Help! Mini Midair FreeFall (Video and Logs posted)

[boblui]

There are two flights, the first one was made with DJI FLY and the second one with Litchi. In which flight was the abnomality observed and at what time ?

 

[JohnNokomis]

The crazy break dancing happened in the Litchi flight at 4:41 into the flight. The AC went from 25 mph to 0.5 mph in 1.9 seconds.

 

[boblui]

The crazy break dancing happened in the Litchi flight at 4:41 into the flight. The AC went from 25 mph to 0.5 mph in 1.9 seconds.

For some reasons the Litchi flight began at 6:08. so I assume that you are referring to the first flight, ie. the one done with the FLY app.

At 4:41, the craft was in sport mode and the elevator stick was moved from max forward to center so the craft braked hard. Such deceleration is just normal.

I think the log data is a bit messed up especially for the Litchi flight. Both the Litchi log and craft .DAT log shows that the craft rolled with no aileron input :

Litchi :



Craft DAT :

 

[slup]

Looked over the flight you did with the DJI FLY app, trying to find signs of uncommanded movements. On several occasions the Mini suffers from a yaw twitching up to 30-60 degrees with only max elevator input.

Looked deeper into this occasion ...



Here below the CCW yawing is seen ... the movement is real as both the compass & the gyro agrees to this. Plotting up the motor rpm's during the same time shows the probable cause ... the rear left motor is maxing out, probably due to having been deformed & doesn't grip the air as it should ... meaning not generating a proper CW torque. That results in that the FC commands both CCW torquing props (front left & rear right) to slow down to counteract the CCW yawing.



Zooming out on the whole flight we clearly see that the rear left motor (dark green) is constantly touching max revs ...



And here we see that the FC commands it much higher than the other 3 motors (green) ... it's on 100% command every time you pitch for speed in Sport mode.



Looked also into how the speeds & pitch angles was affected by this ... but strangely they were kept within the Mini specs. Possibly this can be a kind of borderline case where the prop isn't enough deformed ... or actually flattens when pushed for higher speeds. In this example the yaw was affected ... but with this defect both roll & pitch anomalies can be seen as in your previous flight.

My advice is to change the props to a fresh set & try again ...

 

[Starz]

Will be interesting to see if this is just a bad set of props, or a storage /deformation issue.

 

[BudWalker]

Looked over the flight you did with the DJI FLY app, trying to find signs of uncommanded movements. On several occasions the Mini suffers from a yaw twitching up to 30-60 degrees with only max elevator input.

Looked deeper into this occasion ...

View attachment 111929

Here below the CCW yawing is seen ... the movement is real as both the compass & the gyro agrees to this. Plotting up the motor rpm's during the same time shows the probable cause ... the rear left motor is maxing out, probably due to having been deformed & doesn't grip the air as it should ... meaning not generating a proper CW torque. That results in that the FC commands both CCW torquing props (front left & rear right) to slow down to counteract the CCW yawing.

View attachment 111930

Zooming out on the whole flight we clearly see that the rear left motor (dark green) is constantly touching max revs ...

View attachment 111931

And here we see that the FC commands it much higher than the other 3 motors (green) ... it's on 100% command every time you pitch for speed in Sport mode.

View attachment 111932

Looked also into how the speeds & pitch angles was affected by this ... but strangely they were kept within the Mini specs. Possibly this can be a kind of borderline case where the prop isn't enough deformed ... or actually flattens when pushed for higher speeds. In this example the yaw was affected ... but with this defect both roll & pitch anomalies can be seen as in your previous flight.

My advice is to change the props to a fresh set & try again ...

But the leftBack isn't always commanded at 100% when elevator is full forward. And, sometimes it's the rightBack that is commanded 100%.


Seems that the roll input has some effect in determining if the leftBack and/or rightBack is commanded 100%. But, it isn't totally conclusive as in this interval


There are so many inputs that I and CsvView just aren't smart enough to see an example of what the OP is referring to. @JohnNokomis would it be possible to provide a .DAT where the problem is presented but with as few control inputs as possible. Congratulations though on retrieving the fc_log.log. That's some pretty tricky business. If you're using the Fly App then the mobile device will provide an adequate .DAT. That is, if DJI hasn't synched it away.

DatCon isn't smart enough to know about the time scale in the fc_log.log file. But, the osd_data:flightTime will be identical to that in the .txt log file.

 

[BudWalker]

This doesn't appear to be an uncommanded descent incident. There are no vertical speed commanded vs vertical speed actual discrepancies


Here is an example in a different incident
Mavic Mini uncommanded descent tests
where an uncommanded descent occurred

 

[slup]

But the leftBack isn't always commanded at 100% when elevator is full forward. And, sometimes it's the rightBack that is commanded 100%....

There are so many inputs that I and CsvView just aren't smart enough to see an example of what the OP is referring to....

...retrieving the fc_log.log...

Yeah, the first time I look into a fc_log.log, nice to see that it is an alternative to the mobile device DAT where that is missing ... really lagged down my win computer though, took ages to switch signals in CsvView And as you say, it was slow to find the anomalies, couldn't see any similar behavior as in the previous flight (the first in the thread) where both yaw, roll & pitch went crazy ... here the yaw misbehaves instead. Saw that the rears switched place sometimes regarding the command ... but overall the left seems to bear the worst grunt.

This doesn't appear to be an uncommanded descent incident. There are no vertical speed commanded vs vertical speed actual discrepancies

No that's right ... the flights presented here doesn't have any height instability at all ... & the flight from the fc_log.log doesn't show any roll or pitch instability either. It's this together with the high motor rpm rear that make me think that this possibly can be a kind of borderline case where the lift not yet have been affected but instead affecting the CW/CCW torque enough to affect the balance ... making the yaw twitch.

 

[JohnNokomis]

Ok I have just refreshed the firmware on both the AC and RC using DJI Assistant 2 on my PC. I'm about to change out all props then take it on a test flight. It's weird because I have never recieved a motor speed error like I have in the past. I really appreciate y'all taking time to look through my flight logs.