Mavic 2 Pro strange crash

[Cymruflyer]

True, but I was thinking in terms of it being a metallic tower [not certain exactly what kind of tower it is] but metal structures or buildings can reflect electromagnetic waves causing issues.

Well, the photo shows this to be an old castle and there was very little metal in these old castle structures, if any. And even if there was, the GPS signal is not affected by metal, unless it is a metal roof and side that you are inside, which would block your signal.

The drone was outside in the open, so GPS signal is not going to be affected by any metal structure, as you suggest. If metal was a problem for GPS you would not be able to use it in your car because you are surrounded by mostly metal in that situation, if you think about it.

 

[Cymruflyer]

Very close to the tower (just before contact), GPS count is decreasing from 14 to 10, according to the log file. This is low as to compare to the Mavic 1 (20 satellites instead of 14) but as I said, My Mavic 1 receive Gallileo, not the Mavic 2.
Yes, acting on the right joystick is normally effective during a mission (you can set the mission speed to zero and fly it by acting manually on the right joystick). Speed changes as well as gimbal pitch are transmitted live by the RC during a Litchi mission and acting on the right stick overcome what was registered in the mission settings.

Even 10 or 8 satellites are a lot, with regards to accuracy for a GPS.

Could you please tell us what you had your "Return To Home" minimum altitude, set to? If you did not have it set high enough, that is possibly your problem, if it lost contact with you.

 

[Cymruflyer]

@sar104

What do you think about this strange accident ?
Could it be the setting sun wrongly recorded by the rear camera as an obstacle that stopped the Mavic 2 ?
I had in two occasion the Mavic 1 stopped during missions by the rising or setting sun right in front of the drone. But this was in open space and I can't say whether the stopped drone drifted or not by some metters at this moment.
Anyway, the front sensors should have recorded the very close proximity of the tower and they didn't, apparently.

When such a thing happens, I wonder if you do a quick yaw of 45 degrees, whether that would be enough to turn the drone away from the sun so that the sensors were no longer being affected, in order to continue flying in control again?

 

[Peio64270]

When such a thing happens, I wonder if you do a quick yaw of 45 degrees, whether that would be enough to turn the drone away from the sun so that the sensors were no longer being affected, in order to continue flying in control again?

From my experience with the Mavic 1, it is pretty difficult to move the drone when it is stuck by the false obstacle created by the sun. The first time it happened to me during a waypoint mission, action on the sticks remained without any evident effect until I switched to Sport mode. This time I thought that the drone facing the wall and flying so slowly would stop before hitting it, thanks to the front sensors. When I realized that this would not be the case, it was too late. I pulled the right stick without effect and some tenth of seconds later, I eared the propellers hitting the stones and the drone falling.
But I'm not sure that this was a consequence of the sun on the rear sensors. I had no warning and anyway the drone was set to "stop" on obstacle detection, not to avoid it. I can not explain its slow drift toward the tower.

 

[Peio64270]

Even 10 or 8 satellites are a lot, with regards to accuracy for a GPS.

Could you please tell us what you had your "Return To Home" minimum altitude, set to? If you did not have it set high enough, that is possibly your problem, if it lost contact with you.

RTH was set at 100m (roughly 300'). And contact was not lost, at any moment including after the crash (look at the log).

 

[Peio64270]

To me, the main question not answered is : "Why the same mission was flown perfectly by my old Mavic1 and went in a crash with the Mavic 2 after the drone stopped in view of the home point -last waypoint- where I stood ?". What difference between the two drones can explain this ? The only one I see is the abundance of obstacle avoidance sensors in the Mavic 2. Maybe the OA became confuse with the sun behind and the wall forward, this explaing the rather erratic slow final flight toward the tower while the drone was still in "Waypoints" mode (yellow mark).

 

[Yorkshire_Pud]

I think you should have a readable DAT file, perhaps that contains more info.
From memory the forum won't allow the direct upload of DAT's so you would have to upload it to a hosting site, make the page public and post a link here

 

[Cymruflyer]

From my experience with the Mavic 1, it is pretty difficult to move the drone when it is stuck by the false obstacle created by the sun. The first time it happened to me during a waypoint mission, action on the sticks remained without any evident effect until I switched to Sport mode. This time I thought that the drone facing the wall and flying so slowly would stop before hitting it, thanks to the front sensors. When I realized that this would not be the case, it was too late. I pulled the right stick without effect and some tenth of seconds later, I eared the propellers hitting the stones and the drone falling.
But I'm not sure that this was a consequence of the sun on the rear sensors. I had no warning and anyway the drone was set to "stop" on obstacle detection, not to avoid it. I can not explain its slow drift toward the tower.

Now that is interesting and brought back 2 flights from 2 years ago, of mine.
The first one was when I was trying to fly down to myself through a hole in the branches and leaves of a large tree in my front yard. I sat on the porch steps and could clearly see the hole my Mavic 2 Pro could fly down through. I slowly started the forward and down descent. Sensors stopped its forward movement, so I slowly backed up a tiny bit. I could see the drone clearly, knowing it had the space and I could see the drone's camera view, obviously.

I started very slowly trying to go forward and down as the angle was about 45 degrees down to me from the drone, to get through that hole. It moved a bit then again stopped, so I again tried to move it slightly sideways an inch or two and it started to move but then stopped again due to branches around it.

Had I been able to fly what I wanted, it would have made it through, going slowly forwards and down. However, it stopped, stuck there and it would not let me move it up, down, back, forward, it just stayed there. Suddenly it began to move to my left, its right very slowly, all on its own. I quickly correct that and put it back an inch or two where it was (to the left of it self, or to my right.

It did this one or two more times but I could not do anything to try and back it out and up or forward and down. Eventually it kept doing that sloooow sideways flight and I had no control over it, despite sitting there 30ft in front and below the drone. It did not respond to any control input and eventually the starboard props touch twigs and leaves and it fell to the ground. It damaged a leg and the camera gimbal. I sent it to DJI in the end for repair.

The other time it would not respond was when I was flying along an open trail about 6 feet off the ground in the woods. It was a walking trail and I was just filming it slowly moving forward. I came to a tree that was leaning out into the trail walkway and had a limb going straight out to my right. I could easily fly up a few inches and get over the limb and then I thought I would drop down a few inches and continue the forward flight.

I got it to go up but then sensors stopped it climbing anymore due to leaves and twigs above it. I tried to go a few inches sideways but it would not respond. I tried to back it out of where I had flown it, but again it would not respond. This time it just sat there hovering but I had no control. In each event I never tried putting it into sport mode because I was too close to everything around me and was afraid an input would be too quick and it would crash.

So there it sat for about 3-40 seconds and I had no control over it but it was holding steady. I tried every movement but nothing worked. Thankfully it was only 6ft off the ground and this time I was able to reach up and grab it and turn it off. Had it been higher up and stuck, I would have been in the same situation as before and it would have eventually crashed again.

A friend who is highly experienced told me that the sensors try to keep it from moving into something and that is why it gets stuck. He went on to say, the problem with the Mavic 2 series is that once this happens and it sits there for a many seconds, stuck, the sensors seem to go funny and the drone starts to slowly drift sideways all on its own. If there is a clearing it may fly itself out by going sideways but usually it is stuck due to obstacles and will eventually impact something and crash. I thought you might find this of interest, especially as to what my friend had found, regarding the sensors, when it gets stuck.

 

[Yorkshire_Pud]

I had a M2P go bananas in my first ever flight with it. I had to jump and grab it from the air. Why?
I had been an idiot and left OA on, the drone was surrounded by hedges, trees, the house and me. OA had been switched on by the previous owner and I had not flown a drone with OA before.
DJI switch OA off for a reason when doing automated landings.
I NEVER take off or land with OA on, in fact I rarely use OA at all.

 

[Peio64270]

To answer the two last posts (#Cymruflyer and #Yorkshire_Pud) after my two alerts with my Mavic 1 obviously blocked by the sun, I systematically switched OA off for all my Waypoints flights with Litchi app. As this one was my maiden flight with the Mavic 2 Pro at a pretty low altitude, I decided to leave this mode "On" in the two drones to see how they both manage with potential obstacles in the way (that is why my mission was programed so close of some trees and this tower).

Since the old Mavic 1 did not report any problem and did a perfect flight, I sent confidently the Mavic 2 on the same path. I think now this was my error and that it is likely that the OA algorithm can become confuse when obstacles are truly or fasly (sun) detected in opposite directions.

Lesson learned. In the future, I'll switch off this bloody mechanism.

 

[Peio64270]

I think you should have a readable DAT file, perhaps that contains more info.
From memory the forum won't allow the direct upload of DAT's so you would have to upload it to a hosting site, make the page public and post a link here

Strange enough (again) I have a nice .DAT file with the Mavic 1 but not with the Mavic 2. Only the CSV file and a .TXT file in my phone.

 

[Yorkshire_Pud]

You could perhaps connect the phone to a computer and use the computer to run a search of the phone for "MSDK".

 

[Peio64270]

You could perhaps connect the phone to a computer and use the computer to run a search of the phone for "MSDK".

I find only one corresponding to this date : the one of the Mavic 1 flight.

 

[Yorkshire_Pud]

that's a pity.

 

[Lespaxvera]

Possibly a loss of signal from the RC coupled with RTH being set to 'hover'? The drone then 'drifted' in the wind?

Mavic 2 does not need connection to RC to fly missions. Mission is uploaded to drone. Any DJI drones which uses Go4 app does not need RC connection, NOT like new using Fly app.
Safe flying

 

[Peio64270]

Mavic 2 does not need connection to RC to fly missions. Mission is uploaded to drone. Any DJI drones which uses Go4 app does not need RC connection, NOT like new using Fly app.
Safe flying

You can set your mission to "RTH if connection lost" but this was not my option. Anyway, no problem with the connection at less than 80m distance (as shown by the attached log).

 

[Yorkshire_Pud]

When such a thing happens, I wonder if you do a quick yaw of 45 degrees, whether that would be enough to turn the drone away from the sun so that the sensors were no longer being affected, in order to continue flying in control again?

A plain yaw will not eliminate movement along the original flight path immediately, as I know to my cost of two props, coincidentally with a Mavic 2 Pro ending up hitting wall. The original direction of travel takes some time to bleed off.
You would need to do a banked turn to eliminate the original direction of travel immediately.

 

[3rdof5]

@sar104

What do you think about this strange accident ?
Could it be the setting sun wrongly recorded by the rear camera as an obstacle that stopped the Mavic 2 ?
I had in two occasion the Mavic 1 stopped during missions by the rising or setting sun right in front of the drone. But this was in open space and I can't say whether the stopped drone drifted or not by some metters at this moment.
Anyway, the front sensors should have recorded the very close proximity of the tower and they didn't, apparently.

Your Theory is entirely possible. The P4's also didn't like direct sunlight, a common concern. Seems logical but as you say, why didn't the front sensors react as well? Hmmm....

 

[Touring497]

I do not think it involved GPS. Interference between the Remote Controller and the Drone usually results in the drone drifting or flying aimlessly or without guidance. GPS will not lock it in a 'hover' and, until RTH is established with the help of GPS, the drone just flies aimlessly. VLOS might have happened at the far end of the fortress? The altitude flown seems to have been 50 ft. Were you able to see the drone at all times at that altitude. Your RC-Drone signal is visual-line-of-sight with some help from reflective surfaces. If the signal became obstructed the RTH would initiate and the RTH altitude,etc, would take time to initiate a return. If the signal returned at the point you saw it again at the corner tower perhaps the confusion took over and the drone became a 'fly-away'. I've lost my Mavic 3 due to RF interference in visual control at 50' from where I was standing. I confirmed this with the data from the RC Pro.

Scary prospect for any drone What was the source of the RF interference? Thnx for any info.