Battery Ejection

[Erised]

No it's further damage the quad has crashed.

 

[Jussaguy]

After the kids went to bed last night I finally had some free time to play with the P4 I've had since release day. Got my Beta copy of Litchi all set up and a few test journeys plotted in.

First was a waypoint orbit to a building down the highway about half a mile. Orbited at 130 feet, shot some video, and flew home. Everything went great, I didn't even really bother orienting the antennas and still never went below about 50% on the signal.

Second was a lap around my subdivision with a few sweeps at the corners. I started to lose signal when it was in the far corner from me, but I wasn't too worried since I was flying lower (80 feet) and there are a row of two-story townhouses blocking line-of-sight. Signal dropped. Still not worried, the whole flightplan was maybe five minutes long, and I was more than halfway done, so I should hear it coming in no more than a minute or so.

When that doesn't happen after about four or five minutes I start to wonder if somehow the failsafe wasn't working and it decided to either hover or land if it lost signal. I head over on foot to look for it near where I last saw it on the map, but it's nowhere to be found. If it tried to land it would have possible been in somebody's yard or a tree, and it's after midnight at this point so I don't want to get shot creeping around somebody's yard so I decide to call it a night and look in the morning. On the way home I realize that even if it hung in a tree or landed somewhere out of sight, I still should have been able to get a signal on the remote (It had plenty of battery when it left and had only flown for maybe 3.5 minutes). The fact that I wasn't able to reconnect seems...not good.

Woke up this morning and headed over there again to see what I can find. Sure enough, first driveway past where I saw it on the map...

Well lookie there: Imgur: The most awesome images on the Internet

You may have to blow that up to full-size to tell but on the left edge of the driveway is my P4 (upside down) and in the bottom-right corner are the pieces of my battery. The white pieces between them are just random trash, not anything related to my drone.

Battery closeup: Imgur: The most awesome images on the Internet

The bird closeup: Imgur: The most awesome images on the Internet

Back in my car right-side up: Imgur: The most awesome images on the Internet

What used to be a set of props: Imgur: The most awesome images on the Internet

The only visible damage other than a little scuff: Imgur: The most awesome images on the Internet

Motors all turn fine and new props will lock in place (even on the broken mount) but those things spin so fast I don't want to run it with that chunk missing.

Bad news not evident from the photos: Gimbal is trashed. When I turn it on the gimbal goes to one side and I get "Gimbal motor overload" error. Camera still works, though.

Guess I have an appointment with DJI repair in my future, would appreciate any tips or recommendations on that.

My questions:

1. Why did I crash? I was flying at 70-something feet, the trees there are maybe 50' tops. The video doesn't show anything, it was too dark to see if any branches were close and it just all of a sudden ends while the drone is in the air. If I had hit a tree wouldn't there have been a few seconds of tumbling/falling?

2. How the hell did that battery get so far away? I bet it was 20 feet from the drone, and all of the pieces were together (although I did move them a little closer together for the closeup photo). That seems like the battery ejected mid-air rather than on the ground. An impact hard enough to throw the battery that far would have done more damage to the drone than what it was.

Any ideas?

If I had to guess (and I have yet to read the thread) based on the damage, how you are describing the damage, I think it tried to land and since it was dark, could not sense the tree enough to stop it on descent, and it hit the tree which caused a "spin out" and you crashed from some height straight down after it hit the tree. The battery looks pretty bend so that might be an indicator that it came out of the bird. The bird is in surprisingly good condition from a fall anything more than 10 feet IMO. I'm going with some sort of tree hit as my most likely scenario and the ejected battery is a byproduct of that.

However, we have heard of a couple ejections of battery but that dude had a mutated fuselage and still flew it.

Also, I wouldn't try beta apps at night if I were you. It

 

[storrowdrive]

I have had some odd experiences with the beta version of Litchi on my P4 when executing a flight plan that involved signal loss. I won't use it anymore until the full version comes out.

 

[Jussaguy]

I know you said you cleared up the trees, but at nighttime the object avoidance sensor don't work. I noticed that the other day. Flight log would show that. Feel free to shae flight log and video so we can help investigate the cause of failure/crash.
I know battery ejects pretty easy, but after some sort of crash. I doubt it will eject by itself, but hey, it could happen.
Another important think about the video, if it works as in gopro, in order to save the file, it will cut the last fraction of second, or full second. It happened to me one time I crash my vortex with a fence. It recorded everthing but ince approacjing the fence, it stops. No video of the crash itself.

Let us know. We are curious now

Actually, the reason there is a buffer is because it is saved in a tiny RAM buffer so you can have variations in speed on the card and because it needs to encode to H264 on the fly so there is always a little video in cache that may or may not make it to the bird depending on what breaks in the crash.

 

[Jussaguy]

I'm going to trim the video and upload it later, but it doesn't show much. You can kind of see some branches down below, but the drone was flying almost dead sideways at the time, with the gimbal down only a few degrees, so there wasn't much visibility in the direction of travel. The fact that it was going sideways seems like it would be hard to impact in a way that would throw the battery, and it would have been pointing the other way.

I was watching the flight in Litchi, not DJI GO, and it wouldn't allow me to "stop" the mission while it was disconnected so there is no flight log that I can find, although I haven't really dug very deep.

This is my first DJI, and my experience has been that ejecting the battery has been anything but easy. That's what was so surprising about the distance between them.

There's nothing nearby that the drone could have "hit" in a way that would blow the battery out (like a pole or tree trunk or anything solid). The only thing it could have run into would have been branches, but if that happened I would have expected it to have recorded at least some of that to be on the video.

Friendly word of advice. Stop saying you were using Litchi as it voids your warranty and if it's a DJI issue, even if the battery ejected based on their terms and agreement, you aren't entitled to a warra

I just crashed my P4 too. Was trying out the active track flight mode and tried to get the bird to follow me between a light pole and chain link fence that were only about 10 feet apart. The drone was about 15 feet behind me and 7 feet off the ground when I initiated active track. The obstacle avoidance worked as the drone stopped following me when approaching the light pole and hovered ,but I then tried to backtrack thinking I don't need to test my luck getting any closer to the pole. However while walking backward the drone compensated by also flying backward to keep me in the frame . It hit the chain link fence at 6 feet altitude. Luckily, only damage was 4 bent/destroyed props. No scuffs or dings on the chassis. Put in my spare battery and kept flying. Whew.

However the battery popped out immediately after the collision like you are describing above, even though it was snug and inserted correctly preflight. I realized after the fact that I had "fly backwards" enabled in the dji go app settings that caused this misshap. I knew I should have left this mode alone but It was my second flight since purchasing and I wanted to give it a go. Be careful when using active track and disable the flying backward button.

Just for people to know, the default setting on this I believe is "OFF" and this is why. You can turn on or off the birds desire to go backwards in order to keep you in the field of view by going backwards. At the risk of rubbing it into you you Squemish (as I'm in reality very sorry for anyone who experiences a crash, it sucks big time), I would leave OFF the setting to "go backwards to keep tracked object in frame" because as you see, you lose control of your bird in that mode. Unless you're object is in a place not near anything in sight to turn on this at default setting of "do not go backwards".

I'm pretty sure it defaults at that anyway. Can't remember if I saw it or turned it off. I go though and make any changes to every setting in a new item on the first run. I remember that one but cannot for the life of me remember for sure either way.

 

[Txtwostep]

You're right, the flying backwards option is defaulted to "off" but I had inadvertently turned it on somehow without realizing it. Didn't expect the drone to fly backwards as I approached it and then smash.

Good lesson to learn. Only use Active track in completely open areas with no obstacles or elevation changes.

 

[Squeamish]

Friendly word of advice. Stop saying you were using Litchi as it voids your warranty and if it's a DJI issue, even if the battery ejected based on their terms and agreement, you aren't entitled to a warra

Ehh, I was flying at midnight with 3rd party software, this is my fault.

 

[Ramphex]

Quick question. You said you were flying at 75ft. How is the starting ASL altitude comparible with ASL altitude of where it crashed? Got to keep in mind that 75 feet altitude is from home point so if you flew up hill then you were lower to the ground than 75.

 

[Squeamish]

Quick question. You said you were flying at 75ft. How is the starting ASL altitude comparible with ASL altitude of where it crashed? Got to keep in mind that 75 feet altitude is from home point so if you flew up hill then you were lower to the ground than 75.

About 3-4 feet difference.

 

[Ramphex]

Not too bad. Wish you could get a log, would show complete power loss vs impact and then loss

 

[Squeamish]

Not too bad. Wish you could get a log, would show complete power loss vs impact and then loss

I posted the log from healthydrones.com elsewhere in this thread. Didn't see an impact in there or in the video (which I still have yet to edit/post).

 

[Ramphex]

I posted the log from healthydrones.com elsewhere in this thread. Didn't see an impact in there or in the video (which I still have yet to edit/post).

Ahh I see it now. Looking at the map it could've caught those power lines or the tip of that tree, hard to tell though. My trees go up 70-80 feet so I never fly lower than 120 to also compensate for elevation changes

 

[Squeamish]

Ahh I see it now. Looking at the map it could've caught those power lines or the tip of that tree, hard to tell though. My trees go up 70-80 feet so I never fly lower than 120 to also compensate for elevation changes

Definitely not the poles/lines, and if it was a tree it would have had to have been the very top top of one particular one, guess I'm just lucky!

 

[sar104]

The explanations here are getting really over-elaborate. That's a tall tree - taller than 50 ft judging by GE Street View. Is there any reason to reject the simple hypothesis that it simply flew into the tree, crashed to the ground, and ejected the battery on impact?

 

[Squeamish]

The explanations here are getting really over-elaborate. That's a tall tree - taller than 50 ft judging by GE Street View. Is there any reason to reject the simple hypothesis that it simply flew into the tree, crashed to the ground, and ejected the battery on impact?

That's probably what happened, but that tree is definitely not that tall, it would have had to have been the very top point and even then it doesn't look that close by eye. Plus the fact that there is no collision on the video, it just "ends" during level flight, seemed odd.

 

[sar104]

That's probably what happened, but that tree is definitely not that tall, it would have had to have been the very top point and even then it doesn't look that close by eye. Plus the fact that there is no collision on the video, it just "ends" during level flight, seemed odd.

Well - maybe it hit a sufficiently substantial part of the tree, even if it was right at the top, that the battery ejected immediately. The buffer time would then explain the lack of impact in the video. It's otherwise a curious coincidence that an unexplained failure occurred exactly when it reached the highest obstacle in its path. I guess next time just give it a bit more margin on the altitude. I'm always very conservative with that.

 

[Air Ontario]

That's probably what happened, but that tree is definitely not that tall, it would have had to have been the very top point and even then it doesn't look that close by eye. Plus the fact that there is no collision on the video, it just "ends" during level flight, seemed odd.

If the battery moved inflight, it would look that way or the camera wasn;t facing the obstacle

 

[Ramphex]

Just out of curiousity, when you get your aircraft back up and running, can you measure the altitude of the above mentioned tree. Wonder if the random altitude loss people have been reporting was a factor in this crash.

 

[Jussaguy]

Ehh, I was flying at midnight with 3rd party software, this is my fault.

Fair enough.

 

[Multicoptertec]

Irresponsible flying like this, is exactly why "drones" have such a bad rep. Go fly in the desert or wherever. If your AC hit my car, kids, whatever, you would pay dearly.