Mavic Mini Lost in the Ocean

[Anomander Rake]

Hi all

My Mini decided to descend today whilst I was manually flying the drone to my location. The drone was over water (Sydney beach) but started descending and not moving forward as I was asking it to. It then hit 20% battery but I hadn't enabled RTH so it was all manual. Almost nothing I did could control the drone. I was in S mode, and I was getting readings on my phone of about 0.3m/s speed in fairly calm winds at that altitude. I would control it for 2-3 seconds and then it would keep descending for another few seconds.

I've read the sticky and I've pulled the TXT and LOG. Can someone smarter than me help me figure out if this was a drone error or user error? The drone is gone for good I'm afraid.

thank you for your help

 

[slup]

Unfortunately yet another example of an "uncommanded descent" showing alot of flight instability & maxed rear motor rpm's with the Mini involved ... @sar104 , here you have one with the DAT preserved to your collection.

Alot of wind warnings in the log ... Airdata doesn't show winds over the 8m/s spec though ...



Again the descent starts with pitch/roll instability just after you commanded positive elevator & heading speed starts to pick up ... then suddenly the Mini pitch up, slows down & the down (Z) speed picks up.



Looking into the DAT log shows as usual rear motors near to maxed out & a decrease of front rpm's when the pitch up occurs ... this with full elevator commanded, it's like if the Mini tries a last resort to maintain the negative pitch angle when it can't increase the rear motors anymore ... it lower the fronts.

 

[Meta4]

I've read the sticky and I've pulled the TXT and LOG. Can someone smarter than me help me figure out if this was a drone error or user error?

You descended from 15:34.8 till 15:38.2 and then left the left stick centred.
But the drone continued to descend with no throttle input.
You pushed the left stick up a short time at 15:41.5 but the descent continued.
At 15:47.4, more left stick raised the drone slightly but at 15:57.1 the descending resumed with no left stick input.
Left stick up at 16:01.8 had no effect and the descent continued.
More left stick at 16:07.3 brought the drone up, but it started losing height again at 16:15.7, ignored more left stick up from 16:24.1 and splashed down at 16:29.4.

This is just another of many uncommanded descents causing the loss of a Mavic Mini.
Too many of these have happened and the Mini is not safe to fly.
DJI need to come up with an urgent fix to the Mini disaster.

Even without the drone, DJI should replace yours under warranty as the log data clearly show a serious fault.

 

[sar104]

Unfortunately yet another example of an "uncommanded descent" showing alot of flight instability & maxed rear motor rpm's with the Mini involved ... @sar104 , here you have one with the DAT preserved to your collection.

Alot of wind warnings in the log ... Airdata doesn't show winds over the 8m/s spec though ...

View attachment 94140

Again the descent starts with pitch/roll instability just after you commanded positive elevator & heading speed stars to pick up ... then suddenly the Mini pitch up, slows down & the down (Z) speed picks up.

View attachment 94141

Looking into the DAT log shows as usual rear motors near to maxed out & a decrease of front rpm's when the pitch up occurs ... this with full elevator commanded, it's like if the Mini tries a last resort to maintain the negative pitch angle when it can't increase the rear motors anymore ... it lower the fronts.

View attachment 94142

Yes - that's a good one. I can't add much to your analysis, but there are a couple of strange issues. Firstly - the left front motor current data are very noisy:



I had to filter those for the subsequent graphs, but you can see that the front left motor appears to be pulling significantly more current than the others. I don't think that's related to the main issue though.

Secondly, there is a fairly abrupt transition in behavior during the flight.



The rear motors are higher than they should be for most of the flight, but looking at the last couple of hundred seconds:



The initial descent, starting around 820 seconds, is well behaved. It has no associated horizontal motion, pitch and roll are small, and the rear motors are running at expected speeds. It all goes wrong at 890 seconds, when full forwards elevator is applied. The rear motors immediately max out, the front motors slow down to lower speed than during the previous controlled descent, and the pitch and roll go unstable.



This looks less like VRS and more like simple loss of aircraft control when the FC attempts to apply an attitude that cannot support the aircraft with under-performing rear props.

It also suggests that the emergency response to this condition should be to rotate the aircraft and fly backwards.

 

[BudWalker]

Yes - that's a good one. I can't add much to your analysis, but there are a couple of strange issues. Firstly - the left front motor current data are very noisy:

View attachment 94151

I had to filter those for the subsequent graphs, but you can see that the front left motor appears to be pulling significantly more current than the others. I don't think that's related to the main issue though.

Secondly, there is a fairly abrupt transition in behavior during the flight.

View attachment 94152

The rear motors are higher than they should be for most of the flight, but looking at the last couple of hundred seconds:

View attachment 94154

The initial descent, starting around 820 seconds, is well behaved. It has no associated horizontal motion, pitch and roll are small, and the rear motors are running at expected speeds. It all goes wrong at 890 seconds, when full forwards elevator is applied. The rear motors immediately max out, the front motors slow down to lower speed than during the previous controlled descent, and the pitch and roll go unstable.

View attachment 94155

This looks less like VRS and more like simple loss of aircraft control when the FC attempts to apply an attitude that cannot support the aircraft with under-performing rear props.

It also suggests that the emergency response to this condition should be to rotate the aircraft and fly backwards.

It's been a while and you probably forgot that the motorCurrent data is only partially correct. In particular, the leftFront always shows the problem you described. MotorCurrents are only retrieved when the Experimental option is chosen.

 

[sar104]

It's been a while and you probably forgot that the motorCurrent data is only partially correct. In particular, the leftFront always shows the problem you described. MotorCurrents are only retrieved when the Experimental option is chosen.
View attachment 94156

I had forgotten. Was there any conclusion on whether the current data are still approximately correct? They look reasonable, apart from left front.

 

[BudWalker]

I had forgotten. Was there any conclusion on whether the current data are still approximately correct? They look reasonable, apart from left front.

TBH, I haven't thought about it much. Except that the leftFront current data has the appearance of a SW glitch. Kinda like it's being over written with some other data.

 

[MAvic_South_Oz]

Welcome to the forum, pity it's not an introduction post, rather than this.

Can someone smarter than me help me figure out if this was a drone error or user error?

With the above you should be able to confidently open a ticket with DJI, lucky you.

See how you go sending them the Txt and Dat files, they should be able to see the same results as above, and should replace your mini.
If they give you any grief, send them a link to this thread !!
Also a link the Crash / flyaway section showing many Mini problem threads.

Personally, I would ask them how much to upgrade you to another aircraft, that you simply can't trust a Mini again after learning about its shortcomings.
Go to an Air or Spark, the mini is simply not safe or market ready.

If size is important to you for travel, then a replacement Mini and just be aware of possible issues, but it'd be hard to fly over water again, which is understandably a favourite with many pilots.

Good luck, and please, don't forget to come back and update the forum here with your dealings and results from DJI.

 

[Anomander Rake]

Thank you so much, everyone.
I've emailed DJI support; we'll see how we go.

Cheers

 

[steve495]

Thank you so much, everyone.
I've emailed DJI support; we'll see how we go.

Cheers

Please do keep us informed in this thread!

 

[Prismatic]

I was afraid that a most significant part of your post might be overlooked, down there at the end, so here it is:

<*snip*> ... It also suggests that the emergency response to this condition should be to rotate the aircraft and fly backwards.

The rotation isn't actually necessary, is it? If you have time, perhaps, but new pilots might botch an about-face during a panic. Just hit reverse! It's even possible that would revert the AC to "normal" behavior, but I wouldn't count on it. Once you regain control, you can fly it backwards all the way home.

 

[sar104]

I was afraid that a most significant part of your post might be overlooked, down there at the end, so here it is:

The rotation isn't actually necessary, is it? If you have time, perhaps, but new pilots might botch an about-face during a panic. Just hit reverse! It's even possible that would revert the AC to "normal" behavior, but I wouldn't count on it. Once you regain control, you can fly it backwards all the way home.

I was thinking in terms of actually flying it backwards to the home point.

 

[new2mavic]

I was afraid that a most significant part of your post might be overlooked, down there at the end, so here it is:

The rotation isn't actually necessary, is it? If you have time, perhaps, but new pilots might botch an about-face during a panic. Just hit reverse! It's even possible that would revert the AC to "normal" behavior, but I wouldn't count on it. Once you regain control, you can fly it backwards all the way home.

I remember seeing a post here saying that the poster could not get his Mavic ( forgotten whether it was a Mini ) fly forward but there was no uncommanded descent. He subsequently tried to turned the craft around and flew it backward. The craft was finally brought home that way.

 

[JAW]

Another of many uncommanded descents causing the loss of a Mavic Mini.
Too many of these have happened and the Mini is not safe to fly.

Pretty serious situation!

 

[Prismatic]

I was thinking in terms of actually flying it backwards to the home point.

I thought so, but wanted to clarify that rotation isn't required to halt an uncommanded descent, which is the top priority. Flying home is the next priority, but necessarily secondary.

I can't guess what percentage of Mavic Minis enter uncommanded descent, but it's too high. As others have said, DJI has a mess on their hands. They need to find a solution—in software or hardware—and institute a recall, much like car makers do when some vehicle has a defect in design or manufacturing.

Sooner or later this is going to result in something much worse than a drone sleeping with the fishes.

 

[sar104]

Sooner or later this is going to result in something much worse than a drone sleeping with the fishes.

No doubt.

 

[xspwhite]

Little Mavic Mini lost at sea.

 

[mavicn00b]

Little Mavic Mini lost at sea.

Mavic mini - burial at sea

 

[xspwhite]

Mavic mini - burial at sea

Maybe someone can make some little maciv mini sailor caps.

 

[mavicn00b]

Maybe someone can make some little maciv mini sailor caps.

And a memorial song - ode to the fallen ?