The second time ever flying my Mavic Air, it went off on its own uncontrolled.
I had the drone hovering directly in front of me about 8 ft off of the ground, when it suddenly took off straight up without any controller input. Nothing I could do would stop the drone from ascending on its own. Including down on the joystick, the return to home function, and the auto-land function. It completely ignored any of these commands and continued ascending on its own, until it finally stopped at the max height setting of 400ft.

For the next 15 minutes I tried everything possible to get the drone to come down. Repeatedly using the RTH function, auto-land, joysticks, disconnecting my phone and using just the controller on its own. Nothing could get it to descend from the max height. I still had forward/back, side to side, and gimbal control, but no vertical control. The drone remained pinned at 400ft for the next 15 minutes until eventually the battery died and it plummeted to the ground.

The flight log shows the moment the drone suddenly interprets that it's getting a full up input from the vertical control joystick. The entire rest of the log it shows it's still somehow getting this signal, even though I'm giving no input on the controller, or even when I'm inputting full down in repeated attempts to get the drone to descend. The log also shows it's receiving my requests to both return to home, and to auto-land. It goes into these modes, but continues ascending uncontrolled, apparently being overidden by the phantom joystick signal it's receiving.

I've provided the CSV file highlighting exactly the moment the malfunction starts, notating every attempt I make to RTH and autoland, and showing the clear evidence it is ignoring all of my commands. I've also provided the video from the drone itself, in which you can see I am giving no input on the joysticks when it suddenly starts ascending uncontrolled.

The data log and video cut out early before the eventual crash, because the drone remained in the air stuck at 400ft until the battery died. The DJI customer service keeps repeating that since the log doesn't show the actual crash where it hits the ground, it cannot be covered under warranty. I've explained repeatedly the circumstances, which resulted in the drone being stuck at max height until the battery died, which is why you can't see the actual crash in the log.

I've repeatedly sent them the highlighted portions of the CSV file showing the drone ignoring all commands to RTH or land, and asked for any sort of explaination on how this doesn't constitute a malfuction. I've asked 6 times now and it has been ignored every time. I'm documenting each time I've shown them the evidence and they're refused to acknowledge it.

I have DJI Care Refresh on the product, but I refuse to use one of my repairs when there was such a clear malfuction. Even worse, I'm sure they would just send me the drone back since according to them there is nothing wrong with it. There is no way in hell I'm accepting back a drone that malfuctioned to the point that it fell from 400ft almot striking myself and my dog.

I'm disgusted by DJI's refusal to stand behind their product. I spent nearly $1000 on a drone that I used twice, and now has been back with DJI for over 2 months while they do everything in their power to avoid taking responsibility for thier faulty product.

Here you can see where the drone clearly received the commands for the Return to Home as well as Auto-Land functions, but completely ignores them and continues ascending all on its own.

You stand a better chance of getting useful comments if you post the .txt file so people can look at the actual flight data.
Also post DJI's analysis.

Without being able to see the data, the incident as described sounds like a yaw error situation which is usually caused by magnetic distortion where you chose to power up and launch the drone.
Can you describe your launch surface?
There may be other factors too that would only be known from seeing the data.

Good point. Give more data.

Sorry about that, no problem.

At 1m 59s in the flight log is when it suddenly thinks there is a full up signal on the right stick (vertical control for mode 3) The remains pinned that way for the remainder of the flight, I am not giving any input on the stick, and tried repeatedly pushing it down to get it to descend.

You can also see my attempts to get it to descend using Return to Home and auto land functions at 2m 38s, 2m5 0s, 5m 20.6s, and 5m 37s. I tried these functions and everything else possible many more times after the data log cuts out. It shows the drone gets these signals, but it completely ignores it and continues it's uncontrolled ascent.

CSV File -
https://drive.google.com/open?id=1Y_5aSQW9DCLyDUGiF5oOzoSGG8Nn7yAR

Hi joshfishtail, we are regretful to know the data analysis is inconclusive in your case because no crash data found. I managed to get your case number via your Forum account, your concerns will be escalated to the higher-level team for further follow-up, our manager will contact you after double confirmation, please sit tight, thank you.

DJI Mindy Posted at 4-15 19:10
Hi joshfishtail, we are regretful to know the data analysis is inconclusive in your case because no crash data found. I managed to get your case number via your Forum account, your concerns will be escalated to the higher-level team for further follow-up, our manager will contact you after double confirmation, please sit tight, thank you.

Thank you, Mindy. I really appreciate it.

I have a question, Have you synced your flight distance on the Go4 app? you do not show any flown distance on your Avatar, are you using the Go4 app or a third party app?

joshfishtail Posted at 4-15 19:01
Sorry about that, no problem.

At 1m 59s in the flight log is when it suddenly thinks there is a full up signal on the right stick (vertical control for mode 3) The remains pinned that way for the remainder of the flight, I am not giving any input on the stick, and tried repeatedly pushing it down to get it to descend.

Your flight data shows full throttle input from 1:58.7 until the end of the flight data with the drone stuck at the Max Altitude limit of 400 ft.
Are you sure the problem was in the drone and not the controller?

The evidence in the flight data is obvious and DJI's people should be able to see it clearly.

You haven't posted their analysis to compare with the flight data.
It would be interesting to see if they are even looking at the right flight.

joshfishtail Posted at 4-15 19:01
Sorry about that, no problem.

At 1m 59s in the flight log is when it suddenly thinks there is a full up signal on the right stick (vertical control for mode 3) The remains pinned that way for the remainder of the flight, I am not giving any input on the stick, and tried repeatedly pushing it down to get it to descend.

Hi Josh,

Is it possible to put a link to the DAT (FLY026.DAT)  file for this flight? (its on you mobile device as well)
Mayby in this DAT log is an error written, looks like indeed your up signal was all the time 100% up.
And this up command overrules the autolanding descent.

If the controller is in error (sending all the time 100% up) than in retrospect guess to switch of the controller could have saved your drone. But at the moment of this &#$&!#$, i would not think of that myself!
Cannot imaging that DJI will not help you out, its obvious you tried to land the craft flying so close to your hp!

cheers
JJB

Labroides Posted at 4-15 22:02
Your flight data shows full throttle input from 1:58.7 until the end of the flight data with the drone stuck at the Max Altitude limit of 400 ft.
Are you sure the problem was in the drone and not the controller?

I've repeatedly shown them the exact flight log file, and they asked for the specific date. I also included the file in a drop box at their request.

I've walked them through with an explanation of exactly what was happening at each time segment. I kept pressing and I have been sent through 3 different representatives now, each who just states the crash is not on the file so it is not honored under warranty.

They haven't provided me any sort of analysis that I can post. Neither from the initial data analysis team, nor the flight engineer they said they sent the file to after I started pushing back.

They just keep repeating there is no crash on the log, and I keep repeating the reason (the malfunction caused it to be stuck at max height until battery died), and telling them the exact times in the CSV file where they can see the malfunction taking place that lead to the crash. I've even gone to the extent of highlighting the portions of the CSV file myself and sending them the notated pictures, but they have completely ignored it every time.

It very well could be an issue with the controller, you are right. It's astounding to me I'm having to go on a forum to get any sort of real answers because DJI is refusing to look into it.

JJB* Posted at 4-15 23:40
Hi Josh,

Is it possible to put a link to the DAT (FLY026.DAT)  file for this flight? (its on you mobile device as well)

Thanks so much for the help, I've added the DAT file in to the same drop box found below.

Their response when I stated that it seemed to be getting a 100% up signal the entire time once the malfunction started, was that since I was not on the default controller setting and was on mode 3 instead, that must have caused the confusion. I responded that no, I am certain I was not holding 100% up on my right joystick for 15 minutes straight wondering why my drone wouldn't descend, haha.

The thought to turn the controller off and on again actually did cross my mind, but I wasn't sure if it could reconnect with the drone again when it was 400ft up. I wanted to keep my horizontal control so I could be sure to have it directly above the clearing in the park when it eventually died. I was was horrified watching the battery % slowly approach 0 thinking it was going to nail someone jogging in the park when it fell.

DAT File Added -
https://drive.google.com/open?id=1Y_5aSQW9DCLyDUGiF5oOzoSGG8Nn7yAR

joshfishtail Posted at 4-16 08:09
I've repeatedly shown them the exact flight log file, and they asked for the specific date. I also included the file in a drop box at their request.

I've walked them through with an explanation of exactly what was happening at each time segment. I kept pressing and I have been sent through 3 different representatives now, each who just states the crash is not on the file so it is not honored under warranty.

I think at first they will not give out there results, but if you persist they will , hopefully it might give a clearer picture.

joshfishtail Posted at 4-16 08:19
Thanks so much for the help, I've added the DAT file in to the same drop box found below.

Their response when I stated that it seemed to be getting a 100% up signal the entire time once the malfunction started, was that since I was not on the default controller setting and was on mode 3 instead, that must have caused the confusion. I responded that no, I am certain I was not holding 100% up on my right joystick for 15 minutes straight wondering why my drone wouldn't descend, haha.

Hi,

Its not the correct DAT file, file looking for is the FLY026.DAT, not the 15.

cheers
JJB

JJB* Posted at 4-16 10:44
Hi,

Its not the correct DAT file, file looking for is the FLY026.DAT, not the 15.

Do you know what folder I would go into to find that file? I'm only showing the one FLY015 file in my flight records. Nothing in Flightlogs, Flightrecords, or Logs folders.

joshfishtail Posted at 4-16 11:24
Do you know what folder I would go into to find that file? I'm only showing the one FLY015 file in my flight records. Nothing in Flightlogs, Flightrecords, or Logs folders.

Their updated response today.

Just repeating what they've said that the analysis is complete and I have to pay to move forward. The question he's referring to, which I've asked repeatedly - "It's clear from the CSV file the drone was stuck on 100% ascend, I've highlighted the evidence in the file. Explain to me how this possibly not a malfunction, or alternatively put it in writing to me that this is the normal function of your product to ascend uncontrolled and become stuck at max height."

They ignore it every time and just come back with a generic response without addressing the evidence I'm clearly highlighting to them. They won't mention it.

Sorry about your lost...
If you send them the log FLY015.DAT , they will assume what I have assumed after reading it, you where with a really poor connection (17.1 signal score on a scale of 100) and misregarding that you decided to drove yor bird till over 32.44Mph and you where 80Ft away with only 7% of battery left... NOT TO MUCH TO SEE ON THIS FILE...

What I'm trying to say is that if you have send this file to DJI it will be on your best interest to send the correct one ( DJIFlightRecord_2020-03-12_[16-50-12].txt ) and MAKE shure that they receive it and that they know what's about!

Sorry about my poor English, I'm just a simple Portuguese ;)

Also quite strange that file... ( FLY015.DAT ) it looks like he was recorded 8 minutes after the crash... Strange!

PauloACSilva Posted at 4-19 09:34
Also quite strange that file... ( FLY015.DAT ) it looks like he was recorded 8 minutes after the crash... Strange!

Yeah the more I dig into the files the stranger it is. The video that I had recording as well as the flight log cut out where there's still 30-something percent of battery left, and the FLY026 file is nowhere to be found.

To me this is just further evidence the drone wasn't operating correctly, the way that it seems to have messed up with the logs. From the time it at 15% battery and below I was just trying my best to position it directly above me to ensure it would land in the clearing I was in. Just tiny inputs to keep it directly above me. I am 100% certain I wasn't 80 horizontal ft away and flying at 30mph with 7% battery left. I was nervously staring straight up in the air at it waiting for the inevitable.

joshfishtail Posted at 4-21 11:56
Yeah the more I dig into the files the stranger it is. The video that I had recording as well as the flight log cut out where there's still 30-something percent of battery left, and the FLY026 file is nowhere to be found.

To me this is just further evidence the drone wasn't operating correctly, the way that it seems to have messed up with the logs. From the time it at 15% battery and below I was just trying my best to position it directly above me to ensure it would land in the clearing I was in. Just tiny inputs to keep it directly above me. I am 100% certain I wasn't 80 horizontal ft away and flying at 30mph with 7% battery left. I was nervously staring straight up in the air at it waiting for the inevitable.

I truly believe you, but that's the reading of that file...I have already found logs of my MM flying on places and directions that I never flew , one of the logs send my bird straight to a Island in the Atlantic... Now imagine if I was with the bird on the air and he took it as RTH point??? Just to tell you that on that day I flew only inside my apartment !!!!!!!!  Strange things have happened with almost all birds from DJI and for me most of the problems are mostly due to FW than Hardware, why DJI don't release SDK ? It would allow third parties to improve significantly the software for the MM and DJI then could learn and apply future updates that are much reliable...but that's just me alucinating!!!

PauloACSilva Posted at 4-22 13:24
I truly believe you, but that's the reading of that file...I have already found logs of my MM flying on places and directions that I never flew , one of the logs send my bird straight to a Island in the Atlantic... Now imagine if I was with the bird on the air and he took it as RTH point??? Just to tell you that on that day I flew only inside my apartment !!!!!!!!  Strange things have happened with almost all birds from DJI and for me most of the problems are mostly due to FW than Hardware, why DJI don't release SDK ? It would allow third parties to improve significantly the software for the MM and DJI then could learn and apply future updates that are much reliable...but that's just me alucinating!!!

That point in the Atlantic is famous, as it represents the GPS co-ordinates of 00.0 and 00.0, in other wods, no input from the GPS at all. And it is not surprising that you got that when flying indoors.

And the reason that the last log appears to be missing is probably because the battery shut off before the last log could be saved and/or finalised to the aircraft storage, again, this is quite normal after a crash.

Geebax Posted at 4-22 13:44
That point in the Atlantic is famous, as it represents the GPS co-ordinates of 00.0 and 00.0, in other wods, no input from the GPS at all. And it is not surprising that you got that when flying indoors.

And the reason that the last log appears to be missing is probably because the battery shut off before the last log could be saved and/or finalised to the aircraft storage, again, this is quite normal after a crash.

I'm going to open a tread on this...


Great to know that is "normal" when indoors ;)
On my last flight ( 23/04 ) and after the last update ( V01.00.0500 ), I received a message to repositioning my Antennas, but the bird was really close, analising the data, I can see a lot of losses in the radio signal.
That happened in two different locations , the first one in the left hand side of a bridge and the other on the right hand side... It's a concrete bridge...
Another IMPORTANT and very DISTURBING thing is that in BOTH flights in the .DAT file I can see all the up's and downs of the radio signal but on the .Txt files radio is always 100% even when I received a message to repositioning my antenna on log 063.txt ( weak signal Code: 80016) quite strange that difference between the logs.

PLEASE see how different both logs are related to “radio” signal…  
PLEASE use the HD flight player to better see the diferences…

1 - Any idea about that huge lack of radio signal whit the bird flying so close ( even at 1mt away!!! ) ???
2 - In which Log can whe relie on ???
3 - Is this a major issue with this NEW FW UPDATE ???

Yes, I updated Both MM and controller, and I recalibrated it...

Hope to get some answears for this problem...

Thank's in advance

Log TxT 063 :
https://app.airdata.com/flight/2 ... ENERALNotifications

Log TxT 064 :
https://app.airdata.com/flight/2 ... ENERALNotifications

Log DAT 063 :
https://app.airdata.com/flight/e ... ENERALNotifications

Log DAT 064 :
https://app.airdata.com/flight/7 ... ENERALNotifications

Im trapped in the same hell. Time for small claims court? Not honouring warranty is considered breach of contract in Canada.

HughNorth Posted at 10-31 07:25
Im trapped in the same hell. Time for small claims court? Not honouring warranty is considered breach of contract in Canada.

Any good lawyer would tell you that having them just read the terms and conditions of the warranty to seek breach would cost more than just buying ten new drones.  The courts are not the answer to these types of issues.