have to be lucky to realise what you are doing before the shutdown occurs.
If you do shut the motors of a Mavic Mini down accidentally in mid air then you have, I think, little chance of restarting the motors before the drone hits the ground and you would need a calm mind to a) realise what you had done and b) make the appropriate responses.
You would also need luck on your side because I fear a Mini will tumble in freefall.
If it does then the motors will only 're'start when the drone is within certain tilt limits and you would need to be holding a SECOND CSC whilst the drone passed through that tilt range. Then, even if you get a restart, you might need to apply full throttle to halt the fall.......i.e. you will need a fair amount of height.

 

[mavic_minis_fly_away]

I STRONGLY suggest you return the setting to the default of "Emergency Only".
It is far too easy to inadvertently use the CSC stick position if you are 'chucking the drone about the sky' and, with a delay period that short, you are going to have to be lucky to realise what you are doing before the shutdown occurs.
If you do shut the motors of a Mavic Mini down accidentally in mid air then you have, I think, little chance of restarting the motors before the drone hits the ground and you would need a calm mind to a) realise what you had done and b) make the appropriate responses.
You would also need luck on your side because I fear a Mini will tumble in freefall.
If it does then the motors will only 're'start when the drone is within certain tilt limits and you would need to be holding a SECOND CSC whilst the drone passed through that tilt range. Then you might need to apply full throttle to halt the fall.......i.e. you will need a fair amount of height.

Thanks for the warning. I will do a test soon to see how easy it is to kill the motors. There is a balance I must make between having my drone fly-away and me accidentally killing the motors. I'll probably go with the latter TBH as I fly in a field and so a free-fall event is preferable to a fly-away.

 

[Yorkshire_Pud]

Thanks for the warning. I will do a test soon to see how easy it is to kill the motors. There is a balance I must make between having my drone fly-away and me accidentally killing the motors. I'll probably go with the latter TBH as I fly in a field and so a free-fall event is preferable to a fly-away.

There is no need for you to test it, I have done it myself, the timings come from my own flights AND from the logs of people who have lost their drone as a consequence of an accidental midair motor stop.
You can take that as hard fact.

Besides

If you insist on testing it then switch the drone to the slowest flight mode ( to limit horizontal speeds) and start with the drone hovering at about 6 or 7 ft over cut but not short grass, the free fall about 2ft. Even then if you are unlucky and the drone lands badly you could tear the gimbal ribbon cable.

I will be blunt, if Meta 4 is correct then you would do better to learn how to respond correctly to problems rather than experiment with risky procedures for experimentation purposes alone, especially when they have already been documented in various forms.

 

[mavic_minis_fly_away]

There is no need for you to test it, I have done it myself, the timings come from my own flights AND from the logs of people who have lost their drone as a consequence of an accidental midair motor stop.
You can take that as hard fact.

Besides

If you insist on testing it then switch the drone to the slowest flight mode ( to limit horizontal speeds) and start with the drone hovering at about 6 or 7 ft over cut but not short grass, the free fall about 2ft. Even then if you are unlucky and the drone lands badly you could tear the gimbal ribbon cable.

Best option is for me to glue a guard to one of my joysticks that prevents moving the stick to the danger area. Then I can snap off it in case of emergency.

 

[Kilrah]

Flight data has been analysed and shows that the drone didn't fly away and performed exactly as it was supposed to but operator disorientation caused confusion about what was happening during the flight.

Who would have thought!

 

[Yorkshire_Pud]

Best option is for me to glue a guard to one of my joysticks that prevents moving the stick to the danger area. Then I can snap off it in case of emergency.

I think that would be inappropriate.

As I said, I have had at least two blow aways with my minis and, through following the recommended procedures, I 'saved' the drone on every occasion.

Situations where you will need a mid air shut down are, with the very short delay period and for normal flying, too few and far between to leave the CSC response set to "Anytime".

Anyway, it's your drone and your choice but, if you fly with "Anytime" set as the 'norm', I think you are heading for a wrecked drone.

 

[Meta4]

Best option is for me to glue a guard to one of my joysticks that prevents moving the stick to the danger area. Then I can snap off it in case of emergency.

DJI selected the CSC joystick position because it is one that you will never accidentally activate in normal flight.
In >8 years monitoring drone forums I can only think of two or three incidents where it happened "accidentally".
But in each case, the flight data showed that it was a result of extremely reckless flying and not really an accident.

 

[mavic_minis_fly_away]

A big thank you to Meta4 for his expertise and patience. It is absolutely clear to me now what the logs show.

I still intend to test my mad inverted stick input blow away theory but that's just so I can prove to myself that I am indeed a complete idiot. In the meantime, I am off to rehearse my RTH and do-not-panic skills. Cheers!

 

[mavic_minis_fly_away]

There is or was something in the drone's firmware that limits speed when the gimbal is at certain angles. It's not a bug, there are reasons for it but and it may also appy to the Mini 2 & SE.
You need to be aware of this behaviour and make appropriate adjustments as and when circumstance require.
The gimbal's pitch angle was one thing I wanted to look at in your log.
Firmware update may have chnges the behaviour but I myself do not recollect encountering a problem with it.

Gimbal was at -60 degree pitch most of the time while flying backwards FWIW.

 

[Yorkshire_Pud]

Ahhh, from p11 of the v1.4 manual of the Mini 2.
"During video mode in Normal or Cine mode, the flight speed is limited when the pitch of the gimbal is near -90° or 0° in order to ensure shooting is stable. If there are strong winds, the restriction will be disabled to improve the wind resistance of the aircraft. As a result, the gimbal may vibrate while recording."

Given the similarities between the drones it maybe that it also applies to the Mavic Mini.
I'd need to look through my logs to possibly confirm that one way or the other. But it also seems that the limitation may have been lifted in strong wind.

If you are flying backwards then the drone's nose is raised and the amount of tilt of the gimbal, with respect to the drone, will be increased so as to maintain the 60° with respect to the gravitational horizontal.
It's quite a while since I read a thread concerning this and I do not remember if the problem affected both forwards and backwards flight but which ever direction is concerned the drone's pitch and therefore air speed was limited.

 

[Meta4]

Ahhh, from p11 of the v1.4 manual of the Mini 2.
"During video mode in Normal or Cine mode, the flight speed is limited when the pitch of the gimbal is near -90° or 0° in order to ensure shooting is stable.

In this incident the gimbal was not at that angle and there was no problem with the drone's speed.

 

[Zbip57]

A big thank you to Meta4 for his expertise and patience. It is absolutely clear to me now what the logs show.

Meta4 can be relied on every time.

A pilot's "opinion" of what may have happened is rarely an accurate account of what actually happened. When something totally unexpected occurs during a flight, and panic sets in, that often results in a distorted memory of which button was pushed when or how often.

The flight logs record only factual data. Sometimes they're difficult to decipher, but there are some really smart people on this forum who specialize in figuring out what really happened, and they are usually able to back up their diagnosis with actual data.

We all benefit from learning the outcome and conclusions from such incidents when people are brave enough to share them on this forum. Hopefully similar incidents can then be avoided by others, rather than repeated.

 

[Zbip57]

... from p11 of the v1.4 manual of the Mini 2.
"During video mode in Normal or Cine mode, the flight speed is limited when the pitch of the gimbal is near -90° or 0° in order to ensure shooting is stable."

In this incident the gimbal was not at that angle and there was no problem with the drone's speed.

Understood, this incident was not related to the gimbal pitch issue.

Just to elaborate on that (unrelated) issue, I did a video with my Mini-1 a while ago to demonstrate what this means. The gimbal has a limited mechanical range of motion. Under certain flight manoeuvres involving large pitch angles of the aircraft (i.e. high speed), the camera gimbal may hit its mechanical range limits. The firmware then kicks in to rotate the gimbal pitch angle safely away from those bump stops. That creates an annoying jump in the video.

DJI "fixed" this problem by limiting the flight speed (i.e. pitch angle) of the aircraft, whenever the gimbal is positioned near the limits of it's pitch range (i.e. straight down, or all the way up). That is meant to prevent the camera gimbal from rotating far enough to ever contact the bump stops.

 

[mavic_minis_fly_away]

Just a follow up reply. I tested the emergency motor cut out and it actually takes a few seconds to activate which is a relief.

I did a test flight in 20-25mph winds so I understood better how to handle them. I discovered a few things:

1) RTH in high winds will really ruin your day because the drone will ascend to an unsafe and even windier altitude. i.e. Descending manually with the stick is your best option.

2) In P mode, blow-away is a thing; beyond a certain distance (about 30m??) the drone just stays at that blown-away position rather than fly back to the position it was trying to hold. For less strong winds, it returns to the position it was trying to hold until the next strong gust.

3) Confirmed to myself that I was an idiot and the logs before were not lying about stick position. This has been a sobering and personal reality-changing experience!