Mavic mini 1 stuck at full throttle

[Yorkshire_Pud]

Thank you for that. Have in mid that at this time the wind was quite strong atm. That can cause this behavior? Also do you have any sugestion what can cause this, and any possible fix? Thanks again!

Wind will cause the motors to change speed, the drone changes tilt to fight wind and changing tilt requires individual motors to change speed.
Additionally, when a drone is tilted less of the propellers' thrust is available to keep the drone at the desired height. Basically that means that the motors have to turn faster so that the vertical component of the propellers' thrust matches the drone's weight. The greater the tilt the faster the motors must turn to keep the drone at the desired height.

BUT what I take from the above traces is that the motors went crazily fast whilst the barometer was indicating the erroneous descent to negative heights i.e. I think the control system was trying to fight what it thought was an uncommanded descent.

Hmmm "there are none so blind as those that will not see" I have just realised that the log reports that during the 'descent' of around 217 -220 the log 'thinks'/shows that the throttle was closed during that descent, hmmmmm???.

Looking at the csv downloaded from PH. At 216.9 sec the drone was at 11.8ft and the throttle was centred, according to the log you then partially closed the throttle for around 0.2 sec then centred it again, that partial throttle started the drone descending. ( The VPS height 'agrees' with the barometer height between 216.9 sec and 217.2 sec and possibly 217.3 sec but I think VPS malfunctions after 217.3 sec. Rhetorically, why would the VPS malfunction at that moment? )
HOWEVER, the drone continues to 'descend' for the following 0.5 sec whilst the throttle was centred. The descent should have stopped within that 0.5 sec but the descent rate "OSD.Zspeed [MPH]" actually increases.

We are now around 217.6 sec - 217.7 sec with the drone 'very close' to ground level.
According to the log at 217.7 sec you started to close the throttle and, ignoring the centred blip at 220.7 sec, held the throttle partially closed or fully closed until 222 sec when the drone was supposedly at -86.8ft (not its lowest point).

Obviously something is wrong here since the ground was flat and negative heights were not possible but I am puzzled/confused as to why the log 'thinks' you had the throttle closed.

 

[StaluUu]

Wind will cause the motors to change speed, the drone changes tilt to fight wind and changing tilt requires individual motors to change speed.
Additionally, when a drone is tilted less of the propellers' thrust is available to keep the drone at the desired height. Basically that means that the motors have to turn faster so that the vertical component of the propellers' thrust matches the drone's weight. The greater the tilt the faster the motors must turn to keep the drone at the desired height.

BUT what I take from the above traces is that the motors went crazily fast whilst the barometer was indicating the erroneous descent to negative heights i.e. the control system was trying to fight what it thought was an uncommanded descent.

Hmmm "there are none so blind as those that will not see" I have just realised that the log reports that during the 'descent' around 217 -220 the log 'thinks'/show that the throttle was closed during that descent, hmmmmm???.

Looking at the csv downloaded from PH at 216.9 sec the drone was at 11.8ft and the throttle was centred, according to the log you then partially closed the throttle for around 0.2 sec then centred it again, that partial throttle starts the drone descending. ( The VPS height 'agrees' with the barometer height between 216.9 sec and 217.2 sec and possibly 217.3 sec but I think VPS malfunctions after 217.3 sec. ) HOWEVER, the drone continues to 'descend', for the following 0.5 sec whilst the throttle was centred, the descent should have stopped within that 0.5 sec but the descent rate "OSD.Zspeed [MPH]" actually increases.

We are now around 217.6 sec - 217.7 sec with the drone 'very close' to ground level.
According to the log at 217.7 sec you started to close the throttle and, ignoring the centred blip at 220.7 sec, held the throttle partially closed or fully closed until 222 sec when the drone was supposedly at -86.8ft (not its lowest point).

Obviously something is wrong here since the ground was flat and negative heights were not possible but I am puzzled/confused as to why the log 'thinks' you had the throttle closed.

hello again, About my thorttle control:
- i flew easy for couple of minutes without error
- Then i want to check if full throtrle ( 3.39 on youtube video) will still affect mavic and lock it at full throttle, it does.
- I let of the stick in full throttle for around 2 second, mavic makes weird dance in the air (on 3,40 youtube)
- i let of the stick but mavic is keep gaining altitude (i cound make some stick movments but i doub and i prob. just left the stick and pray) or i just spam throttle down, dont realy remember .
- Then when all stopped, and mavic stopped gainign altitude i hold throttle down for landing by my own, and i landed the drone.

 

[BudWalker]

@StaluUu thanks for providing the .DAT.

The problem appears to be a HW fault in the Z axis accelerometer. There are several data that indicates the Mini was ascending, including the pilot's description. However, IMU:alti (red) is showing the Mini is descending.


IMU:alti is the result of the Flight Controller's fusing algorithms and is dependent on IMU:velD. Comparing IMU:velD against GPS:velD it can be seen that IMU:velD is incorrect.


Finally, IMU:velD is dependent on IMU:accelZ

If the Mini is accelerating upwards then accelZ should be negative. Also, the accelZ data starting about 222.3 secs looks a bit suspect.

Looks to me like the Z axis accelerometer gets stuck positive causing the FC to think the Mini is loosing altitude. To compensate the FC commands higher thrust causing the Mini to ascend.

The questions remains what causes the accelZ to get stuck. I could speculate but what would be the point? It's clear, at least to me, that the problem is a HW fault.

 

[Yorkshire_Pud]

Didn't/doesn't the P3 suffer an occasional stuck accelerometer? I seem to remember that but think it prevented launch or gave a warning at the start of the flight. I recollect the thinking was a hard knock or long storage could cause the problem.

Whatever the P3 problem was, for me, the fix was to give the drone's body a gentle smack. It's been a while since I flew a P3 and longer since I encountered the problem and it's not applicable to this flight but .... just reminiscing.

 

[BudWalker]

Didn't/dosen't the P3 suffer an occasional stuck accelerometer? I seem to remember that but think it prevented launch or gave a warning at the start of the flight. I recollect the thinking was a hard knock or long storage could cause the problem.

Whatever the P3 problem was, for me, the fix was to give the drone's body a gentle smack. It's been a while since I flew a P3 and longer since I encountered the problem and it's not applicable to this flight but .... just reminiscing.

Yes, I remember those days. But, wasn't it the case that the accelerometer was always stuck? Before any flight was attempted. Not after a flight had been in progress.

 

[Yorkshire_Pud]

But, wasn't it the case that the accelerometer was always stuck? Before any flight was attempted. Not after a flight had been in progress.

That is my recollection too, which is why I said "but think it prevented launch or gave a warning at the start of the flight."

 

[Zbip57]

.... just reminiscing.

Ah, the good old days...

It was the accelerometer sticking in the Naza-M flight controller of the original Phantom. The solution was to give to the Naza a good whack.

 

[StaluUu]

Hello, sorry for long time no reply. Meantime I was very sick, also ther was heavy winds in Poland, and I was on holidays

As an engineer I troubleshot the drone i made frame support with help of 3d printer i mounted mavic 1m above the ground and start to full throttle it. I notice heavy vibration when on full throttle. Previous owner swapped the props but swap them like that half prom stock DJI, half AliExpress copy that caused unbalance on the drone (hard to notice on first eye glance). I swap all props for fresh DJI stock. Also i figure out that a small plastic piece was missing near propeller that can cause pressure difference when on full throttle. Later i mounted fishing line to mavaic and started testing it outside (yes, I flew the drone with line like a kite ), and the problems went off. In my opinion, mostly the vibration due to different props were messing with drone due to ESC error. Thank you all for your help. I hope someone will come across this post and fix his own drone. BR, Michał

 

[Yorkshire_Pud]

Hello, sorry for long time no reply. Meantime I was very sick, also ther was heavy winds in Poland, and I was on holidays

As an engineer I troubleshot the drone i made frame support with help of 3d printer i mounted mavic 1m above the ground and start to full throttle it. I notice heavy vibration when on full throttle. Previous owner swapped the props but swap them like that half prom stock DJI, half AliExpress copy that caused unbalance on the drone (hard to notice on first eye glance). I swap all props for fresh DJI stock. Also i figure out that a small plastic piece was missing near propeller that can cause pressure difference when on full throttle. Later i mounted fishing line to mavaic and started testing it outside (yes, I flew the drone with line like a kite ), and the problems went off. In my opinion, mostly the vibration due to different props were messing with drone due to ESC error. Thank you all for your help. I hope someone will come across this post and fix his own drone. BR, Michał

Attached is a partial plot of motor rpm for a Mavic Mini on a day where I have the log marked WIND WIND WIND, these were the highest rpm's on that flight. As you can see one motor reached 14,800 rpm, I do not remember if that if the fastest I have seen but hover rpm is around 10,000 rpm. If you work out the centripetal accelerations they are quite impressive, at the blade tip, 14,691g @ 14,800rpm, 6707g @ 10,000rpm. A little bit of imbalance in the wrong place gives a fair bit of force at those speeds