I have read about and seen videos about this before, but today I experienced it by myself.

My Mavic was on its way back to me in a distance of about 50m and an  altitude of about 12-15m. I stopped above grassland near to a small  street and then it begun descending for several meters quite quickly. I  was able to compensate with the throttle stick and fly back for landing.

At home I immediately uploaded the log file to Airdata.com for analysing  the incident. I found no error or warning messages, all seemed to be  ok. But when I took a close look to the barometric and sensor altitudes I  found out that the barometer reflected the descending while the sensor  altitude stayed same. Even when I manually increased altitude it took  quite long until the sensor altitude did react and increase, too.
Lighting was ok, sun was shining, so I have no other explanation that  either  a hardware problem of the sensor based altimeter or a serious  firmware bug.

My Mavic Mini is about 2 months old, with most recent firmware and had no issues during flights until today...                

Are you able to share the link to the log so others can take a look and possibly help you analyze?

Was it cold? some users flying in cold weather experienced this.
Posting the Log can help you further in getting some more info.

Yes, if you are able too, please share the log so it can be analyzed for you and myabe figure out what went on.

In my thought it was caused by that Authorization bug... but we'll only know after log

Hello and good day GerdS. I am sorry for the trouble this has caused and thank you for reaching out. Can you please try to fly your DJI Mavic Mini to a different location to see if the issue will still persist? There might have been some interference on the place where you have flown your drone previously that is why this issue happened. Please keep us posted on the said issue for us to be able to assist you further. Thank you.

Have added the log files. Interesting part is near to the end starting at about 16m 06.7s.
There is nothing special at the location I have flown today. Best weather conditions, sunny, calm wind, about 5°C.

It was the second flight with the battery not fully charged, because of just before with the first flight with this (full) battery I had the feeling of a uncommanded sideways drifting, so I came back for landing and doing a compass calibration to be on the safe side.

GerdS Posted at 12-30 12:49
Have added the log files. Interesting part is near to the end starting at about 16m 06.7s.
There is nothing special at the location I have flown today. Best weather conditions, sunny, calm wind, about 5°C.

zip doesn`t open, error.

Upload your log using >https://www.phantomhelp.com/LogViewer/Upload/

cheers
JJB

JJB* Posted at 12-30 13:20
zip doesn`t open, error.

Upload your log using >https://www.phantomhelp.com/LogViewer/Upload/

The zip opened ok for me but it only goes upto 9 Mins 33 Secs.

grahamjohnson10 Posted at 12-30 13:23
The zip opened ok for me but it only goes upto 9 Mins 33 Secs.

When inspecting my log file at Airdata.com the log starts with 09m 10.9s, because there was a previous flight without power cycling before, see here: https://app.airdata.com/share/HROLtk

GerdS Posted at 12-31 01:56
When inspecting my log file at Airdata.com the log starts with 09m 10.9s, because there was a previous flight without power cycling before, see here: https://app.airdata.com/share/HROLtk

Can you post the Airdata link?
I can't make any sense of what's in that zip file.

See my last post for the link.

when I took a close look to the barometric and sensor altitudes I  found out that the barometer reflected the descending while the sensor  altitude stayed same. Even when I manually increased altitude it took  quite long until the sensor altitude did react and increase, too.
Where are you finding your "sensor altitude" data?
What time in the flight data do you see sensor altitude reacting slowly to increasing altitude?

Here again all my log files of that incident, hopefully complete this time.

https://www.dropbox.com/s/3ww4wy21t0dsio3/Logs.zip?dl=0

This clearly sounds and look like my issue yesterday, which I could replicate today myself after refreshing all FW, APP and performed COMPASS and IMU calibration as suggested by users and DI support.

Hopefully DJI will analize this further to narrow the issue and apply a FW fix.

GerdS Posted at 1-3 09:53
Here again all my log files of that incident, hopefully complete this time.

https://www.dropbox.com/s/3ww4wy21t0dsio3/Logs.zip?dl=0

Link to Phantom Help Output of your Flight record .txt file

DJIFlightRecord_2019-12-30_[14-33-33]

Can you point out the previous case you say you have seen?

i've had this happen way too many times where the drone descends all by itself. today the app crashed  while recording and this happened. i gained altitude  when i lost the app and while reloading the app you will see it descend almost in the water. fix it dji!!!!

It is possibile the VPS stopped further descent. It was already dangerously low... with no idea how fast it was traveling and for how long I con only suggest (as indicated in the manual) to stay well clear of the water.

Anyway... had that happened to me (fly APP crash) being so low I'd have immediately initiated an altitude increase of at least 5 seconds. I am quite accustomed to this behavior as I have a problem with my RC and I am in for a USB OTG cable change.

Today I have experienced even two uncommanded descendings in just one flight! Luckily I was able to get altitude again by full throttle and forward speed instead of falling into an acre.

DJI, what is the matter with you, why does this critical bug still not get fixed by updated firmware?

With this bug I cannot trust into the Mini for flying over critical regions, e.g. water.

I have uploaded the logs to https://app.airdata.com/share/rgvsII. The descendings begin at 1m40s and 8m11s.

Here I have uploaded the .log and .txt files: https://drive.google.com/open?id ... 8DAHUsukXfRjbjydP2U

GerdS Posted at 4-16 10:41
Today I have experienced even two uncommanded descendings in just one flight! Luckily I was able to get altitude again by full throttle and forward speed instead of falling into an acre.

DJI, what is the matter with you, why does this critical bug still not get fixed by updated firmware?

Hi Gerds,

a typical example of this MM phenomenon, in sport mode with rapid pitch and up/down stick inputs you will see this 'uncontrolled' height changes.

This is not the uncontrolled descent into crash (up stick and no never a up reaction...) but IMO beacuse of the underpowered MM in power and processing power.

I have the same and depends on the wind how much it will climb or decend flying such moves.

But in normal mode my and guess your MM fly`s perfect. It up to all MM owners if they accept this or not.

cheers
JJB

JJB* Posted at 4-16 11:03
Hi Gerds,

a typical example of this MM phenomenon, in sport mode with rapid pitch and up/down stick inputs you will see this 'uncontrolled' height changes.

Both descendings happened while almost hovering, there was no aprupt manoevering before. I was just in Sport mode because flying with goggles (no easy access to the tocuscreen), so I won't run into a problem with wind. When I gave full throttle the Mini was very sluggish on ascending, almost ignoring the command.

JJB* Posted at 4-16 11:03
Hi Gerds,

a typical example of this MM phenomenon, in sport mode with rapid pitch and up/down stick inputs you will see this 'uncontrolled' height changes.

@JJB Would you mind looking at this Flight Log.


That was from that gator flight video, my mini is still set to 40° of pitch and I couldn't reproduce others drops and went on to flying normally. I figured having a skilled eye look it over would be a good idea to find out if my mini is also experaincing these height drops.


I have the ,Dat if you need it.

ABeardedItalian Posted at 4-16 11:49
@JJB Would you mind looking at this Flight Log.

Hi,

Great flight, just minor minor red lines on the blue altitude line.
Enjoy your good flying MM!

cheers
JJB

GerdS Posted at 4-16 11:15
Both descendings happened while almost hovering, there was no aprupt manoevering before. I was just in Sport mode because flying with goggles (no easy access to the tocuscreen), so I won't run into a problem with wind. When I gave full throttle the Mini was very sluggish on ascending, almost ignoring the command.

oke, what happend (IMO) your craft did climb during forward speed, this with no UP stick at all.
After releasing FWD stick MM descend to its original height before this uncontrolled climb.

I see this more in logs flying in sport mode with full or almost 100% fwd stick.
Guess than manual up steering when craft wants to descend gives a sluggish respond.

Hope that DJI take this serious !

BTW my own MM does the same sometimes, flying with some wind is factor for climbing like this.

cheers
JJB

GerdS Posted at 4-16 10:41
Today I have experienced even two uncommanded descendings in just one flight! Luckily I was able to get altitude again by full throttle and forward speed instead of falling into an acre.

DJI, what is the matter with you, why does this critical bug still not get fixed by updated firmware?

Your Mini data shows an extremely unreliable drone.
There's a lot more evidence of problems in the data than you described.
It cannot maintain altitude.
Sometimes it climbs very slowly and other times it loses altitude despite giving it full throttle.

I started to analyse the data but gave up after a while because there are problems showing all through it.
Here's what I found.
At 1:12.7 you gave the Mini full throttle in Sport Mode and held full or high throttle until 1:35.7.
The climb was very sluggish - only 0.5 m/sec.
From 1:35.7 you left it hovering and initially it gained 4.4 metres but then began to lose height and dropped 8.4 metres before you applied throttle which slowed the descent briefly.
The loss of altitude continued when the left stick was centred again.

From 3:01 you applied throttle and 30 seconds later the drone had climbed less than a metre.
From 3:29.9 you went full throttle till 3:46.5 but the Mini sank 8.2 metres

This drone is not safe to fly and not fit for purpose.
You cannot have any confidence in it and should demand a refund.

I had read about this issue and had my first encounter with this bug today and nearly lost my Mavic Mini for good.  I have flown extensively in colder conditions over frozen lakes, snow, ice etc without issue until today. It was +10 °C, sunny with a light breeze. 13+ satellites, no compass or IMU issues, no wind warnings until it was about to land on its own.

I flew up over the river and hovered while checking safety settings and it was holding altitude fine. I noticed even with full throttle it was not climbing very fast as I flew at about 140 feet along the river and then stopped again. As I began flying straight I noticed it was dropping altitude on its own which I have seen from another user on Youtube.  I realized that it was still falling despite full throttle input and tried switching to sport mode for more power but it had no effect. I scrambled for a place to land and managed to control the drift enough to put it over a snow ledge along the river bank where it proceeded to hit the snow and raise again before it was set down. I recovered it and it had no damage. I swapped to a fresh battery, completed a compass calibration as I thought maybe there was some kind of magnetic issue in the area and let it spin up and idle for a bit to see if there would be any errors and there were none. I let it hover, did some turns and hard inputs in sport mode without issue then put it back in p mode and climbed a bit and it seems to be ok.

I began descending under control and let off the stick to complete a hand landing however it just continued to lower into the ground on its own and bounced off the snow and climbed again. I was able to raise it up and then hand land it.
I synced my logs with DroneLogBook.com however the flight log stops after about half way into the flight for some reason. For now, I have uploaded the video and a screen recording of the DJI Flight Log replay showing the on-screen data and stick input for reference.

AIR DATA LINK: AIR DATA
Hover test after battery swap / compass cal: Air data 2

Max power limit reached warnings while descending from 13m. This is the log the corrosponds to the second video: Air data 3



Issues start about 1:45 on the screen recording. (Flying forward and descending on its own)

I will play the devil's advocate here. The specs show it to be good down to 0* C. Obviously the snow/ice below it is lower than that. Maybe it was not really 10*C ? Combine that with the surface below being moving water or featureless snow, maybe too much to handle?
The more experienced guys here are busy conversing in a useless political fling-fest topic right now, so I am afraid I am all you have ;-) No 'fence.

GaryDoug Posted at 4-16 18:33
I will play the devil's advocate here. The specs show it to be good down to 0* C. Obviously the snow/ice below it is lower than that. Maybe it was not really 10*C ? Combine that with the surface below being moving water or featureless snow, maybe too much to handle?

It was above zero most of the day. It is now dark out and still +4 degrees. The snow is melting. People are out running in t-shirts (albeit this is Canada). I have flown 4 or 5 batteries in -10 to -15 extensively over snow without any problems. The temperature above zero should not be an issue for this aircraft. Why would it just thunder into the ground with good sats and not respond to throttle input to climb? Even if it was due to being over snow, surely the VPS sesors are not programmed to take priority over all stick and GPS input.. I was getting 3.8 volts or so while in the air. No voltage errors due to weather - which I have seen before during a hover test in about -25 months ago.

I didn't say the devil was right ;-)

tyler947 Posted at 4-16 18:13
I had read about this issue and had my first encounter with this bug today and nearly lost my Mavic Mini for good.  I have flown extensively in colder conditions over frozen lakes, snow, ice etc without issue until today. It was +10 °C, sunny with a light breeze. 13+ satellites, no compass or IMU issues, no wind warnings until it was about to land on its own.

I flew up over the river and hovered while checking safety settings and it was holding altitude fine. I noticed even with full throttle it was not climbing very fast as I flew about 140 feet along the river and then stopped again. As I began flying straight I noticed it was dropping altitude on its own which I have seen from another user on Youtube.  I realized that it was still falling despite full input and tried switching to sport mode for more power but it had no effect. I scrambled for a place to land and managed to control the drift enough to put it over a snow ledge along the river bank where it proceeded to hit the snow and raise again before it was set down. I recovered it and it had no damage. I swapped to a fresh battery, completed a compass calibration as I thought maybe there was some kind of magnetic issue in the area and let it spin up and idle for a bit to see if there would be any errors and there were none. I let it hover, did some turns and hard inputs in sport mode without issue then put it back in p mode and climbed a bit and it seems to be ok.

Your Mini flew normally until 2:26.1 when it started losing height.
It lost 35 ft before you reacted by giving it full throttle at 2:33.3 but the Mini lost another 20 ft despite full throttle being applied.
It appeared to recover briefly but at 2:44.4 it was losing height again, and again kept losing height despite full throttle being applied.
It stabilised at 2:48.4 but started sinking again 10 seconds later and again failed to respond to full throttle input, sinking from 100 ft to 37 ft before finally climbing very slowly for a bit and then losing height (still with full throttle) from 3:21.7.
This had nothing to do with any external influence on the drone.
It was entirely due to the fatal flaw that is built-in to the Mavic Mini.
Too many of them cannot maintain altitude.

Your Mini is an unreliable lemon, like many others.
I'm amazed that DJI released such a flawed drone.
I'm more amazed that despite four months of daily incidents like this, DJI still haven't acknowledged a problem or done anything to fix this disaster.

My first DJI drone was a Phantom 2 in maybe 2013, followed by a Mavic Pro by end of 2016 with meanwhile more than 110h total air time at 573 flights and not a single incident of any kind. The Mavic Pro still flies on its first three batteries (Fly More package) and just the props have been replaced without need by the Platinum version.

When I bought the Mini I expected the same reliability as used to by the older drones but this in not the case here, it is a big step backwards. And when I remember to have seen the uncommanded descending already in an early promotor's video happening with a prerelease unit, DJI had already more than 6 months time to fix that serious issue.

So I must state from todays view, that the Mini is the most worse maintained drone DJI ever has produced and does not fulfill the quality expectations we all meanwhile have with DJI products. Such issues as we see here will damage the high reputation of DJI significantly. It's the first drone I would not recommend for buying, especially not for beginners!

Labroides Posted at 4-16 19:32
Your Mini flew normally until 2:26.1 when it started losing height.
It lost 35 ft before you reacted by giving it full throttle at 2:33.3 but the Mini lost another 20 ft despite full throttle being applied.
It appeared to recover briefly but at 2:44.4 it was losing height again, and again kept losing height despite full throttle being applied.

Thank you for taking the time to look at my flight logs. I have e-mailed DJI Support with a link to all my logs, the DAT file, videos from my drone and video I took with my phone where I located the Mini to show the conditions.  
Hopefully DJI support will also look at my logs and determine if I need to replace my drone through warranty. This is definitely not the reliability I would expect from DJI products.

I guess just replacing the drone by another one is not a real solution. All hardware/sensors seem to work properly as the barometer is reflecting the uncommanded changes in altitude. And if it is a firmware issue, all drones will suffer from this bug, more early or later.
I have done further test flights today, both in P-mode and sports mode. Was not able to provoce uncommanded descending today, but at least have seen uncommanded ascending again with both modes, as soon as approaching full speed.
Now I have recalibrated the IMU although I have few hope that this will fix the issue. But usually this is one recommendation from DJI support if telling them a problem. Will have another test flight tomorrow, where I will also use a different battery.

GerdS Posted at 4-17 13:40
I guess just replacing the drone by another one is not a real solution. All hardware/sensors seem to work properly as the barometer is reflecting the uncommanded changes in altitude. And if it is a firmware issue, all drones will suffer from this bug, more early or later.
I have done further test flights today, both in P-mode and sports mode. Was not able to provoce uncommanded descending today, but at least have seen uncommanded ascending again with both modes, as soon as approaching full speed.
Now I have recalibrated the IMU although I have few hope that this will fix the issue. But usually this is one recommendation from DJI support if telling them a problem. Will have another test flight tomorrow, where I will also use a different battery.

But usually this is one recommendation from DJI support if telling them a problem. Will have another test flight tomorrow, where I will also use a different battery.
DJI have no idea and are no help.
Their people might suggest recal of IMU just to say something and get rid of you.
If it was anything easy and obvious like IMU recal/firmware/battery it would have been solved long ago.
This has been happening all year and they still haven't been able to do anything to fix what is a very serious problem.

Labroides Posted at 4-17 16:05
But usually this is one recommendation from DJI support if telling them a problem. Will have another test flight tomorrow, where I will also use a different battery.
DJI have no idea and are no help.
Their people might suggest recal of IMU just to say something and get rid of you.

I did a recal of the IMU, compass, and refreshed the firmware with each battery installed. I noticed yesterday that upon take off it would drift backwards so I also replaced the rear props. Testing in my front yard it is holding position well although seems a bit shaky given how little of a breeze there is. No touch-and-go with the ground again so far. Will be heading out to a field to see if it holds altitude while flying forward.

tyler947 Posted at 4-17 16:11
I did a recal of the IMU, compass, and refreshed the firmware with each battery installed. I noticed yesterday that upon take off it would drift backwards so I also replaced the rear props. Testing in my front yard it is holding position well although seems a bit shaky given how little of a breeze there is. No touch-and-go with the ground again so far. Will be heading out to a field to see if it holds altitude while flying forward.

Well I flew two batteries in all the modes over snowy fields are varying altitudes and only noticed it climb a few meters while flying forward at full speed. Also when taking off it did not drift backwards after the prop replacement. I compared the replacement props to the ones I took off and can not tell any difference at all. I changed too many variables to be sure what helped so I do not know if it was the firmware refresh 3x (1 per battery), the IMU cal, or the prop swap but for now it is working as expected. The only warning message I had was due to wind at higher altitude - no max power / esc issues.

tyler947 Posted at 4-17 18:37
Well I flew two batteries in all the modes over snowy fields are varying altitudes and only noticed it climb a few meters while flying forward at full speed. Also when taking off it did not drift backwards after the prop replacement. I compared the replacement props to the ones I took off and can not tell any difference at all. I changed too many variables to be sure what helped so I do not know if it was the firmware refresh 3x (1 per battery), the IMU cal, or the prop swap but for now it is working as expected. The only warning message I had was due to wind at higher altitude - no max power / esc issues.

"I changed too many variables to be sure what helped so I do not know if it ..."


You could do a few more test flights as is to be sure problem is gone.

Then put swapped out props back on.  

Do a few test flights.

Problem returns, most likley the props.  Otherwise it was firmware refresh or IMU calibration.

I want to ask you to do a little test with yours to check if this problem is affecting all Mavic Minis more or less:

1. Go to a flat location with at least 200m free space in one direction.
2. Take off in P mode and go to a preferred altitude, keep the altitude as shown within the Fly app in mind. Don't touch the throttle stick anymore from now on.
3. Fly straight away approx. 200m with full speed, then slow down for hovering.
4. Read the altitude there, then wait about 10s and check for uncommanded altitude loss.
5. Turn 180° and fly back to your home point with full speed.
6. When back, check the altitude as displayed by the Fly app. Is it still the same as was at the beginning, or how far is it off?
7. Switch to Sports mode and repeat from step 2.

Please let me know your results.

For my MM difference in altitude was up to 10m in sports mode.

I repeated this test today, although quite gusty, with my old Mavic Pro, both in P and S mode. What shall I say? Altitude was rock solid and was exact same between start and end of test, if there was any derivation it was less than 1m. Same I would expect from Mavic Mini.

The Mavic FCs execute many control loops and one of them is responsible for continuously adjusting altitude. Something goes very wrong there in MM firmware, but for some MMs the problem does not show up that seriously, why ever...

I might now contact the DJI service for exchanging my MM and with some luck, the replacement drone might not show this bug so drastically as my current one. But on the long term this would be no solution as long as the firmware bug causing this issue remains unfixed. There is no guarantee that this bug will not show up again in future, caused by aging parts.

The problem is now finding a really competent contact person at DJI who is able to adress this problem to the responsible developer for finally fixing.
Any help or contact info is welcome.

GerdS Posted at 4-19 08:24
I want to ask you to do a little test with yours to check if this problem is affecting all Mavic Minis more or less:

1. Go to a flat location with at least 200m free space in one direction.

heres mine doing a drop to the ground after flying in sport mode, this was after only a few days flying