Phantom 4 Crash, DJI at fault?

[Sawyer]

Guys, my Phantom 4 crashed. I was filming the side of a building in Houston when all of a sudden my drone begins moving forward on its own. I pull back on the joystick, nothing happens. The drone continues to move forward and for whatever reason the obstacle avoidance sensors on the front of the drone also fail, resulting in my drone flying into the wall and falling 6 stories. The odd thing is when the drone hit side of the building, it didn't bounce off and fall but instead kept pushing forward into the wall like it was magnetic or something. Crazy. I've uploaded the TXT file and this image. Am I at fault? Thanks

 

[Meta4]

Guys, my Phantom 4 crashed. I was filming the side of a building in Houston when all of a sudden my drone begins moving forward on its own. I pull back on the joystick, nothing happens. The drone continues to move forward and for whatever reason the obstacle avoidance sensors on the front of the drone also fail, resulting in my drone flying into the wall and falling 6 stories. The odd thing is when the drone hit side of the building, it didn't bounce off and fall but instead kept pushing forward into the wall like it was magnetic or something. Crazy. I've uploaded the TXT file and this image. Am I at fault?

I uploaded your txt file to the Phantomhelp.com log viewer so you can see better.
It looks like this: DJI Flight Log Viewer - PhantomHelp.com
There are some strange warning/error messages at the start.
I haven't seen these before - where did you launch from?
The frequent GPS no match messages are a false alarm and can be ignored.
The flight looks fine until you descend and start to lose sats around 12:00.
The further you descend, the less GPS sats your Phantom can see.
Being a P4, the sat count includes GPS and Glonass sats and the Phantom needs to have a minimum of 6 of one type or the other.
That usually means about 9 sats as a minimum
You are losing sats because you are close to the building and it's blocking the view of much of the sky.
Although the flight record doesn't show atti mode until 12:11.8, I suspect the Phantom was already there because the number of sats was jumping up and down for a few seconds with the last time you had 9 being 12:09.5
If it's just a momentary loss, the GPS will try to carry on.
Without enough satellites, your Phantom is in atti mode from then on and will have no horizontal position holding ability.
Control then would have been like driving on ice with no brakes.
WIthout GPS, the avoidance sensors are unable to precisely control the Phantom.
It looks like the impact is at 12:11.3 when the Phantom starts to tumble and fall.
The compass error messages shortly after are due to tumbling after the crash.

Flying close to an obstacle without GPS is always risky.

 

[sar104]

Your recollection may be incorrect, according to the flight data. It was descending slowly and moving towards the building, and attempted obstacle avoidance at 730 seconds. What does the video show? It kept moving very slowly the north east, possibly under the influence of the wind which was out of the south but probably doing strange things around the building, and then hit the building. There was only one impact, at 731.4 seconds, as seen on the pitch data.

It also appears that you did not use back elevator to attempt to avoid the building - the elevator and aileron are both centered and you had not applied either since the 700 second mark. You were applying throttle, both up and down in the last few seconds. Perhaps you were confused?

 

[Sawyer]

Dropbox - IMG_0056.MOV
LINK to Crash Clip.

Thanks for all the help guys. Obviously everything happened extremely quickly or at least that's what it seemed, but I don't recall the obstacle avoidance function working at all as the drone flew towards the building. In the notification image above, it says "attitude to large. exit forward obstacle sensing." What exactly does this mean? I'm assuming this is why the sensors shut down right before impact.

I was in an alley way between two buildings, but I have flown in downtown areas a few times and never had my drone lose signal so abruptly before.

 

[zgembit]

Sorry to hear about your crash mate, I have been there!
I can relate to this issue with my own experience of this message "current RTH will pass a no-fly zone" that I got when not even close to a such zone. What happened to me is: I got that message, lost connection with aircraft, then aicraft started RTH but in the direction away from me! The original RTH point (where I was standing) was erased and new one appeared where the aircraft lost connection. At that point the craft was approx. 1,5km from me and flying away. Only my luck I could re-connect and stop RTH and bring it back manually.

Some people here on forum tend to read those flight rapports as they were holy bible, which they are not. They are probably as holy and accurate as DJI´s crappy firmware, support and customer service. Point is, listen to all the tips but dont let people tell you that everything was your fault.

 

[Meta4]

What happened to me is: I got that message, lost connection with aircraft, then aicraft started RTH but in the direction away from me! The original RTH point (where I was standing) was erased and new one appeared where the aircraft lost connection. At that point the craft was approx. 1,5km from me and flying away.

That would be a very rare and unusual thing to happen. Could you supply flight data to see how it happened?

Some people here on forum tend to read those flight rapports as they were holy bible, which they are not. They are probably as holy and accurate as DJI´s crappy firmware, support and customer service. Point is, listen to all the tips but dont let people tell you that everything was your fault.

I'd trust the evidence of the flight record every time.
They have proved to be very accurate and an important way to learn what really happened to help prevent problems in future.
Please point out any case you are familiar with where this was not so.

 

[zgembit]

That would be a very rare and unusual thing to happen. Could you supply flight data to see how it happened?

I'd trust the evidence of the flight record every time.
They have proved to be very accurate and an important way to learn what really happened to help prevent problems in future.
Please point out any case you are familiar with where this was not so.

I will post flight data on this described issue, it just got pointless at least for me to spend time on solving more issues.
Are you still surprised that unexpected issues and problems do happen with DJI drones?

Point a case were you can say with 100% accuracy that it was not DJI´s fault or of that flight report firing some wrong warnings? Why would you blindly believe in dji´s flight repport? How can you know the compass is 100% functional, GPS, electronics that control sticks? It is a same company that makes that "flight rapport maker" that makes drones and unfinished firmware many people have lots of issues with.
I have seen this warning message, RTH will pass no-fly zone, on my screen a couple of times but its never appearing in warning log for e.g.

 

[sar104]

I will post flight data on this described issue, it just got pointless at least for me to spend time on solving more issues.
Are you still surprised that unexpected issues and problems do happen with DJI drones?

Point a case were you can say with 100% accuracy that it was not DJI´s fault or of that flight report firing some wrong warnings? Why would you blindly believe in dji´s flight repport? How can you know the compass is 100% functional, GPS, electronics that control sticks? It is a same company that makes that "flight rapport maker" that makes drones and unfinished firmware many people have lots of issues with.
I have seen this warning message, RTH will pass no-fly zone, on my screen a couple of times but its never appearing in warning log for e.g.

You made the assertion that we should not trust the flight logs so go ahead and point to some evidence to support that, rather than asking someone to prove a negative. Flight logs have certainly identified hardware, firmware and software issues, so trusting the basic logging of measured parameters is not in any way equivalent to asserting that the equipment, itself, is infallible.

Furthermore, time and time again it has been shown that the data were correct and explained the event under examination. Since log analysis became commonplace and easy, I cannot recall a single case where the log data turned out to be wrong.

 

[zgembit]

Yes I red couple of threads where pilots´ experiences what happened was not same as in flight log, couple of experts here are just blaming on pilots misstakes. Sure, we all make misstakes. It is that overdriven belief in something that is more of a will or need to believe in then it is a thing that you can 100% say it happened or not.
You don't write firmware for dji drones, you didn't write software that makes flight repports. You cannot say 100% gps malfunctioned and fired wrong warning etc. DJI as a company they are, are not giving you backing in that sense, because they make a lots and lots of misstakes and also repair lots of drones for free that are falling from the sky. And many of the times they will not repair drones.

Example from real world: I bought brand new drone, got remote with bad left stick sensor pushing on yaw, even if not touched. Sometimes sensor pushed the yaw alone, but sometimes it did not, unpredictable and completely random! Sensor/stick was failing. My luck it was not on right stick where forward and backward movement is controlled, otherwise it could alone crash in something and it would look like I did that. How would I argument for that for DJI I won't even try to imagine.
(I sent the RC to holland and got wrong RC back! Sent it again and got right one, whole process took approx 1 month, brand new drone!) Terrible!

 

[sar104]

Yes I red couple of threads where pilots´ experiences what happened was not same as in flight log, couple of experts here are just blaming on pilots misstakes. Sure, we all make misstakes. It is that overdriven belief in something that is more of a will or need to believe in then it is a thing that you can 100% say it happened or not.
You don't write firmware for dji drones, you didn't write software that makes flight repports. You cannot say 100% gps malfunctioned and fired wrong warning etc. DJI as a company they are, are not giving you backing in that sense, because they make a lots and lots of misstakes and also repair lots of drones for free that are falling from the sky. And many of the times they will not repair drones.

Example from real world: I bought brand new drone, got remote with bad left stick sensor pushing on yaw, even if not touched. Sometimes sensor pushed the yaw alone, but sometimes it did not, unpredictable and completely random! Sensor/stick was failing. My luck it was not on right stick where forward and backward movement is controlled, otherwise it could alone crash in something and it would look like I did that. How would I argument for that for DJI I won't even try to imagine.
(I sent the RC to holland and got wrong RC back! Sent it again and got right one, whole process took approx 1 month, brand new drone!) Terrible!

You are missing the point and your example is not applicable. The data files contain output from multiple and, in some cases, redundant sensors. If a sensor is bad then it is usually obvious because it disagrees with other data. And while I've read plenty of reports from pilots where they disagreed with the logs, in all cases when it was resolved, the logs were correct.

 

[Bakersfield Quad]

When an airliner crashed, they use the same sort of data to find answers, it works for them why not us?

 

[zgembit]

You are missing the point and your example is not applicable. The data files contain output from multiple and, in some cases, redundant sensors. If a sensor is bad then it is usually obvious because it disagrees with other data. And while I've read plenty of reports from pilots where they disagreed with the logs, in all cases when it was resolved, the logs were correct.

Maybe you are missing the point, and example is applicable and can be actual in every sense.
I simply question those flight reports and DJI products as well, and also people who blindly read them and tell to the other pilots that they mostly should blame themselves.

 

[sar104]

Maybe you are missing the point, and example is applicable and can be actual in every sense.
I simply question those flight reports and DJI products as well, and also people who blindly read them and tell to the other pilots that they mostly should blame themselves.

No, I'm not missing the point, and this discussion is completely fruitless. And stop posting nonsense. You are demonstrating the worst kind of ignorance by rejecting objective evidence in favor of subjective opinion and recollection of events. Your example description, apart from being barely coherent ("stick sensor pushing on yaw" - what on earth is that supposed to mean?), appears to be a hardware control problem. In what sense is that a failure of the data logging system?

 

[zgembit]

No, I'm not missing the point, and this discussion is completely fruitless. And stop posting nonsense. You are demonstrating the worst kind of ignorance by rejecting objective evidence in favor of subjective opinion and recollection of events. Your example description, apart from being barely coherent ("stick sensor pushing on yaw" - what on earth is that supposed to mean?), appears to be a hardware control problem. In what sense is that a failure of the data logging system?

You can ask for clarification if you having hard time understanding something instead of posting ignorant messages yourself.

 

[sar104]

You can ask for clarification if you having hard time understanding something instead of posting ignorant messages yourself.

My questions were rhetorical - I assumed that was obvious. I had no expectation of getting any useful information from you. And "ignorant" doesn't mean what you apparently think it does.

 

[zgembit]

Then you understood both what I ment and you know what "ignorant" mean, you are not so dumb!!

 

[Morningdewd]

Just watch Judge Judy, most people even in the face of overwhelming evidence still wont accept the fact there wrong and will never admit fault , while others tho few and far between can actually go "ya know now that I look at it I screwed up"

 

[RémyHeli]

Logs are maybe not perfect, however background is that Meta4 and sar104 are very experienced and helpful. If I'm correct they i.e. helped pilots to find lost drones by analysing available logs and with thinking out of the box.

 

[sar104]

Just watch Judge Judy, most people even in the face of overwhelming evidence still wont accept the fact there wrong and will never admit fault , while others tho few and far between can actually go "ya know now that I look at it I screwed up"

True, unfortunately. If you look at the subsequent responses of those who come here asking for help with flight analysis, they do fall into those two broad categories though mostly, I would estimate, into the latter. You would think, from the wide range of conclusions from the flight logs ranging from simple pilot error to DJI equipment failure, that it would be obvious to anyone that flight logs are not some kind of DJI trick to hide their liability when things go wrong, but then, of course, there are still those who are unable even to comprehend the process and how or why it works.

 

[Meta4]

Yes I read couple of threads where pilots´ experiences what happened was not same as in flight log, couple of experts here are just blaming on pilots mistakes.
.... You cannot say 100% gps malfunctioned and fired wrong warning etc.

The kind of hypothetical examples you suggest would be shown up in the recorded flight data which comes from many different sensors.
A malfunction in one, would cause problems that would show up in the other recorded data.

I believe your mistrust of the recorded flight data is not justified.
I've analysed many flight data from many incidents and found it to be a lot more reliable than what the flyer thought was happening.
Eyewitness testimony can be very unreliable, an affected by what the observer thought they observed.
This is particularly true, particularly when the observer is disoriented or has misunderstandings of the technology they are using.
If you can point to a single case where you believe that the recorded flight data is unreliable, I'd be most interested to see it.

With plane crashes, accident investigators study the evidence and recorded flight logs to work out the actual cause of the incident to help prevent similar events from occurring and contribute to pilot knowledge.
It's exactly the same with drone incidents.
The recorded flight data is the best tool we have for finding out exactly what happened, rather than what someone (who may or may not be reliable) thinks might have happened.
Sometimes the answer is pilot error and sometimes it's a malfunction, but Phantoms have proved to be very reliable and pilot error, disorientation and misunderstandings ares much more common.