Hello,
just wanted to get some information about the situation which i had yesterday. Been flying over water for almost 2min with a good signnal when sudenlly i lost my connection with the aircraft and could not reconnect with it. I press RTH right away but unfortunetly my drone didnt came back. After waching the flight log and the last video image on find my drone, i see that the drone started to slowly go down towards water, then for a split of a second the view is disable and after few moments video signal is back but i see my drone falling fast from much higher high then I was flying it first place. Could someone take a look and give me an advice what was wrong in this situation.
Flight log: https://we.tl/t-lVHZznOmtZ
Flight from find my drone: https://we.tl/t-KMuKKCQupB
Last video from find my drone: https://we.tl/t-P0qeJxnSKD

Kind regards,
Simonas

Hi there. Thank you for reaching out and we're deeply sorry for the incident that happened to your unit. In this case, we strongly recommend you to contact our team at https://www.dji.com/support to open up a flyaway case. They will provide you with further details about this case once it is open. Thank you for your valued support.

Here's a summary of what your flight data shows:
https://www.phantomhelp.com/logviewer/6KJ2KO7II6SMJDR943G0/
The recorded data suggests that the drone was less than 1 metre above the water when you lost signal for 23 seconds.
After signal was restored, the height sensor was giving an impossible false reading of -411 metres

It looks like you flew the drone too close to the water while in Sport Mode and crashed it.

.

Labroides Posted at 11-21 02:51
Here's a summary of what your flight data shows:
https://www.phantomhelp.com/logviewer/6KJ2KO7II6SMJDR943G0/

in the second link which i attached there a video of fly path and sticks movement, at that moment when the signal was lost it doesnt show left stick pointing down, what is more the height at this point was something arround 3 meters above water. In the video which is in last link the drone dives down from something like 15-20m i would say.

after few moments video signal is back but i see my drone falling fast from much higher high then I was flying it first place.
What height do you think you saw the drone falling from?
The drone was >1200 metres away when it was lost.
Can you see it at that distance?

Labroides Posted at 11-21 03:02
after few moments video signal is back but i see my drone falling fast from much higher high then I was flying it first place.
What height do you think you saw the drone falling from?
The drone was >1200 metres away when it was lost.

No, I didnt see it that far. But what can i see is a last video captured of it and i clearly didnt make it to fly that high from where it was falling

alisauskas.simonas@gmail.com Posted at 11-21 03:00
in the second link which i attached there a video of fly path and sticks movement, at that moment when the signal was lost it doesnt show left stick pointing down, what is more the height at this point was something arround 3 meters above water. In the video which is in last link the drone dives down from something like 15-20m i would say.

in the second link which i attached there a video of fly path and sticks movement
The joystick input is visible in the recorded flight data too.

at that moment when the signal was lost it doesnt show left stick pointing down, what is more the height at this point was something arround 3 meters above water.
At 2:31.0 the drone was flying at full speed in Sport Mode (11.6 metres/sec) and the VPS sensor indicated the drone was only 0.3 metres above the water.
The height you see on your screen comes from a barometric sensor and is not accurate.

In the video which is in last link the drone dives down from something like 15-20m i would say.
Your estimate is very wrong.
The wideangle lens of the drone's camera makes things look further away than they actually are.
At 6 seconds into the video, the drone is just about touching the water.
When the video resumes, the drone has skimmed the water surface, become very wet, started to climb to RTH and fallen due to water in the drone and/or battery.

Your video was slow to download, but now that I've seen it, it confirms what the data shows.

Hi Simonas,

had a look at your log and video.

Flying real low over water....
at 2m30.9s in flight (flightmode Sport) you lost connection.
Last VPS height just before disconnect 0.3 meter...before that the VPS height values are incorret. (in red *!*)
Battery level 40%

at 2m54.3s connected again, drone in GoHome mode.
Battery level at connection 37%, so 3% used for....

What happend imo in the 23 seconds...at disconnect drone speed to zero, hover, wait few seconds and start to climb.This happend bc the flightmode at reconnected again is Go Home.
So in the 23.4 seconds disconnect time your drone started the RTH.

3% batt juice is not used for touching water, flips out and back again...


End of your video shows drone falling down from some height, about 2 or 3 seconds falling down, so must be from height 10 to 15 meter height.
In the logs no error wich could explain why it did fall out of the sky.

cheers
JJB

alisauskas.simonas@gmail.com Posted at 11-21 03:04
No, I didnt see it that far. But what can i see is a last video captured of it and i clearly didnt make it to fly that high from where it was falling

No, I didnt see it that far.
You were flying in Sport Mode less than a metre above the water.
The water was not flat calm.
Flying so close to an obstacle when you cannot see the drone and how far it is from the obstacle is asking for trouble.

But what can i see is a last video captured of it and i clearly didnt make it to fly that high from where it was falling
What you see in the end of the video is what the data shows.
... that the drone was in RTH and had climbed a little before it shut down because of the water it scooped up when you crashed into the water's surface.

Crashing into the water is what caused loss of signal and the crazy data shown by the barometric sensor after the crash.

Big pilot error.

Thanks for the responses. But it still doesnt explain me the fact why it kept going down. While i was analyzing video and flight log at the same time I noticed that when i pushing the drone go down with left stick only for -1s ( in video this is seen on 3-4s of the clip, and after that I did not touch the left stick but the drone was still going down (6-8s of the clip), so that is pretty interesting for me

And by looking deeper in the flight log when stoped to yawn i actually press left stick up for the drone to get higher, but after that it still kept going down

JJB* Posted at 11-21 03:42
Hi Simonas,

had a look at your log and video.

In the logs no error wich could explain why it did fall out of the sky.
The IMU height data is a pretty big clue

If you are suggesting the drone didn't hit the water (I can't tell from your wording) .....
What would have caused the sudden Upward vision sensor error + loss of signal at exactly the same time, when the drone is that close to water surface (with small wind waves)?

alisauskas.simonas@gmail.com Posted at 11-21 04:52
And by looking deeper in the flight log when stoped to yawn i actually press left stick up for the drone to get higher, but after that it still kept going down
[view_image]
[view_image]

Hi,

It is normal for drone to vary in height when the remote pilot uses rapid RC inputs!
Don`t expect a DJI drone to keep height steady within 1 cm or 2.....

See my chart of your data, right side of the chart stops at the disconnect.

I do not see any out of the normal height variations.

Sorry for your loss, but it was not a clever thing to do to fly that low in Sport mode out of eye sight with many rapid RC inputs.

cheers
JJB

Labroides Posted at 11-21 06:03
In the logs no error wich could explain why it did fall out of the sky.
The IMU height data is a pretty big clue

well   ever seen a drone in sport mode fly into the water....jumpes back, get into hover wait for seconds to engage RTH??

Moment of disconnect, drone stopped flying into hover, waited, into RTH mode en started to climb.
Vertical speed first record at disconnect = 0.9, next records to 0.3, this in RTH.
When does this happen?  when drone stops at 20 meter to yaw back to home direction.
Distance to home is there....

So the only strange value = -411.7   no idea why, but mayby some water ingress to the baro sensor flying so close to the water. (and we do not have data how the drone braked at disconnect, guess really really close to or at the water)


so imo the last records shows craft is flying, not in the water.

cheers
JJB

JJB* Posted at 11-21 06:32
well   ever seen a drone in sport mode fly into the water....jumpes back, get into hover wait for seconds to engage RTH??

Moment of disconnect, drone stopped flying into hover, waited, into RTH mode en started to climb.

It's still hard to tell what you are trying to say, your wording is not clear at all.
But I'll respond assuming you are suggesting the drone didn't hit the water, recover briefly, enter RTH, but fall shortly after because of water damage.

well   ever seen a drone in sport mode fly into the water....jumpes back, get into hover wait for seconds to engage RTH??
Yes, I've seen cases where a drone partially touched the water and climbed away, but not always successfully.
The cases I've seen didn't involve loss of signal, so no RTH.

Moment of disconnect, drone stopped flying into hover, waited, into RTH mode en started to climb.
Vertical speed first record at disconnect = 0.9, next records to 0.3, this in RTH.
When does this happen?  when drone stops at 20 meter to yaw back to home direction.
Distance to home is there....
It's hard to tell what you think you are saying here.

Signal was lost and there's no data for 23 seconds, so some guesswork is needed to work out what happened for those 23 seconds.
The video shows that the drone was going to hit the water.
The VPS suppports this and so does the sudden upward sensor error.
So the only strange value = -411.7   no idea why, but mayby some water ingress to the baro sensor flying so close to the water. (and we do not have data how the drone braked at disconnect, guess really really close to or at the water)
The IMU is inside the drone.
It doesn't breate in water when flying low.
It gets wet when the lower parts of the drone get wet.

so imo the last records shows craft is flying, not in the water.
Like the video, the last records show the drone struggling before it finally falls.

When connection is restored at 2:54.4, the GPS location is 33 metres towards the home point from where signal was lost.
All 15 location data points for 2:54.4 to 2:55.7 after signal was restored are within 1 metre.
If the GPS data is correct, the drones horizontal velocity is almost zero.

We have no reliable height data after signal was lost.
OSD.isSwaveWork = False and the IMU altitude data is garbage
It's reasonable to guess that this is a result of collision with the water's surface and the upward sensor error happening at the same time indicates water splashing on the drone's upper surface.

After losing signal, the Air 2S would have taken 6 seconds before Failsafe RTH initiated and commenced retracing its path until signal was regained.
That accounts for the signal being restored after 23 seconds and the drone having recovered from skidding across the water and fling back 33 metres at low altitude.
But the water damage was progressively affecting drone functions.
It was very sick and unable to acellerate or make good speed and ended up falling into the water.

The video shows that the camera looking up at an odd angle with an arm prominent in view and falling.
Was this also a result of a water collision?
The video freezes at 8 seconds which corresponds with the end of data.

The battery drain for was very low <3 amps after signal was regained.
Definitely not what you would see if the drone was in normal flight.
The video makes it obvious that the drone was struggling to fly properly.

If you can explain this without the drone hitting the water, I'd be interested to see how.
It would require multiple failures to happen at the same time ... the time that the video and data indicate the drone was just about to hit the water.

Labroides Posted at 11-21 18:53
It's still hard to tell what you are trying to say, your wording is not clear at all.
But I'll respond assuming you are suggesting the drone didn't hit the water, recover briefly, enter RTH, but fall shortly after because of water damage.

Hi Labroides,

Interesting your reply, bc we do have the same idea what happend.
Only you do not fully understand my writing...

1)how to explain drone was in RTH after crasing into water and flips out for 23 seconds?
2)
IMU inside drone, correct  but baro sensor is not a watertight sensor. Minus 411 due to water ingress to the sensor.
3)
OSD.isSwaveatwork did show FALSE before the crash as well, 'normal' behaviour flying low over water in Sport mode, so imo no garbage data at all.Only the 4 records just before the disconnect it was True, reading 0.3 vps height. (and True in other parts of this flight as well)
4) Upward vision sensor error also 2x at start of the flight, if one sensor would give as first an error touching water it would be not the upward sensor but downward or side sensors.
So why only the upper sesnsor to give an error? did DJI forgot the other warnings...

After losing signal 6 seconds for Failsafe RTH, correct. So yo do understand what i wrote.
So we agree i guess:

• disconnected flying low at 0.3 meter with speed 11.5 ms, in fwd&roll motion, but did it took 33 meters to stop from 11.5 to zero speed? if all data is correct, than it is 33 meter.
• drone speed to zero, my guess is that during this rapid braking action drone touches water
• drone managed to get dry feet and hover, wait 6 second, into RTH. (cannot think of another reason drone to be in RTH)
• drone gained height to x meter, (for sure in RTH, cannot think of another reason to gain height)
• errors in drone, thus falling down. indeed low in power draw.
  

Not sure , at first sight, if the last records on the log = the falling down in the video. Drone pitch roll and roll angles not as in the video. But as we see a leg of the drone, camera was approx 80 degeers offset from normal position, so correct.
Camera offset most likely due to first minor impact with water. (minor bc it survived this impact)
Don`t understand why the last records showing a positive vertical speed, while losing height. Must be the water 'damage' to the system.

You ask me "Can i explain without hitting the water..."  well i did already in my previous post 16 (...AT...)
See the chart of this flight for the VPS height data.
Green = sensor measuring, White line = not measuring.
I do lot of testing and imo the TrueFalse value on WaveatWork is correct to use.
So for me no garbage, but respect your opinion.


cheers
JJB

JJB* Posted at 11-22 01:30
Hi Labroides,

Interesting your reply, bc we do have the same idea what happend.

3)  OSD.isSwaveatwork did show FALSE before the crash as well, 'normal' behaviour flying low over water in Sport mode, so imo no garbage data at all.Only the 4 records just before the disconnect it was True, reading 0.3 vps height. (and True in other parts of this flight as well)
Is wavework was showing true for quite a bit of the early flight before the drone was at speed and tilting >30°

disconnected flying low at 0.3 meter with speed 11.5 ms, in fwd&roll motion, but did it took 33 meters to stop from 11.5 to zero speed?
No, not 33 metres to stop.
As I tried to point out, the drone came back 33 metres in Failsafe RTH, retracing its path before it reconnected.

if all data is correct, than it is 33 meter.
    drone speed to zero, my guess is that during this rapid braking action drone touches water
    drone managed to get dry feet and hover, wait 6 second, into RTH. (cannot think of another reason drone to be in RTH)
    drone gained height to x meter, (for sure in RTH, cannot think of another reason to gain height)
    errors in drone, thus falling down. indeed low in power draw.
As the drone retraced the path it took initially, it probably wasn't climbing much at all and may even have been splashed some more on the way back (?)

Not sure , at first sight, if the last records on the log = the falling down in the video. Drone pitch roll and roll angles not as in the video. But as we see a leg of the drone, camera was approx 80 degeers offset from normal position, so correct.
Camera offset most likely due to first minor impact with water. (minor bc it survived this impact)
camera survived, but gimbal control system might have been water affected.

Don`t understand why the last records showing a positive vertical speed, while losing height. Must be the water 'damage' to the system.
Gyro sensors that give vertical speed are part of the IMU, like the altitude sensor which was water affected.
Would it be surprising for the gyro to also be giving bad data?

Labroides Posted at 11-22 02:18
3)  OSD.isSwaveatwork did show FALSE before the crash as well, 'normal' behaviour flying low over water in Sport mode, so imo no garbage data at all.Only the 4 records just before the disconnect it was True, reading 0.3 vps height. (and True in other parts of this flight as well)
Is wavework was showing true for quite a bit of the early flight before the drone was at speed and tilting >30°

Yes, see the green line parts in the chart previous post,  about 12x VPS height measuring with pitch down > 30 degrees.

Thanks, i forgot about the retrace route in a RTH.
Do you know what happens flying below < 1 meter baro height and failsafe RTH kicks in ?
(flying close to home drone will climb to 2 meter after that it will fly home at current height).
But what in this situation? will drone retrace route at 0.3 meter ? (i hope not)  of will it climb first, if yes to what height?
iaw Manual it will just retrace route...do you know this ?

Checked the latlon data to the map.  see the chart   red line is 33 meter retrace route.
Purple line is the few records in goHome after the disconnect.
Confusing data,  as GoHome records (after retrace) are not the left side of the red 33 meter line.looks like drone did not retrace its route.
Seen also on the video play back video.And while flight a upward sesnsor error, from start of the flight.

So question is wich data in the log after the disconnect can be trusted ?  Baro height obvious incorrect.
But what about the other data ?  is the vert speed incorrect or not ?  pitch and roll angles correct?  etc ect. For being incorrect all data (not baro) looks normal, for sure if you check the values in relation to each other.

As you said  camera survived, yes it did record video. But not normal to see leg in the video, data shows that camera was approx 80 degrees offset to drone heading.

cheers
JJB

JJB* Posted at 11-22 05:53
Yes, see the green line parts in the chart previous post,  about 12x VPS height measuring with pitch down > 30 degrees.

Thanks, i forgot about the retrace route in a RTH.

Do you know what happens flying below < 1 meter baro height and failsafe RTH kicks in ?
No ... I haven't tested or seen data to show this.

Also, something not related to the crash, but still, worth mentioning. Just how low the cell voltages are from 1 minute 07 sec to 1 minute 40 seconds. Sometimes reaching 3.2V. DJI drones will force land at 3.0V... Why is the voltage so low? Flying in Sport mode might have caused that, but I would guess that it wouldn't go as low as it did in this flight. Also, starting a flight with a battery not fully charged is generally not a good idea, especially a battery that had auto-discharge happen to it.