Djo Mavic 2 Pro crashes against building in tripod mode

[Donvictorio]

That worked. The instructions to get the DAT file are in the post linked in my signature.

Mobile device DAT files (DJI GO 4 & DJI Fly)​

​

The DAT file naming convention, based on the date and time of the start of the file, is: YY-MM-DD-hr-min-sec_FLYXXX.DAT., where XXX is the flight recorder file index from the HOME_dataRecorderFileIndex field in the txt log.​

​

These are retrieved by the same method as the TXT logs. Under both iOS and Android they are in a subfolder, MCDatFlightRecords, in the folder that contains the TXT logs.​

Is that the file you require? That's the last flight's record txt I have in the flight's record folder

 

[sar104]

Is that the file you require? That's the last flight's record txt I have in the flight's record folder

No - that's the txt log that you already supplied. You need to look inside the directory DJI » dji.go.v4 » FlightRecord » MCDatFlightRecords for the DAT file ending FLY095.DAT.

 

[Donvictorio]

No - that's the txt log that you already supplied. You need to look inside the directory DJI » dji.go.v4 » FlightRecord » MCDatFlightRecords for the DAT file ending FLY095.DAT.

I can't see such files.

Should I uninstall dji app and reinstall again as suggested in your post?

 

[sar104]

I can't see such files.

Should I uninstall dji app and reinstall again as suggested in your post?

You mean the directory is empty? If so then reinstalling is not going to restore files that are not there - it only helps in terms of recording future files.

 

[Donvictorio]

You mean the directory is empty? If so then reinstalling is not going to restore files that are not there - it only helps in terms of recording future files.

It seems I don't have the folder you mention at all.

Is it possible that when I synchronized dji go4 with the cloud, that such folder has been deleted by the app?

 

[MaxE1955!]

Fewer things more gut wrenching then seeing your expensive drone not respond to stick inputs! Especially in close proximity to obstacles. Sorry for your bad luck.

 

[sar104]

It seems I don't have the folder you mention at all.

Is it possible that when I synchronized dji go4 with the cloud, that such folder has been deleted by the app?
View attachment 96826

That's a known problem only with the DJI Fly app, not DJI GO 4, and it doesn't remove the directory, just the contents. I haven't seen a case where there is simply no MCDatFlightRecords directory.

So without that file there are a few observations that can be made. The velocity record at the end of the flight (after 1381 seconds) is purely inertial since the aircraft had lost GNSS positioning confidence. It's also clearly wrong, as compared to the video record, which confirms what you described. The problem, in terms of the warranty claim, is that there is no expectation or promise by DJI that the aircraft will be able to hold stable, controlled flight if the GNSS solution is lost. If it had dropped into ATTI mode then your stick inputs would have arrested the uncommanded movement, but it doesn't make that switch immediately and so it crashed before that happened. DJI can point to pilot error for flying in that environment and cite the user manual to back that up.

Your only option is to try to argue that the firmware is poorly programmed and that it should engage ATTI sooner, or point out that the IMU solution was incorrect. I doubt that either argument will work because the former is a stated limitation and the latter is an unrealistic expectation from the IMU data. If you had been able to retrieve the DAT file and it turned out that the backup IMU disagreed with the active IMU then it would have been a stronger argument.

 

[steve_lyt]

if you had switched to p mode the drone would have avoided crashing, as all the sensors would have been active

 

[Donvictorio]

Sar104 and Meta4, thank you very much for the time and effort you put to see into my case.

Although, I do not see how could the dat file be retrieved and guarantee will not cover this crash I am very grateful for your inputs and the discussion.

I know it's easy to judge by looking at the flight record and rely on what drone thought it was doing and simply say it was pilots fault.

But as reasonably experienced pilot I saw all the situation, I immediately noticed the incorrect movement and I applied counter measures, just to see that they would not bring any response.

My personal opinion is that dji should improve firmware on situations where gps quality is not good. It's usually done by using dual systems, median calculations and also building up model which in emergency drone would go to and remain smooth and safe (in my industry, once gps is lost or rejected due to quality, device go into "dead reckoning")

I wish you all the best

Vic

 

[sar104]

if you had switched to p mode the drone would have avoided crashing, as all the sensors would have been active

No - it would have made no difference.

 

[sar104]

Sar104 and Meta4, thank you very much for the time and effort you put to see into my case.

Although, I do not see how could the dat file be retrieved and guarantee will not cover this crash I am very grateful for your inputs and the discussion.

I know it's easy to judge by looking at the flight record and rely on what drone thought it was doing and simply say it was pilots fault.

But as reasonably experienced pilot I saw all the situation, I immediately noticed the incorrect movement and I applied counter measures, just to see that they would not bring any response.

My personal opinion is that dji should improve firmware on situations where gps quality is not good. It's usually done by using dual systems, median calculations and also building up model which in emergency drone would go to and remain smooth and safe (in my industry, once gps is lost or rejected due to quality, device go into "dead reckoning")

I wish you all the best

Vic

There are no easy ways to deal with a loss of GPS if the inertial system loses track of velocity and the vision system is out of range. Dead-reckoning was exactly what the FC was trying to do, and it didn't work. I'm not sure what you mean by "dual systems" in this context. The traditional method is for the FC to give up and switch to ATTI but, as I mentioned, that doesn't always happen fast enough. The other alternative is to allow the pilot to switch to ATTI, as is possible with the Phantoms. Realistically that is only robust approach, and so I have ATTI mode programmed into the Tripod mode switch position.

 

[steve_lyt]

No - it would have made no difference.

tripod mode flew into a low tree, when I pass the p mode, p mode will stop before it hits anything, it doesn't stop object that move towards the drone.

 

[sar104]

tripod mode flew into a low tree, when I pass the p mode, p mode will stop before it hits anything, it doesn't stop object that move towards the drone.

Yes, but the aircraft can't do anything if it has incorrect information on its own velocity.

 

[RayOZ]

Probably could descend lower so the VPS can hold position. Are you sure you didn’t get confused with your left/right with the drone facing you?

 

[sar104]

Probably could descend lower so the VPS can hold position. Are you sure you didn’t get confused with your left/right with the drone facing you?

It's clear from the txt log and the video that the aircraft was facing the operator. VPS would have kicked in if it had been lower and it was descending, but unfortunately not low enough.

 

[Donvictorio]

There are no easy ways to deal with a loss of GPS if the inertial system loses track of velocity and the vision system is out of range. Dead-reckoning was exactly what the FC was trying to do, and it didn't work. I'm not sure what you mean by "dual systems" in this context. The traditional method is for the FC to give up and switch to ATTI but, as I mentioned, that doesn't always happen fast enough. The other alternative is to allow the pilot to switch to ATTI, as is possible with the Phantoms. Realistically that is only robust approach, and so I have ATTI mode programmed into the Tripod mode switch position.

Is it difficult to set tripod mode to be atti mode?

In the future, I will probably buy another drone and would love to have this option available for emergency situations like the one we talk about.

By "Dual(or triple) systems" i meant for example mounting two GPS antennas on opposite places, comparing results or using bearing and distance between drone and remote controller as another positioning reference. I am talking about redundancy here. English isn't my primary lang but I hope I made my point clear.

 

[Donvictorio]

Probably could descend lower so the VPS can hold position. Are you sure you didn’t get confused with your left/right with the drone facing you?

Not much lower i could have gotten due to trees along the sidewalk, but surely few meters of clearance was there (the clearance I keep for safety).

At some point I did get confused if I was really pushing it to the right direction. But now I see i was pushing it to the right which is correct.
And by the way, all people you see in the photo are involved in drone recovery. This area is normally off limits, taped off

 

[sar104]

Is it difficult to set tripod mode to be atti mode?

In the future, I will probably buy another drone and would love to have this option available for emergency situations like the one we talk about.

By "Dual(or triple) systems" i meant for example mounting two GPS antennas on opposite places, comparing results or using bearing and distance between drone and remote controller as another positioning reference. I am talking about redundancy here. English isn't my primary lang but I hope I made my point clear.

Mavic 2 Atti mode

Redundant GPS antennas won't help if you simply don't have a view of enough satellites. Bearing to the RC would require directional antennas or a phased array - not practical for this technology, and not going to work well anyway in an environment with multiple reflecting buildings.

 

[Keule]

Is it difficult to set tripod mode to be atti mode?

Check this thread: Mavic 2 Atti mode

By "Dual(or triple) systems" i meant for example mounting two GPS antennas on opposite places, comparing results or using bearing and distance between drone and remote controller as another positioning reference.

With limited skyview like your incident showed, also my Matrice 600 with 3 GPS antennae will not get a reliable GNSS lock. ATTI mode is the only way to operate an AC with no/limited skyview.

 

[m80116]

I just want to let you know that @Donvictorio has my sympathy for something I feel it shouldn't have happened and it normally doesn't happen, and for his wilfulness of his opinions.

As I lack the ability to dig deeper into the matter I can only say that I've seen Mavic Pros performing incredibly in more than challenging situations, and I refuse any explanation that "he shouldn't have flown in those conditions because that manual says so", otherwise one what has it got a Mavic PRO for ?

Whatever it was I am pretty sure it can be isolated and discerned against all odds, being an FC biased data processing, glitch or hardware fault. I am sorry for what happened.