My Phantom fell from the sky and survived...

[2fast4u]

I've been following the forums and you just hope this doesn't happen but then it does, still I'm very, very lucky. I would appreciate your support in trying to understand what happened.

Great conditions at dusk, little wind, second flight with the DBS antenna. Took off with 95% battery, flew almost 3km out, then about halfway back. I paused to take a couple of pictures, unfortunately no video of the incident because of that. Then turning the bird to fly it back home and without any warning it fell from the sky. I saw a couple of messages flash on the screen but only after it started falling. I had the critical low voltage error (also described in this post:
Phantom 3 PRO Crashed with "Critically Low Voltage" message | Page 3 | DJI Phantom Forum ) and I also saw something with velocity but I don't recall exactly what and I don't see that mentioned in the flight log.

I was about 200m / 650 feet up when it happened. Because trying to take control didn't do any good I did the only thing I could think of: press RTH. I just watched it tumble through the sky in the app. And then... it stabilized, probably around 50m or so in height. It had initiated auto landing because of the critical low voltage error. I looked down for a suitable place, steered it clear from a couple of trees and around 15m in height I lost the connection. This all happened about 1.5km from where I was standing, so I jumped in my car and drove to the landing site. I found the bird about 10m from where I had intended to land, upside down, also about 10m from the waterline of a small canal next to it. Everything in one piece just a little dirt on the propellers. Phew...

So I feel I'm definitely the luckiest Phantom pilot of the day, but I'm a bit scared to take off again before I understand this. I did an IMU calibration yesterday (level surface, cool bird), I rechecked it just now, it says it's fine. When looking at the log the only thing that is abnormal is the number of cell errors on the battery, but then again I did not have any warning in the app before it came tumbling down. I'm 100% sure the battery was properly secured.

Any ideas?

The log can be found here: HealthyDrones.com - Innovative flight data analysis that matters

Thanks

 

[dottat4]

I see cell deviations but not super huge enough to warrant this. Mine went splat when this happened. It made engine sounds on the way down too. I find it hard to believe these are all defects too. Having just received my bird back today with explanation I have asked dji for more info. There's something else erroneous at play here with how the bird detects and takes action upon perceived critically low battery.

Sent from my Nexus 6 using Tapatalk

 

[kphantom]

Interesting how the temperature curve rose substantially around the time the issue occurred. It's unclear if the battery shorted and heated up, or the flight control system was just reacting to issues and revved up the engines to try and gain control. The bird was spinning fast or the compass was going haywire, as there are numerous high degree changes in short time periods.

Could the battery have shorted briefly?

Note the major deviations.

 

[dottat4]

Interesting how the temperature curve rose substantially around the time the issue occurred. It's unclear if the battery shorted and heated up, or the flight control system was just reacting to issues and revved up the engines to try and gain control. The bird was spinning fast or the compass was going haywire, as there are numerous high degree changes in short time periods.

Could the battery have shorted briefly?

Note the major deviations.

What's really more disturbing in these cases is the blackout effect. We all know that you can't simply turn a phantom on and off. Yet somehow this problem seems to do just that.

Sent from my Nexus 6 using Tapatalk

 

[kancur]

You started with 95 battery. But the voltage readings say around 3.6V per cell. That definitely isn't right and I believe that's the cause of other problems. The temperature rise is suspicious too. Definitely send a log to dji.

 

[Maxdroneomaha]

Sounds like what happened to me. Sent mine back to dji. Got it back few days ago...no explanation as to why it occurred. Dont listen to these people who try and tell you battery wasnt put in correctly. All part of the fun i guess. Good luck

 

[dottat4]

Sounds like what happened to me. Sent mine back to dji. Got it back few days ago...no explanation as to why it occurred. Dont listen to these people who try and tell you battery wasnt put in correctly. All part of the fun i guess. Good luck

Agreed. They know something is up. When you got yours back did they have a new battery in it like they did with mine? Pretty sure after they saw the bird they figured I would need a new battery.

Sent from my Nexus 6 using Tapatalk

 

[kphantom]

Sounds like this might be a bad battery. How old is it? How many cycles?

 

[bartold]

Agreed. They know something is up. When you got yours back did they have a new battery in it like they did with mine? Pretty sure after they saw the bird they figured I would need a new battery.

Sent from my Nexus 6 using Tapatalk

Just got mine back last week for a crash like this. Battery got damaged pretty good. They returned it with no new battery. I called and was told that DJI's battery policy has changed. They are no longer replacing batteries.




I was able to speak to my supervisor regarding the missing battery. Logistics determine that they were not going to cover the battery.
The battery no longer qualified for repairs. Logistics determine thru the flight date that this would not qualify for a replacement battery.
Thank you for your time and patience and have a great weekend.

 

[2fast4u]

Battery had 25 charges. I'll contact my dealer and open a case with DJI keep you posted but from what I've seen so far they have replaced birds that crashed because of this but no explanation and batteries are not replaced somehow.

 

[Dunbeazy]

Upload your flight data to healthy drones and see what the battery voltage difference was. DJI replaced my battery 2 weeks ago after I submitted data to them showing the discrepancy in voltage per cells. In all it took about 3 weeks to receive a new 1.

 

[2fast4u]

Opened a case with DJI and awaiting feedback. After studying the replay on my ipad in more detail I've noticed things are a bit more complicated. The copter first starts to spiral / spin out of control but maintains altitude. Then after 20 seconds it starts to descend still spiraling. At 30m, 13 seconds later, it recovers, because RTH was initiated it then rises again to the pre set 100m. Then a critical voltage error occurs, 45 seconds after it first started spinning. Auto landing then goes fine untill it is about 15m in height then it spirals out of control again and lands upside down. It was still on, lying there blinking when I found it. Discovered one stress crack as a result. Strange sequence of events, I wonder what DJI will find.

 

[dronebohire]

Prop wash
The prop is also exerting torque on the air, to be precise, the engine torque. This force also has to be accounted for by a corresponding acceleration of the air passing through the prop. So the cylinder of air flowing through the prop doesn't go straight back, but rotates as well. Again, the rotation speed comes straight from Newton's third law, and is torque divided by mass flow rate and prop diameter. An informal way of looking at it is that the prop is dragging the air around with its rotation.
It not your battery or fault of the phantom. Increasing throttle will probably not help the situation. What can you do about it, not a lot, time you realise there's a problem your phantom has hit the ground. It will probably never happen again. If it happens again keep throttle or reduce, then if possible move the unit to a different position left or right when it settle then throttle up slowly but no guarantee. That it will work. Hope that helps.

 

[witold]

Clearly something is wrong. I wouldn't fly it. Send to DJI and hope for the best. There are internal flight logs on the memory card that is inside the shell on the board that - from what I understand - are more extensive than the logs we get on the tablet. (can someone confirm this?)

Don't listen to the BS about prop wash. That hasn't been an issue since some early versions of P2 firmwares that allowed quick descent. If your P3 is stock and you're not flying at some high altitude locations, this is not a concern.

 

[2fast4u]

Indeed prop wash doesnt seem likely. Opened a case with DJI, filled out all the forms, shipping tomorrow. DJI isn't very forthcoming with info but on the RMA form they have mentioned "engine failure in mid air" as problem description. Seems more likely and then a critical low voltage error on top of it, can imagine that as well. They will do a full check of the bird once received so we will see if that brings anything new. Will keep the thread updated if there is any more info.

Btw I was unable to get the log from the bird (mc data error), so I supplied the logs from the iPad. Would indeed be interested to know if there is a difference.

 

[RF Guy]

What nobody is mentioning is that - flying out far distances puts you into unknown RF environments. There are microwave communications tower paths over many areas. Metro areas especially are getting REALLY RF congested. 2.4ghz 1 watt un-licensed data links are being used all over now. Any AM of FM station towers can completely blank-out a Phantom receiver. 60hz fields from power transmission lines... really bad. Folks think that just increasing antenna gain will give you solid command/control over very long distances. Not so. The Phantom's receivers and processors are not protected from massive overload by strong RF fields.

 

[2fast4u]

Small update: DJI confirmed it was an engine failure. They will repair the Phantom under warranty. Needs a new shell, new GPS module (apparently from crash impact), probably also one engine, that last bit hasn't been confirmed strangely enough. I'll comment on the end result and quality of the support process later, expected turnaround time is 2 weeks