My phantom 4 fell out of the sky today :(

sar104 · Apr 26, 2017

CharlieU said:
Try this one.
Airdata UAV - Flight Data Analysis for Drones

That works. Strange data though - there appear to have been large uncommanded altitude excursions towards the end, both up and down. And then the record ends at an altitude of 246 ft. What happened after that? Did it just continue to descend, still powered up, without recording any further data? You don't appear to have been recording video so nothing to help us there. I have no idea what is going on here.

sar104 · Apr 26, 2017

After a bit more thought - barometer failure could have led to that result. The excursions all appear to be vertical. Was it trying to hold station with flawed altitude data, and was the end of the record - apparently at 246 ft - when it actually hit the ground?

sar104 · Apr 26, 2017

Looking further at the altitude record and differentiating wrt time, the excursions at the end seem unphysical, with ascent and descent rates peaking at over 40 ft per second. I suspect a faulty barometer.

jsposito · Apr 26, 2017

tcope said:
After-Sales Service Policies - DJI

DJI - Contact Us

Sorry friend so off topic but where did you get that cool skin for your bird ?

CharlieU · Apr 26, 2017

sar104 said:
That works. Strange data though - there appear to have been large uncommanded altitude excursions towards the end, both up and down. And then the record ends at an altitude of 246 ft. What happened after that? Did it just continue to descend, still powered up, without recording any further data? You don't appear to have been recording video so nothing to help us there. I have no idea what is going on here.

1 was recording video. The flight was just to check out extremely grainy video. The video quit at the time the radical assent and descent ended. Did not record descending for crash. Litchi disconnected and then connected again shortly. I saw on the monitor the P4 descending on a rooftop. It didn't hit the rooftop but hard landed on a dirt pathway. I linked the video to DJI. This is a different video on another day with a P4. This was taken about 300ft from where my bird went down.

Now what do you think? 19 mph wind gust at the time my troubles began. I don't think so.

sar104 · Apr 26, 2017

CharlieU said:
1 was recording video. The flight was just to check out extremely grainy video. The video quit at the time the radical assent and descent started. Litchi disconnected and then connected again shortly. I saw on the monitor the P4 descending on a rooftop. It didn't hit the rooftop but hard landed on a dirt pathway. I linked the video to DJI. This is a different video on another day with a P4. This was taken about 300ft from where my bird went down.

Now what do you think? 19 mph wind gust at the time my troubles began. I don't think so.

OK - Litchi may not have set the video flag in the log. Do you have any thoughts on my suggestion that it was a barometer failure? I also doubt that it was a wind problem, unless that somehow caused a barometer issue.

K9VXV · Apr 26, 2017

I see a wobble like that when I am flying in sport mode and get high winds. It freaks me out and I usually stop and then kick it back to P mode and continue along. Is it possible for the Phantom to flip over in air from high wind and then not be able to recover because of it or am I just being silly for worrying when it does that?

CharlieU · Apr 27, 2017

sar104 said:
OK - Litchi may not have set the video flag in the log. Do you have any thoughts on my suggestion that it was a barometer failure? I also doubt that it was a wind problem, unless that somehow caused a barometer issue.

I don't know much about barometer failure. I tried flying the P4 after the crash. It had wires hanging out of the landing gear but it flew with the same battery without recharging. Airdata UAV - Flight Data Analysis for Drones

Here is a later flight, same P4 with different battery before sending to DJI. Airdata UAV - Flight Data Analysis for Drones

Here is the flight of the crash.

sar104 · Apr 27, 2017

CharlieU said:
I don't know much about barometer failure. I tried flying the P4 after the crash. It had wires hanging out of the landing gear but it flew with the same battery without recharging. Airdata UAV - Flight Data Analysis for Drones

Here is a later flight, same P4 with different battery before sending to DJI. Airdata UAV - Flight Data Analysis for Drones

Here is the flight of the crash.

Video shows as unavailable. The other two flight logs don't show anything anomalous. My guess remains the barometer.

CharlieU · Apr 27, 2017

I'm wondering why it would fly ok after the crash of it was barometer failure. I'm not questioning your assement, I'm just in the dark about the barometer. I'm thinking it could have been hit by one of the eagles in the immediate area Not saying, just wondering. I have seen videos of eagle strikes.

sar104 · Apr 27, 2017

CharlieU said:
I'm wondering why it would fly ok after the crash of it was barometer failure. I'm not questioning your assement, I'm just in the dark about the barometer. I'm thinking it could have been hit by one of the eagles in the immediate area Not saying, just wondering. I have seen videos of eagle strikes.

That's a good question - I don't have an answer. However, if you looked at the graph I plotted, the reported altitude changes occur at rates that are too large to have been real (> 40 ft/s) and, I assume, the reported altitude at impact (when logging stopped) is + 246 ft, which is also wrong, presumably. That suggests, unambiguously I think, that the barometer was reading incorrectly even if something else caused that problem.

I would be tempted to give it some soak tests - power it up and just leave it sitting on the ground and see if there are any unusual variations from zero in reported altitude. If that doesn't reveal anything then put it into a low hover and repeat the test.

CharlieU · Apr 27, 2017

sar104 said:
That's a good question - I don't have an answer. However, if you looked at the graph I plotted, the reported altitude changes occur at rates that are too large to have been real (> 40 ft/s) and, I assume, the reported altitude at impact (when logging stopped) is + 246 ft, which is also wrong, presumably. That suggests, unambiguously I think, that the barometer was reading incorrectly even if something else caused that problem.

I would be tempted to give it some soak tests - power it up and just leave it sitting on the ground and see if there are any unusual variations from zero in reported altitude. If that doesn't reveal anything then put it into a low hover and repeat the test.

Thank you for all of your input. I'm getting an education here.
I would love to do those tests but too late. I sent the bird to DJI for repairs. They covered it under warranty and sent me back what appears to be a new P4. All the stickers were on it. I did not have DJI Care. I did purchase it for my new P4 Pro.

CharlieU · Apr 27, 2017

sar104 said:
That's a good question - I don't have an answer. However, if you looked at the graph I plotted, the reported altitude changes occur at rates that are too large to have been real (> 40 ft/s) and, I assume, the reported altitude at impact (when logging stopped) is + 246 ft, which is also wrong, presumably. That suggests, unambiguously I think, that the barometer was reading incorrectly even if something else caused that problem.

I would be tempted to give it some soak tests - power it up and just leave it sitting on the ground and see if there are any unusual variations from zero in reported altitude. If that doesn't reveal anything then put it into a low hover and repeat the test.

I wish I had your knowledge before I sent it in for repair. Flying one of these is a 4 year college course.

CharlieU · Apr 27, 2017

CharlieU said:
Thank you for all of your input. I'm getting an education here.
I would love to do those tests but too late. I sent the bird to DJI for repairs. They covered it under warranty and sent me back what appears to be a new P4. All the stickers were on it. I did not have DJI Care. I did purchase it for my new P4 Pro.

I corrected the link to the crash video. It does show the altitude changes.

sar104 · Apr 27, 2017

CharlieU said:
Try this link to video.

Thank you for all of your input. I'm getting an education here.
I would love to do those tests but too late. I sent the bird to DJI for repairs. They covered it under warranty and sent me back what appears to be a new P4. All the stickers were on it. I did not have DJI Care. I did purchase it for my new P4 Pro.

OK - the video is very interesting. The altitude variations at the end are at least somewhat real, even if the rates are possibly a bit lower than indicated by the log data. And the video does end in mid-air, quite possibly up around 250 ft, which does not support my hypothesis at all. There is a 3 second discrepancy in the log length and the video length however, given that you appear to have started video at 0.43 s into the log.

My confidence in the barometer error has decreased - I'm really not sure what happened. However, DJI covered it, so presumably the found something in their version of the flight logs that made them at least suspect some kind of failure.

JWH · Apr 27, 2017

K9VXV said:
I see a wobble like that when I am flying in sport mode and get high winds. It freaks me out and I usually stop and then kick it back to P mode and continue along. Is it possible for the Phantom to flip over in air from high wind and then not be able to recover because of it or am I just being silly for worrying when it does that?

I had a heart stopping moment when I slightly clipped a tree. The drone did a complete flip, then flew normally. I didn't know they could do that. It all happened so quickly that I didn't have time to react. It righted itself. I certainly won't test it again! I changed the props just to be safe even though they appeared undamaged. I keep a wider distance from trees too. I had my warning shot!

Duchunter · Apr 27, 2017

JWH said:
I had a heart stopping moment when I slightly clipped a tree. The drone did a complete flip, then flew normally. I didn't know they could do that. It all happened so quickly that I didn't have time to react. It righted itself. I certainly won't test it again! I changed the props just to be safe even though they appeared undamaged. I keep a wider distance from trees too. I had my warning shot!

Oh yeah man! You can throw the dang thing up in the air and it will recover before it hits the ground.