First crash – indoors

[calmar81]

Dear fellow pilots,

unfortunately I had my first crash yesterday ... and of course with my latest, most expensive drone, the Mavic 2 Pro. I do not really know what happened and since I mainly want to rule out a technical defect, I would be very happy if someone could explain to me how it came about.

I took some pictures for my dad in a hall with 2 tennis courts. The floor covering was renewed and he asked me to take some photos for the club homepage. Everything worked wonderfully. Despite low-hanging lighting, low altitude, safety net, etc. The obstacle detection has worked great, absolutely quiet flight behavior and just as you know from your drones, very accurate flight characteristics and response to my remote control.

When I was done with the recordings, I wanted to show him the drone's point of view and put my white DJI goggles on his head ... Just like in the ad. I flew on and wanted to show him the hall from one corner...



... so that he has a good overview. The drone still responded normally. So I fly near that corner and turn the Mavic towards the center of the hall. The next thing I notice is how the drone drifts slightly to the left ... I tried to use the right stick to steer to the right and to move it away from the wall and the curtain fastened in front of it ... nothing happens. Instead, the drone continues to drift to the left, the rotor blades caught in the curtain and it crashes to the ground. (Link to the drone's video footage) I immediately turned off the engines, but 3 of the 4 rotors were already destroyed. Nothing else happened to the drone as far as I can see, except that the front left foot, where the position light is mounted, is slightly bent inwards and has a gap where it's screwed to the leg with 3 screws ... Very tough little drone.



When I got back home, I installed new rotors and checked the flight characteristics and all system functions. It flies perfectly and I notice no faults or defects in software or hardware.

My questions:

• Was it perhaps due to the DJI goggles? Can they also control the drone's horizontal movement, not just the rotation?
• If so, why the sensors didn't work? I ran the flightlog and in the end it seems to avoid an obstacle (the big wooden beam, or rather the ceiling, because it sinks slightly?). But if so why does it avoid the obstacle in the direction in which both the left side and the rear sensors would have to recognize the wall?
• Why could not I control the drone anymore? I purposely did not steer to the wooden beam (to the front) but to the right.
• Was it a pilot error or a problem of the sensors, software, hardware?
• Where can I get a new foot for the drone or would you send it in completely for a check? As I said, it flies normally and I get no error messages from IMU, sensors, gimbal, camera, etc. and I see no optical damage. Everything else seems fine. Unfortunately, I have no DJI Care and do not know the procedure

Thank you all very much in advance. I'm very thankful for every tipp you might give me!
And thanks for all the content here in the forum. Love being a part of it.

Christian

 

[calmar81]

For those of you, who want to read / decompile / analyse the flight data, I have packed 3 files I downloaded from the DJI Go 4 App on my iPhone via iTunes

 

[Ping^Spike]

Looking at the floor covering / pattern in there mate I'm guessing it was as simple as the downward VPS not being able to get a lock

 

[GavieboyDji]

For those of you, who want to read / decompile / analyse the flight data, I have packed 3 files I downloaded from the DJI Go 4 App on my iPhone via iTunes

GPS when indoors doesn't work so you are relying on vision sensors and when flying in normal mode the side sensors aren't on they only work in tripod mode.

 

[TrayBoz]

Sorry to hear about your crash. It's unbelievable how many crashes happen indoors. Often they are pilots' first crash ever and/or only crash ever. Most experienced pilots don't fly indoors because it's far too unpredictable. GPS can be spotty indoors, so your mav is likely to switch rapidly between ATTI mode and GPS mode, making it extremely difficult to control and cause erratic flight.

Do a forum-wide search on this site for the words "Indoors" and "crash".

I know, the DJI Ads make it look like you can fly indoors just fine. They also give the impression that you can race your mavic through a dense forest and it will automatically avoid all branches. It's called MARKETING.

I'm sure that now we'll hear from a few pilots saying "I've flown indoors lots of times and never crashed!"
Good for you. You're probably the type that relies on the pull-out method of birth control.

 

[calmar81]

Looking at the floor covering / pattern in there mate I'm guessing it was as simple as the downward VPS not being able to get a lock

Hm ... maybe. But I don't understand why the mavic didn't react to my control at all.

Thank you, @TrayBoz
I' wasn't flying in a dense forest There was plenty of space and also 5-7 sattelites visible through the plastic windows. I never wanted my mavic to avoid any trees or small obstacles. It did not even have to avoid the wall... I was on the sticks... sight to the drone. It just didn't want to move right, when I controlled it to do so. So you mean, rapid switching from ATTI to GPS or any Marketing ******** affects my drone more than me pulling that stick to the right? I also do not know why you are so personal now. Maybe I'm to inexperienced... fine. But I'm here to learn.

Again: I'm fine with everybody saying, it was my fault... but I want to talk about my case and my perception here. I did a forum search for "crash", "indoor"...not only on this forum here. But none of what I found explained, what I encountered. Sorry.

 

[calmar81]

GPS when indoors doesn't work so you are relying on vision sensors and when flying in normal mode the side sensors aren't on they only work in tripod mode.

I just had a look in the flight record: There were 7-10 satellites shown in the top bar of the app, when the crash occured. Normally I have around 18 sats. But it was changing quite rapidly, as @TrayBoz suspected. I don't know exactly at what satellite count the mavic shifts to ATTI mode.

But even if there were too few satellites, the sensors should have worked... And even if they didn't: Why didn't my mavic react to my control to fly right?

 

[TrayBoz]

Sorry if I sounded too negative or like I was blaming you. Most people crash when they try to fly indoors - even very experienced pilots - that was my point.

I haven't tested this on my Mav 2 Pro yet, but my original Mavic wouldn't let me launch unless I had at least 10 sats locked in. I would always wait until at least 12 were connected, just to be sure.

I HAVE noticed that with my Mav 2, if I have less than 13 satellites connected, it tends to drift slightly with NO inputs from my sticks. If I wait a few minutes and get 13-22 satellites connected, it will hover rock-steady in the same position without budging. (outside of course)

I simply said rapid switching between ATTI and GPS modes makes your aircraft fly and respond very erratically, and make it nearly impossible to control - regardless of what you're doing with the sticks.

And I was just saying that the words "CRASH" and "INDOORS" occur together in threads a LOT on this forum.

If you feel you have to tempt fate and fly indoors, at least put on propeller guards.

Good luck.

 

[Ping^Spike]

But even if there were too few satellites, the sensors should have worked

No mate, they wouldn't - simply because of the patterned surface you were flying over.

The downward VPS can't lock on to water, it can't lock on to astroturf and it sure as **** can't lock on to whatever repeated-pattern that is you've got going on the floor there

 

[calmar81]

No mate, they wouldn't - simply because of the patterned surface you were flying over.

The downward VPS can't lock on to water, it can't lock on to astroturf and it sure as **** can't lock on to whatever repeated-pattern that is you've got going on the floor there

Thanks again for your answer, I really appreciate it! But for my understanding: You mean even if tracking position during my approx. 12 min flight before with photos and panoramas taken on altitudes like 50cm to 5 meters, all of a sudden the patterned floor could cause a problem? Maybe I was at an altitude where there was a moiré-like flicker in the image of the VPS.

But even if the positioning wasn't possible, why didn't the rear sensor take my drone away from the wall. And even if the VPS is overruling the Vision Sensors front / back (is that so?) ... why didn't my drone just react to my stick-command "get away from that ****ing wall", pulling the right stick to the right... away from the wall.

 

[FLPilotshark]

Never fly your drone inside anywhere...END OF STORY.

 

[Jackcutrone]

His question about why the flooring would make it crash when he had been flying successfully over the same floor for several minutes seems compelling to me. Usually people on the forum are nice enough to take the time to examine the flight data and not just say don’t fly indoors. As I read his original post the only change was that he put his goggles on his father. Were they locked or were they set unlocked so that touching the touch pad affected the flight?

I am not all that experienced either but I can understand why going in and out of gps mode to atti would affect the flight but presumably that too should have been happening throughout the flight most of which was successful since he was able to complete his photographic mission before the nonresponsiveness to the controls began.

He deserves a better answer. I just don’t have the skills to look at his flight log and provide one.

 

[sar104]

For those of you, who want to read / decompile / analyse the flight data, I have packed 3 files I downloaded from the DJI Go 4 App on my iPhone via iTunes

Those are not the correct flight logs. To get the txt log and the mobile device DAT file, follow the instructions in the links below:

TXT log: DJI Flight Log Viewer - Phantom Help

Mobile device DAT file: How to retrieve a V3.DAT from the tablet

 

[calmar81]

His question about why the flooring would make it crash when he had been flying successfully over the same floor for several minutes seems compelling to me. Usually people on the forum are nice enough to take the time to examine the flight data and not just say don’t fly indoors. As I read his original post the only change was that he put his goggles on his father. Were they locked or were they set unlocked so that touching the touch pad affected the flight?

I am not all that experienced either but I can understand why going in and out of gps mode to atti would affect the flight but presumably that too should have been happening throughout the flight most of which was successful since he was able to complete his photographic mission before the nonresponsiveness to the controls began.

He deserves a better answer. I just don’t have the skills to look at his flight log and provide one.

Thank you very much for your nice reply. I had expected more reactions like yours and was a little disappointed with some of the messages above ... until now. Thanks

@sar104 Thank you! I thought I had the right ones in that ZIP-package. I attached the .TXT file to this message. Unfortunately, my FlightRecords/MCDatFlightRecords folder just shows one .DAT file with a wrong time stamp.

 

[sar104]

@sar104 Thank you! I thought I had the right ones in that ZIP-package. I attached the .TXT file to this message. Unfortunately, my FlightRecords/MCDatFlightRecords folder just shows one .DAT file with a wrong time stamp.

OK - it's pretty much what one might expect. The aircraft has marginal and varying GPS reception and is trying to hold position with a combination of VPS and GPS. GPS health, which is really a measure of positional hold ability, drops to zero just before the event.



Much of the advice above is correct - the Mavics, in general, are poor indoors because of their tendency to try to hold position based on GPS when they see it. I only fly mine indoors in ATTI mode for that reason. It's harder to fly, but much more predicatable.

 

[calmar81]

Thanks again @sar104 !
I try to make sense of the diagram, but of course I see and understand that the GPS signal strength drops to zero. Of course, I learn from this event...

But is it clear from the flight log, why the movement of the right stick had no effect? Or why the rear sensor wasn't able to avoid the crash? So was the drone just "confused" and wasn't able to interpret my command or the sensor readout. Is the crash cause of a technical mistake or more of the drone's technical limitation?

I mean... I get the point, that positioning is done via GPS and VPS... and a bad GPS signal + patterns on the ground can of course cause the drone to drift. But Obstacle Avoidance to the back should have worked? When I playback the flight log, it says at flight time 12:27 "obstacle avoided. check flight route." It was maybe the ceiling above or the wooden beam in front. But because of the message I assume that the obstacle avoidance system should have worked also for the wall?

please excuse me for being so annoying and repeating my questions all the time. I just want to understand the event and also my drone's systems better to learn for future flights.

 

[sar104]

Thanks again @sar104 !
I try to make sense of the diagram, but of course I see and understand that the GPS signal strength drops to zero. Of course, I learn from this event...

But is it clear from the flight log, why the movement of the right stick had no effect? Or why the rear sensor wasn't able to avoid the crash? So was the drone just "confused" and wasn't able to interpret my command or the sensor readout. Is the crash cause of a technical mistake or more of the drone's technical limitation?

"Confused" may be an appropriate term. It's difficult to know exactly what happened because we don't know the details of how the FC processes and prioritizes the various inputs, but it looks pretty much like what you might imagine when it gets conflicting data.

 

[calmar81]

"Confused" may be an appropriate term. It's difficult to know exactly what happened because we don't know the details of how the FC processes and prioritizes the various inputs, but it looks pretty much like what you might imagine when it gets conflicting data.

Okay, thank you again!

I think I will send it in to DJI Germany for a data analysis and checks if everthing is right with its hardware. I think that then I have more security for future flights or at least I feel safer that the drone is technically still in top condition.

And I will keep you updated on the outcome... If it's my fault for not knowing better or having better experience, what will be fine for me, or if it was some technical issue or limitation.

Thanks again to you all ... "fine folks", how our moderator keeps greeting all of us after registration

 

[SpinItUp]

Net. 2nd service! Advantage DJI
Sorry just reading the weekly digests[emoji12]

 

[dbyantis]

Dear fellow pilots,

unfortunately I had my first crash yesterday ... and of course with my latest, most expensive drone, the Mavic 2 Pro. I do not really know what happened and since I mainly want to rule out a technical defect, I would be very happy if someone could explain to me how it came about.

I took some pictures for my dad in a hall with 2 tennis courts. The floor covering was renewed and he asked me to take some photos for the club homepage. Everything worked wonderfully. Despite low-hanging lighting, low altitude, safety net, etc. The obstacle detection has worked great, absolutely quiet flight behavior and just as you know from your drones, very accurate flight characteristics and response to my remote control.

When I was done with the recordings, I wanted to show him the drone's point of view and put my white DJI goggles on his head ... Just like in the ad. I flew on and wanted to show him the hall from one corner...

View attachment 53797

... so that he has a good overview. The drone still responded normally. So I fly near that corner and turn the Mavic towards the center of the hall. The next thing I notice is how the drone drifts slightly to the left ... I tried to use the right stick to steer to the right and to move it away from the wall and the curtain fastened in front of it ... nothing happens. Instead, the drone continues to drift to the left, the rotor blades caught in the curtain and it crashes to the ground. (Link to the drone's video footage) I immediately turned off the engines, but 3 of the 4 rotors were already destroyed. Nothing else happened to the drone as far as I can see, except that the front left foot, where the position light is mounted, is slightly bent inwards and has a gap where it's screwed to the leg with 3 screws ... Very tough little drone.

View attachment 53798

When I got back home, I installed new rotors and checked the flight characteristics and all system functions. It flies perfectly and I notice no faults or defects in software or hardware.

My questions:

• Was it perhaps due to the DJI goggles? Can they also control the drone's horizontal movement, not just the rotation?
• If so, why the sensors didn't work? I ran the flightlog and in the end it seems to avoid an obstacle (the big wooden beam, or rather the ceiling, because it sinks slightly?). But if so why does it avoid the obstacle in the direction in which both the left side and the rear sensors would have to recognize the wall?
• Why could not I control the drone anymore? I purposely did not steer to the wooden beam (to the front) but to the right.
• Was it a pilot error or a problem of the sensors, software, hardware?
• Where can I get a new foot for the drone or would you send it in completely for a check? As I said, it flies normally and I get no error messages from IMU, sensors, gimbal, camera, etc. and I see no optical damage. Everything else seems fine. Unfortunately, I have no DJI Care and do not know the procedure

Thank you all very much in advance. I'm very thankful for every tipp you might give me!
And thanks for all the content here in the forum. Love being a part of it.

Christian

There are lots of reports of crashing a GPS controlled drone, indoors. It is possible that it was handling well for a little while, in a brighter area where the optical sensors can work, as you moved to the corner, it is possible that it started to get GPS lock, as that happens the drone is switching between modes and will drift until it finds a stable sensor, alternating between ATTI and P-mode.