About two months ago, I took a flight with my brand-new Mavic 3 Pro. The drone had approximately 2 hours of flight time. I am an experienced pilot, having flown DJI drones since the F550, and I've owned all Phantom and Mavic models. I even flew near the North Pole where the compass wasn't functioning correctly. I own a boat charter company and fly from my boat about four times per week for my business, so I have a significant amount of experience in various flying conditions.

Here's some context and information about my experience.

I was on my powerboat (which has no mast) and took off as usual. Everything seemed fine with the drone, no obstacles in sight, and I was flying over the ocean.

However, when I reached my flying altitude of about 10 meters, I received an error message indicating a "Sensor error." The drone suddenly drifted rapidly to the right and plunged into the sea. It felt like it dropped like a stone for no apparent reason. It was a frustrating experience with my brand-new Mavic 3 Pro turning into a swimming lesson.

There was nothing I could have done to prevent it. I managed to retrieve the drone from underwater after a few minutes, but due to the saltwater exposure, both the drone and the battery were irreparably damaged.

Fortunately, the drone is still covered by DJI warranty.

Immediately after takeoff, the drone displayed the following errors (error codes: 16010010, 1601002F, 16010041), and then it plummeted into the water.

I subsequently sent the drone to the DJI Europe repair center, and after some initial difficulties, they finally received it and conducted an analysis.

Their conclusion was that I had collided with an obstacle. I vehemently disputed this, insisting that it was impossible and that I was not being dishonest. However, they later explained that the drone had been in ATTI mode, which raised more questions.

What baffles me is the inconsistency in the reports. Initially, the engineers claimed a collision with an obstacle, but now the explanation has shifted to the drone being in ATTI mode. I am seeking a clear explanation for why the drone suddenly fell from the sky, and currently, there isn't a definitive answer.

I was confident that DJI would replace my drone under the one-year warranty, but instead, they asked me to pay 600€ for repairs or to use my DJI Care and pay 150€.

I've provided evidence of the drone's sensor error and have found other individuals on forums who encountered similar situations—sensor errors followed by drone crashes (https://mavicpilots.com/threads/sensor-error-16010041-and-aircraft-crash-immediately-after-firmware-update.132602/).

I had no choice but to use my DJI Care Refresh to have a chance to fly my new drone again.

I am deeply disappointed with DJI's handling of this situation. Now, I lack confidence in flying this drone, as I don't know if it might suddenly fall from the sky at any moment.

Let me know if other experienced similar issues. Appently it can happen with all mavic 3 models.

Here is a screenshot of the error messages I got right before the crash

And here is DJI repair center explanations

Flying over water will cause the sensor error. That's your fault when flying over water. You take the risk of it crashing into the sea. Well at least they replace it with a little fee compare to buying new $2199.

You got it wet.  It responded.  Not a happy drone.  : )

Hi, there. We are sorry for the inconvenience caused. This is not the service we want to have. We are able to locate your case number using your forum account. We'll forward this to our relevant team for further assistance. Rest assured that this will be taken care of and handled accordingly. Thank you for your patience and understanding.

djiuser_zKR5HFX3a2EC Posted at 9-16 16:10
Flying over water will cause the sensor error. That's your fault when flying over water. You take the risk of it crashing into the sea. Well at least they replace it with a little fee compare to buying new $2199.

Flying over water will cause the sensor error. That's your fault when flying over water.
You have no idea what you are talking about.
Flying over water does not cause a "sensor error".
If it did, I'd know about it as I've flown thousands of miles over the sea without any errors.

but now the explanation has shifted to the drone being in ATTI mode. I am seeking a clear explanation for why the drone suddenly fell from the sky, and currently, there isn't a definitive answer.

I've never seen this error message and can't tell why it might cause the drone to fall.
Being in atti mode isn't a problem and definitely doesn't cause a drone to fall.
And you should ignore the person who suggested flying over water causes problems.
It sounds like you'd know that anyway.

It's hard to work out what actually happened from the words, but if you can post the recorded flight data someone may be able to interpret it and give you a second opinion about the incident.

Go to: https://www.phantomhelp.com/LogViewer/Upload/
... where you'll find instructions.
That will give a report you can post a link for the report here.
Or just post the .txt file.

@fansbb4da246 do you remember if you sent DJI any logs?
If so do you remember what logs you sent?

Do you still have the controller and or the phone used with the flight?
If so upload the .txt flight log from the screen device to https://www.phantomhelp.com/LogViewer/Upload/ and paste the URL of the resulting page here.
The txt probably has a name similar to DJIFlightRecord_2023-07-30_[15-2x-yz].txt

Could you check if you have  any .DAT flight logs on the screen device, if so do you have one named ''2023-07-30_15-26-23_FLY033.DAT''?
If so can you upload it to a file hosting website, make its page public and post the address here?

is recommended to upload your logs here as per Labroides. Dji has been known to reverse their decision when you go back to them with solid proof and can state your argument better. I fly over water a lot, long distance flights and have never encountered a similar issue before.

Hi there, we are sorry for the accident. Please allow me to explain and confirm further information with you.

The analysis result of your case is Flight Date: 2023-07-30 15:26:23 GMT+8    FLY-33

1. The aircraft worked under ATTI mode after taking off（GPS signal bar 0）
2. Flight Time T=48.02s, due to the unstable platform, the aircraft tilted forward for 14.78 degrees and tilted leftward for 58.3 degrees before taking off, the aircraft rolled over when it was adjusting its attitude after taking off, and then it crashed with the obstacle

Conclusion: The incident was caused by non-manufacturing factors so it is not covered under warranty.

Could you please double-check if the time and date are correct? Besides. if you have the recording video of the accident, please also provide it to us, we will ask our engineers to check it further.

fansbb4da246 Posted at 9-16 16:00
And here is DJI repair center explanations

[view_image]

Hi,

Interesting chart from DJI.

It does not show height of your drone, unfortunately.
What it does show is that at takeoff (drone in air) already the pitch and roll angles are not normal values belonging to a normal takeoff.

If you share your flightlog, the height values can be seen.

I doubt if 10 meters was ever reached, but height data in the log will tell.

BTW sensors does not go into an error flyign over water! I fly a lot (low) over water, never had any sensor errors doing so.
Sensors are not so clever to understand if drone is flying over water or not.Only their measurement or return signal interact with software.
Flying over highly reflective underground is noticed by sensors, but what it actually is ? ; still water, ice, snow, solar panels...

Taking off in ATTI mode is possible and will not result in loosing height, just horizontal stability not there (if vision is not working that is)

cheers
JJB

Hello everyone,

Thank you all for your feedback. Here is the Flight Log:

https://www.phantomhelp.com/LogViewer/0YKIYONPW69TXE8FI3W7

From what I can gather from the log, the drone did not acquire satellites when I initially took off, and it was flying in ATTI mode. However, at 4.8 seconds into the flight, as I was ascending, it suddenly switched to P-GPS Brake and veered to the right without any input from the controls. It then quickly obtained 5 satellites. You can observe that all the control sticks were in neutral positions while the drone drifted to the right and lost altitude until it eventually crashed into the ocean. Additionally, the IMU altitude was negative at the start of the flight, though I'm unsure if this could be a contributing factor. Nevertheless, the log clearly indicates that the drone experienced an uncontrolled flight, and I couldn't regain control over it.

As I mentioned to DJI, I did not collide with any obstacles; I was in the middle of the ocean and witnessed the drone's descent. I want to emphasize that I am being completely honest in this matter. I had DJI Care coverage, and if the incident had been my fault, I would have willingly paid the 150€ . However, considering the additional 60€ for shipping to return the drone, this amounts to a substantial cost when I am not at fault.

Here is the response I received from DJI engineers:


Hi there, we are sorry for the accident. Please allow me to explain and confirm further information with you.

The analysis result of your case is Flight Date: 2023-07-30 15:26:23 GMT+8    FLY-33

1. The aircraft worked under ATTI mode after taking off（GPS signal bar 0）
2. Flight Time T=48.02s, due to the unstable platform, the aircraft tilted forward for 14.78 degrees and tilted leftward for 58.3 degrees before taking off, the aircraft rolled over when it was adjusting its attitude after taking off, and then it crashed with the obstacle

Conclusion: The incident was caused by non-manufacturing factors so it is not covered under warranty.




Ultimately, I can only identify two possible causes:

1 - The drone's IMU did failed due to inadequate calibration on the boat.
2 - The drone touched an invisible UAP


At the end I got my replacement drone.  What I am only expecting from DJI is to admit that I am not a Liar, and that the drone got a malfunction, and so refund me my used DJI care. That is all I am asking for.

Please feel free to share any insights or recommendations regarding this situation.
Thanks

First thing I'd do is question how the drone could have taken off at negative height, i.e. -27ft?
The motor start or perhaps the launch is supposed to set the height to zero, that's the entire basis of the drone's barometric height-measurement regime.
The only possible thing I can think of is that this is perhaps the second 'flight' in ONE drone switch on, where the first flight was VERY short, 99% battery at the '2nd' restart, and your boat was moving a lot, the movement perhaps confusing the drone.

I will have to check an old pair of logs to see if, in those, the 'height' carried over from one flight to the next ( mid air motor stop and restart with a P3 ) but even if it happens, your negative height suggests that the 'second take off' was 27ft below the first, that seems impossible to me on a "motor boat".

Sean-bumble-bee Posted at 9-18 08:05
First thing I'd do is question how the drone could have taken off at negative height, i.e. -27ft?
The motor start or perhaps the launch is supposed to set the height to zero, that's the entire basis of the drone's height-measurement regime.
The only possible thing I can think of is that the this is perhaps the second 'flight' in ONE drone switch on, where the first flight was VERY short, 99% battery at the '2nd' restart, and your boat was moving a lot (the movement perhaps confusing the drone.

YEP. Checked the logs, the original launch height did carry over to the 'motors restarted' flight log i.e. all heights shown in the second log were relative to the original take off point. So......with regards to your flight/s was this the second flight of a single drone switch on, with the drone being moved about quite a lot between the two 'flights'?

Second question, there are two periods of ATTI indicated, the first from 2.3 seconds to 4.7 seconds, during this period the VPS height increases from 0ft to 34.4ft so the VPS was seeing something, the question is, what was it seeing? I guess the second period could be due to either the fall or the drone being too high, I haven't checked the range/height limit for VPS sensing.

ATTI occurs when the drone has insufficient GPS AND no VPS.
No VPS can be due to a number of reasons, poor light, inappropriate surface/surface-pattern, too much height, the latter didn't apply to the first ATTI. You would have to tell us about the former two. If the surface was appropriate and well lit then the drone should not have been in ATTI mode.
Was the boat moving horizontally, i.e. being driven? What way did the drone move relative to the boat and what way was the drone facing relative to the boat?

That I can see you gave the drone ONLY throttle and rudder, so, question 3, what caused the roll shown in the log? ATTI mode would mean that WITHOUT aileron and elevator the drone should have ZERO pitch and roll. It should not attempt to hold position so there's no reason for it to roll .............. or am I completely up a gumtree???????

Caribbean Drone Posted at 9-18 07:45
Hello everyone,

Thank you all for your feedback. Here is the Flight Log:

Hi,

Had a look at your flight too, and yes   your drone reached some height. (13.3 meter IR height measured)

But
First records shows IMU error, so not wise thing to start a flight wich such error.

Guess due to this IMU error the baro height, wich is normally reset to zero at starting motors, did not reset the baro height to zero.
I never have seen that height from previous flight is there, at clsoing down the rotors the baro height is also set to zero.

Error in the log "vision sensor blocked"  So with zero satellites and no vision  : ATTI mode.

But at 4.8s FlightMode shows P-GPS !  must be fault in DJI software or dycrypting software, with no sats and no vision only mode possible is ATTI.

At the end Sensor Fault (not describing wich sensor!). Drone looses height rapidly.

My conclusion :
Pilot error and Drone IMU error.
Pilot: not clever to start a flight with IMU errors at start.
IMU error: well  i have no idea why your drone had a  IMU error. (but as DJI replaced the drone....)

Interesting to read that DJI saw roll angle of 58.3 degrees before takeoff, in the log (see screenshot) only normal values at start / takeoff.
Max roll right angle : 45.2 degrees at 6 sec in flight.

cheers
JJB  [ Charts by FRAP ; the 'best' software for data and visual flightlog analysis (www.jjbfrap.eu) ]

Sean-bumble-bee Posted at 9-18 08:50
YEP. Checked the logs, the original launch height did carry over to the 'motors restarted' flight log i.e. all heights shown in the second log were relative to the original take off point. So......with regards to your flight/s was this the second flight of a single drone switch on, with the drone being moved about quite a lot between the two 'flights'?

Second question, there are two periods of ATTI indicated, the first from 2.3 seconds to 4.7 seconds, during this period the VPS height increases from 0ft to 34.4ft so the VPS was seeing something, the question is, what was it seeing? I guess the second period could be due to either the fall or the drone being too high, I haven't checked the range/height limit for VPS sensing.

Hi

Good question

That I can see you gave the drone ONLY throttle and rudder, so, question 3, what caused the roll shown in the log? ATTI mode would mean that WITHOUT aileron and elevator the drone should have ZERO pitch and roll. It should not attempt to hold position so there's no reason for it to roll .............. or am I completely up a gumtree???????

Looking only at data, at takeoff in ATTI only RC up command. Drone Pitch and Roll both almost flat values. Normal for take in ATTI ; with only UP and no software correction to position are made.


But at 4.7s in the log P-GPS, uh ?  not possible ofcourse! (zero sat, vision blocked). So why in the log ?  for the logic software 1 or 0, so drone is in "P-GPS mode"
GPS mismactch position errors. Seen before in logs, this will make steeer input to the drone.
Speed increases rapidly.


In one of the last records : craft flying in zone....what kind of zone ?  Flying in a No fly zone in ATTI mode = descending.


cheers
JJB

JJB* Posted at 9-18 10:28
Hi

Good question

I think there are all sorts of questions raised by this flight/incident.

BTW the launch height DID carry over between the motors stopped and motors restarted portions of the P3 flight I mentioned I am looking at the 2nd log now.
I also think I have seen height carry over from one log to the next where a disconnection has caused two separate .txt flight logs to be created. That two logs where created is odd enough in itself, I will see if I can find them .... just to check.

Sean-bumble-bee Posted at 9-18 10:53
I think there are all sorts of questions raised by this flight/incident.

BTW the launch height DID carry over between the motors stopped and motors restarted portions of the P3 flight I mentioned I am looking at the 2nd log now.

True, 2 flightlogs created with a disconnect in the middle, baro height is not set to zero as motors keep running. with rotors start or stopped, baro set 2 zero.

If log is started after earlier start that the time in sec is not started at zero, so afaik this is the first log for this flight.

JJB* Posted at 9-18 11:30
True, 2 flightlogs created with a disconnect in the middle, baro height is not set to zero as motors keep running. with rotors start or stopped, baro set 2 zero.

If log is started after earlier start that the time in sec is not started at zero, so afaik this is the first log for this flight.

"If log is started after earlier start that the time in sec is not started at zero, so afaik this is the first log for this flight."

If I understand you correctly, good catch, in my restart log "OSD.flyTime" starts at several minutes.

Caribbean Drone Posted at 9-18 07:45
Hello everyone,

Thank you all for your feedback. Here is the Flight Log:

Hello! Thank you for getting back to us, and we sincerely apologize for any inconvenience this has caused. We will forward your request to the relevant team for review. Please keep an eye on your email for any updates.

JJB* Posted at 9-18 09:41
Hi,

Had a look at your flight too, and yes   your drone reached some height. (13.3 meter IR height measured)

Hi,

I have a couple questions for you.

1. If the drone is in ATTI mode prior to launch, should the DJI software detect that and prevent launch?
  Same thing with GPS quality 0 or 1 or 2.  Should the DJI software detect that the drone has GPS quality of  0 or 1 or 2 and prevent launch?

2.  Or should  DJI instead flash a warning on screen that launching with GPS quality 0 or 1 or 2 may void warranty for flight..proceed at own risk…launch not recommended…and pilot must agree before launching will be unlocked?
3.  Should DJI be more transparent about some of the things that can cause GPS weak signal...e.g. unintentional Multipath or intentional Multipath jamming?

JJB* Posted at 9-18 09:41
Hi,

Had a look at your flight too, and yes   your drone reached some height. (13.3 meter IR height measured)

In my defense, I never received any IMU error messages. I'm not foolish; I wouldn't fly with an uncalibrated IMU over water. When I powered up the motors, everything on my DJI Remote Controller showed green. However, just a few seconds after takeoff, the drone experienced the failure and plummeted. Nothing I could have done

DroneApe Posted at 9-18 15:44
Hi,

I have a couple questions for you.

Regarding 1) & 2)
I can't answer with respect to the Mavic 3 but with the Mini 1 & 2 if there is no GPS and poor lighting, which amounts to no VPS and the drone taking off in ATTI mode,  they do prevent 'uninformed' take off, you have to touch the warning and then acknowledge/accept responsibility for the flight ... or some such acceptance. I get that virtually every time I launch them indoors, so much so that I no longer read the 'acceptance blurb'. In my case fortunately once they get airborne they 'realise' that there is enough light for the VPS to work and fly perfectly - I think the drone's body puts the VPS camera in deep shadow when it is sitting on the table top etc,,The same thing might happen with the M2P/Z but I can't remember but the M2P/Z does frequently switch its landing light on just after take off, presumably to assist the VPS cameras.
Have a look at https://forum.dji.com/forum.php? ... 1&extra=#pid3132878

It wouldn't surprise me if the Mavic 3 Pro does the same thing.

Sean-bumble-bee Posted at 9-18 18:19
Regarding 1) & 2)
I can't answer with respect to the Mavic 3 but with the Mini 1 & 2 if there is no GPS and poor lighting, which amounts to no VPS and the drone taking off in ATTI mode,  they do prevent 'uninformed' take off, you have to touch the warning and then acknowledge/accept responsibility for the flight ... or some such acceptance. I get that virtually every time I launch them indoors, so much so that I no longer read the 'acceptance blurb'. In my case fortunately once they get airborne they 'realise' that there is enough light for the VPS to work and fly perfectly - I think the drone's body puts the VPS camera in deep shadow when it is sitting on the table top etc,,The same thing might happen with the M2P/Z but I can't remember but the M2P/Z does frequently switch its landing light on just after take off, presumably to assist the VPS cameras.
Have a look at https://forum.dji.com/forum.php? ... 1&extra=#pid3132878

I should have wrote  GPS quality 1 or 2 instead of GPS quality 0 ATTI so as not to confuse.  

DroneApe Posted at 9-18 15:44
Hi,

I have a couple questions for you.

Hi

If the drone is in ATTI mode prior to launch, should the DJI software detect that and prevent launch?
ATTI is detected, audio warning + prompt on the screen. BUT message on the screen is only seen for few seconds, easy tro miss.IMO DJI should change the flightmode (top left FlyApp) into ATTI, bc N-mode is seen while drone is in ATTI mode ! ( bad UI )
You can fly in ATTI, so no need to prevent a launch. BUT mayby a confirmation prompt for the pilot should be nice "sure wanna takeoff in ATTI mode ?"

Same thing with GPS quality 1 or 2.  Should the DJI software detect that the drone has GPS quality of  0 or 1 or 2 and prevent launch?
Top left GPS icon shows 3 colours ; just wait for WHITE.  [ white does not say anything about receiving sat numbers, it is number+quality = white ]
Red = not enough sats for GPS lock [ in the csv 0,1 or 2 ]
Yellow = enough sats for GPS lock, not enough to set HomePoint [ in the csv 3 ]
White = enough sats for GPS lock,  HomePoint has been set. [ in the csc 4 or 5 ]

Or should  DJI instead flash a warning on screen that launching with GPS quality 0 or 1 or 2 may void warranty for flight..proceed at own risk…launch not recommended…and pilot must agree before launching will be unlocked?
Yes, would be nice for non experienced pilots, see my first remark. (mayby as an option disable/enable extra takeoff warnings in the safety menu)

Should DJI be more transparent about some of the things that can cause GPS weak signal...e.g. unintentional Multipath or intentional Multipath jamming?

Nah, enough on the www to find. Problems is, imo, that many drone pilots do not read and understand the manual, and have no idea that drone flying is not that simple.
(flightplanning, battery management in flight, wind calcualtions, risky envirioments, etc etc)


cheers
JJB

Caribbean Drone Posted at 9-18 15:56
In my defense, I never received any IMU error messages. I'm not foolish; I wouldn't fly with an uncalibrated IMU over water. When I powered up the motors, everything on my DJI Remote Controller showed green. However, just a few seconds after takeoff, the drone experienced the failure and plummeted. Nothing I could have done

Hi,

IMU and or ATTI warnings are prompts on the screen, but dissapear after short period.
As we all look at our drone at take-off   easy to miss disapearing prompts on the screen.

Same for the flightmode seen in the FlyApp, it shows N-Mode while drone is in ATTI....DJI should change this asap.

cheers
JJB

Sean-bumble-bee Posted at 9-18 11:50
"If log is started after earlier start that the time in sec is not started at zero, so afaik this is the first log for this flight."

If I understand you correctly, good catch, in my restart log "OSD.flyTime" starts at several minutes.

Good catch ....

Too many checks to remember, reading a CSV.

My memory is nicely saved in FRAP.   ;-)

cheers
JJB   [ Charts by FRAP ; the 'best' software for data and visual flightlog analysis (www.jjbfrap.eu) ]

DroneApe Posted at 9-18 20:44
I should have wrote  GPS quality 1 or 2 instead of GPS quality 0 ATTI so as not to confuse.

And I should probably written "poor GPS" instead of "no GPS". I will have to have a look at the OSD.gpsLevel column in some of my indoor flight logs, to see what its value was at motor start i.e. just after the "acceptance".

JJB* Posted at 9-18 23:41
Good catch ....

Too many checks to remember, reading a CSV.

That options list makes interesting reading, cheers.

djiuser_zKR5HFX3a2EC Posted at 2023-9-16 16:10
Flying over water will cause the sensor error. That's your fault when flying over water. You take the risk of it crashing into the sea. Well at least they replace it with a little fee compare to buying new $2199.

Huh.

djiuser_zKR5HFX3a2EC Posted at 2023-9-16 16:10
Flying over water will cause the sensor error. That's your fault when flying over water. You take the risk of it crashing into the sea. Well at least they replace it with a little fee compare to buying new $2199.

Flying over water will cause sensor errors... What a BS man. Diving into salt water will cause sensor errors,
I have used my sensors many times above water, never had any error.

Poor light / visibility, IMU, compass or system failure are more likely the cause of this accident. Take off without proper booting the system could also be a reason for this crash (drone in ATTI).