Mavic Air Fly Away

[sar104]

Could it simply be the wrong log?

No - it's the correct log, unless he put a second Mavic into the water.

 

[skiptv]

So to achieve the farthest flight. You need to. Turn off your camera/ no video. Turn off your sensors while flying your drone completely out of site. What’s the point. You can’t get video of the flight. You can’t see the drone to know where it is in relation to objects and with the sensors off there is a good chance you’ll plow into a tree power pole or hard to detect power line. Sounds like a great way to throw away $800 bucks

 

[Air Beetle]

+1 and its illegal in the US and screws the rest of us

 

[JDawg]

No - it's the correct log, unless he put a second Mavic into the water.

LOL. Sorry but that is funny.

 

[Trickydicky]

I would say that you are dealing with two very smart and capable people (@laurens23 and @sar104).

I can't disagree with this, they have both given me far more insight into what was going on with the drone in a few posts compared to a month with DJI getting nowhere.

 

[Trickydicky]

And given that the altitude reading suggests a splash down

Which altitude reading suggests a splash down?

I was actually contemplating flying my new drone over the lake to see if I could see the first one, so I might well be coming on here again with another drowned drone.

 

[Trickydicky]

So either he was mistaken, having consumed too much lowland whisky perhaps,

How did you know?

 

[sar104]

Which altitude reading suggests a splash down?

The zero altitude reading.

How did you know?

When you have eliminated the impossible, whatever remains, however improbable, must be the truth.

 

[Trickydicky]

So until that disconnect I see nothing particularly unexpected in the data.

So nothing in the data suggests a crash into the water before the first disconnect (except the height data).

The zero altitude reading.

But you've already told me not to put too much faith in the height readings when flying over water, and I think you're entirely right. Was the drone in the water at the zero altitude reading or the -2.7metre reading at first disconnect, neither of these corresponds to the actual level of the water which was about 0.5metres below take off level. Surely all you can really say is the drone was continuing to descend at first disconnect. But I've also just noticed that the final rc throttle reading was zero at first disconnect. So I had effectively stopped the drone when it disconnected. So how did it manage to continue to move forward and continue descending and the roll over, all with no input from me. I'm not sure you can argue that it hit the water before first disconnect because it wouldn't have continued to move forward and descend if it had. Other than the final roll angle is there any other bit of data at 563.3 which suggests the drone has crashed. The limited data set I have shows the drone still flying.

If the drone crashed into the water and disconnected, would you really expect it to reconnect a couple of seconds later? The electronics would be fried.

I still maintain the drone developed a fault of some sort, how else to explain it flying away.

Time for a single malt I think!

 

[sar104]

So nothing in the data suggests a crash into the water before the first disconnect (except the height data).

It's a combination of data and information that suggests the water impact:

1. The zero altitude;
2. The disconnect;
3. The brief reconnect with a roll value inconsistent with flight;
4. That it did not return to home after the disconnect.

I would say that your observation that it remained above the water and then flew away is the sole counter-evidence and, as I mentioned previously, is inconsistent with those data.

But you've already told me not to put too much faith in the height readings when flying over water, and I think you're entirely right. Was the drone in the water at the zero altitude reading or the -2.7metre reading at first disconnect, neither of these corresponds to the actual level of the water which was about 0.5metres below take off level. Surely all you can really say is the drone was continuing to descend at first disconnect.

I was referring to VPS height, not baro/IMU altitude. That's usually pretty good over the duration of short flights - within a couple of meters. Arguing about 0 m vs -0.5 m vs -2.7 m is not productive - there is several tenths of a second latency on the transmitted telemetry.

But I've also just noticed that the final rc throttle reading was zero at first disconnect. So I had effectively stopped the drone when it disconnected. So how did it manage to continue to move forward and continue descending and the roll over, all with no input from me.

The log only recorded one data point (i.e. 0.1 s) after the throttle went to zero - that's not long enough to register in altitude change. You centered the elevator over a period of 1.1 s, and that is reflected in the change in pitch. If it had struck the water and suffered a rapid disconnect then the data showing the impact would be missing. The single set of data 1.6 s later is unremarkable except for the roll value.

I'm not sure you can argue that it hit the water before first disconnect because it wouldn't have continued to move forward and descend if it had. Other than the final roll angle is there any other bit of data at 563.3 which suggests the drone has crashed. The limited data set I have shows the drone still flying.

Again - I'm not arguing that it hit the water since you have stated unequivocally that it did not - I'm just pointing out why the data suggest to me that it did. Only the roll value indicates it was down. But GPS health had dropped to 3 due to it only reporting 6 satellites locked at that point, down from 18 just before the disconnect 1.6 s earlier - also consistent with getting wet.

If the drone crashed into the water and disconnected, would you really expect it to reconnect a couple of seconds later? The electronics would be fried.

No - it will disconnect if the antenna goes below the water surface, but it takes time for water to get into the main body and cause significant problems. I've seen several documented cases of Phantoms and Mavics touching down enough to submerge the legs and even the camera, and then take off again.

I still maintain the drone developed a fault of some sort, how else to explain it flying away.

If we believe the final data point in terms of GPS health and number of satellites locked, then it would have rapidly switched to ATTI mode - that would have prevented RTH and it would have moved with the wind. That was out of the NE at the time. Did it head off to the SW? I wasn't clear on that from your previous posts.

Time for a single malt I think!

When isn't it?

Can you get the DAT file from your mobile device? That has a lot more data, although still just at 10 Hz.

How to retrieve a V3.DAT from the tablet

 

[Trickydicky]

Thanks for the detailed response sar104, it is really appreciated.

The log only recorded one data point (i.e. 0.1 s) after the throttle went to zero

I think the point I was trying to make was simply that both the throttle and elevator were at zero at first disconnect. Accepting the lag in transmitted telemetry I can see that the drone would continue to move forward as the speed gradually reduced, so the distance it moved between first and second disconnects can sort of be explained, although it is a bit far if you do the maths. But I can't quite get my head around the difference in height between first and second disconnect. I had originally thought that I must have been continuing to apply rc throttle but now it seems I wasn't. So that final height reading must be suspect as well.

it will disconnect if the antenna goes below the water surface,

Ok, I can understand this but the drone was descending at almost 3m/s at first disconnect. With no throttle or elevator being applied after it is unlikely that the antenna would see light of day again if it had hit the water. Do you think it might reconnect if it was under water? Sorry, I think the reason I keep at these things is because you said there were four different bits of data that suggested the drone was in the water and in isolation I would agree, but when you look at the other data it just doesn't seem to add up.

If we believe the final data point in terms of GPS health and number of satellites locked, then it would have rapidly switched to ATTI mode - that would have prevented RTH and it would have moved with the wind. That was out of the NE at the time. Did it head off to the SW? I wasn't clear on that from your previous posts.

I did suspect the lack of satellites would impact the drones ability to return home. I would have said the heading was nearer W to WSW, but I would also have said that in my memory the drone was flying with purpose, not drifting on the wind - obviously I could be wrong. At final disconnect the drone cannot be flying at 8.5mph (extrapolating the speeds just before first disconnect suggests it would stop just before second disconnect) so this is another suspect reading. I guess it's possible that the drone stopped just before final disconnect and then started to drift away with the wind. At the time I guess I would be looking from drone to controller wondering what was going on. Does ATTI mode impact on the drones ability to avoid obstacles. If it does then the drone is definitely in the water at the end of the lake.


I think this is the DAT file for the flight, wasn't sure what to do with it so it's a dropbox link.

Dropbox - 2018-05-24_14-30-29_FLY034.DAT

Can I ask you another question which is sort of related and may help me in future flights. I keep reading about people getting disconnected but still being able to operate the drone. The drone can't be disconnected if you can still operate it, can it?. So is it just the video feed that disconnects. If you get a disconnect what do you think is the best way of recovering. I've read again that you turn off the controller and turn it on again. (sorry that's two questions)

 

[Trickydicky]

I would say that your observation that it remained above the water and then flew away is the sole counter-evidence and, as I mentioned previously, is inconsistent with those data.

Sorry sar104, just picked up on this and it's exactly why I keep coming back at these particular bits of data

1 zero altitude, so the drone is in the water at around zero altitude, give or take, yet continues to descend to -8.7 metres after hitting the water and moves forward 7 metres in the next 1.6 seconds, this is counter evidence

2 the disconnect, looking at the forums there seem to be many reasons for disconnects, not just crashing

3 the reconnect with a roll value inconsistent with flight - and also a speed of 8.5mph and a height of 0.4 metres and all the other data which suggests the drone is still flying, this is counter evidence

4 that it did not return home - lack of satellites so a switch to ATTI mode. ( although I can concede that you have to ask why it didn't pick up all the satellites, but maybe the reconnect time was too small to give it time to reconnect to all the satellites)

So really what I'm arguing is that it's not just what I saw happen, the data suggests that there is other evidence that the drone didn't hit the water

Time for a single malt! I guess I owe you one.