I was flying perfectly fine, bringing the drone back to land at very slow speed (manual control).

When it was maybe 2 meters from me, it had to go over a small barrier, but there was more than enough clearance on all sides. While flying near the barrier, it suddenly shot up in the air (maybe by 2 meters) as if I had pushed the left stick full up (I didn't), hit a tree, stayed in the tree for a couple of seconds (cutting branches and leaves), and then crashed down on the ground. It carried on spinning on the ground on itself until I shut down the motors manually.

Very heavy damage made to the gimbal and lens, the drone body, one front sensor broken, all propellers shattered, battery is heavily scratched etc.

I am also surprised it didn't shut itself down once on the ground. It was completely out of control.

I have sent a message and the crash log to DJI, but my question is: are they likely to repair (or replace) a drone under warranty due to those types of herratic behaviours??

The flight log and crash log don't seem to show much useful information or specific errors (apart from motor obstructed and speed error, once crashed)...

I am so annoyed that the drone is pretty much destroyed due to its own random behaviour, as opposed to a piloting mistake!

You have sensors on the bottom of your drone that will cause drone to jump up when flying over obstacles , only other thing I can think of is low battery RTH kicked in causing aircraft to try to gain altitude.

How high from the obstacle was the tree branch?  I've had this happen as well when flying over a park table, but it only jumped up around 2 foot or so.  

user error - NEXT please

If the logs show that it wasn't your fault then yes, it's very likely that DJI will replace it under warranty.

If you upload the flight log to PhantomHelp and provide a link to it here, people will be able to see for themselves what happened. Many pilots claim to not be at fault when they really are to blame, so you're likely to encounter some skepticism unless you're willing to post the log to support your description of the event.

Strobing_NYC Posted at 2018-10-8 14:51
user error - NEXT please

Overly harsh but probably correct.

OA kicked in and did what it's supposed to do ?  

Looking at the detailed logs:

* I was bringing the drone to land at very low speed and going down slowly (altitude 8m)
* Suddenly, the drone appeared to initiate a Go Home mode, at 18% battery. Was 35m from home. However, drone didn't climb higher
* I stayed in control of the drone (direction, altitude) and brought it down to 1.6m high, at 15m from home (was planning to land closer than the take off point, so drone was only about 4m from where I wanted to land)
* Suddenly, drone went up to 3.4m high, at high speed, and hit the top of the tree...

Here are the relevant rows from the detailed log:
https://docs.google.com/spreadsh ... dh65QZlXBOw/pubhtml

Surely, the bottom sensor didn't trigger such high fast climb (or shouldn't do), and drone shouldn't start climbing so fast close to its home location/destination, even in RTH mode.
I am assuming the bottom sensor would either stop the drone from going forward further, or just raise its altitude slighly.

BTW, I have been flying RC helicopters and quadcopters (without assistance, stabilisation, GPS etc), for more than 10 years. I had full control of the drone, and it just went on a crazy vertical climb suddenly, on its own. All the sensors and automation created the crash, not me!

Very disheartening having a crash that causes damage, but other than ruining your day, it's never as bad as is seems. DJI drones are very repairable. The first crash I had, the gimbal broke and I had it professionally repaired. Every crash since then I've bought parts on Amazon and completed my own repairs by watching YouTube videos. Anyway, I'm only writing this to let you know that if DJI doesn't help you, it will still be ok.

Send it into DJI care refresh,  will not be your first crash,been there done that.

Hi TedBob, sorry to read on the crash accident, since you have contacted our support, the team will help to start a ticket for the data analysis, we will provide the warranty service if it turns out to be a non-pilot error according to the data analysis result, please keep us updated if you need any assistance during this process.

MrArcher Posted at 2018-10-8 14:41
How high from the obstacle was the tree branch?  I've had this happen as well when flying over a park table, but it only jumped up around 2 foot or so.

According to the flight log, it went up by 1.8m (or almost 6 feet) without any inputs from me. Seems to be quite a climb if it was just triggered by Obstacle Avoidance sensors.

TedBob Posted at 2018-10-8 16:59
Looking at the detailed logs:

* I was bringing the drone to land at very low speed and going down slowly (altitude 8m)

Hi TedBob,

If you like upload your log here > https://www.phantomhelp.com/logviewer/Upload/

cheers
JJB

TedBob Posted at 2018-10-8 21:26
According to the flight log, it went up by 1.8m (or almost 6 feet) without any inputs from me. Seems to be quite a climb if it was just triggered by Obstacle Avoidance sensors.

IIf the forward sensors  detected the barrier then it would automatically rise  to 5m above the obstacle. Maybe this is what happened.

That sounds exactly what happened.  No random behavior at all.  The 5M avoidance bump was cut short when it hit the tree, and seeing the MP doesn't have upward sensors, it just keep blazing upward trimming the tree until it faulted out.  Users need to understand how the avoidance system works, particularly in tight areas.  totally user error.  Hopefully, DJI will be nice enough to offer some help.

That is why I have it set to stop when the sensors detect an object instead of try to go up or around the object.

MrArcher Posted at 2018-10-9 11:00
That sounds exactly what happened.  No random behavior at all.  The 5M avoidance bump was cut short when it hit the tree, and seeing the MP doesn't have upward sensors, it just keep blazing upward trimming the tree until it faulted out.  Users need to understand how the avoidance system works, particularly in tight areas.  totally user error.  Hopefully, DJI will be nice enough to offer some help.

According to the user manual:

* when an obstacle is detected 15m away
* drone should stop and hover first
* then start climbing and stop when 5m above the obstacle

I think 5m is a ridiculous height to avoid an obstacle, but anyway: I was controlling the drone manually over the fence, and it was going over it slowly (with enough clearance) when it suddenly went up at very high velocity. I don't believe that's the expected behaviour explained in the user manual. I surely wouldn't expect the drone to shoot up 5m higher when it's already flying on top of something.

I think what might have happened is that the front sensors didn't detect the fence (I was flying in the evening, low light), and it's only when the drone flew above the fence that the ultrasonic sensors detected something underneath (at maybe a distance of 50cm), and gave the order for the drone to go up by 5m at very high speed...

Strobing_NYC Posted at 2018-10-8 14:51
user error - NEXT please

this is a pretty shetty response, if you have nothing worth saying just dont even bother with your low life trolling comment.  why the need to be such a j3rk when OP is clearly not asking for it.

DJI Mindy Posted at 2018-10-8 18:22
Hi TedBob, sorry to read on the crash accident, since you have contacted our support, the team will help to start a ticket for the data analysis, we will provide the warranty service if it turns out to be a non-pilot error according to the data analysis result, please keep us updated if you need any assistance during this process.

Hi Mindy,

I raised a support ticket via the Android App, and also uploaded the flight log on the form when completing it. However, I didn't get any email notification or anything else after submission, just a message on the app that ticket had been submitted. Is that expected? How long does it typically take for DJI to give an initial response?
Thanks

TedBob Posted at 2018-10-9 12:11
According to the user manual:

* when an obstacle is detected 15m away

I agree it was bottom sensors that may have caused it to jump, but can’t be certain, it might become clearer if you follow jjb link and put up your flight log.
This of course could have been a malfunction, so pay no heed to some posters.

hallmark007 Posted at 2018-10-9 12:43
I agree it was bottom sensors that may have caused it to jump, but can’t be certain, it might become clearer if you follow jjb link and put up your flight log.

I have already put a link to a spreadsheet with all the relevant detailed logs when the issue happened (see several posts above). There is no mention of "obstacle detected" or similar warning/messages, when the drone went up vertically.

It would be quite worrying to know that the drone could shoot up 5m higher when the bottom sensor suddenly detect something underneath.

TedBob Posted at 2018-10-9 12:46
I have already put a link to a spreadsheet with all the relevant detailed logs when the issue happened (see several posts above). There is no mention of "obstacle detected" or similar warning/messages, when the drone went up vertically.

It would be quite worrying to know that the drone could shoot up 5m higher when the bottom sensor suddenly detect something underneath.

There is lots of info missing from that, most prominent is your sheet shows no controller input.

See link to logs online: https://www.phantomhelp.com/logviewer/7D1V379H2PUL2WT8NMIZ/

Issue started at around 10m 34s. Sudden jump in altitude, without any user input or warning message, followed by a crash on the ground...

TedBob Posted at 2018-10-9 12:52
See link to logs online: https://www.phantomhelp.com/logviewer/7D1V379H2PUL2WT8NMIZ/

Issue started at around 10m 34s. Sudden jump in altitude, without any user input or warning message, followed by a crash on the ground...

Your log shows aircraft RTH, you decided while it was in RTH to lower aircraft for some reason, aircraft was still some 50ft from home when it encountered obstacle causing it to rise up and hit the underside of the tree.
I’m not sure what your reasoning for continually lowering aircraft on its RTH , but looking at the trees on google maps it looks like a tight enough spot to try to get through with obstacle sensors on, you said your aircraft jumped 5 metres, but log shows around or just less than 5 ft.
I think aircraft acted correctly in this situation. But dji May see it differently, so you could get a warranty claim.
I would say in low light trying to steer through the trees with OA on had its risks attached, might have been better to let Aircraft RTH on it’s own.

hallmark007 Posted at 2018-10-9 13:36
Your log shows aircraft RTH, you decided while it was in RTH to lower aircraft for some reason, aircraft was still some 50ft from home when it encountered obstacle causing it to rise up and hit the underside of the tree.
I’m not sure what your reasoning for continually lowering aircraft on its RTH , but looking at the trees on google maps it looks like a tight enough spot to try to get through with obstacle sensors on, you said your aircraft jumped 5 metres, but log shows around or just less than 5 ft.
I think aircraft acted correctly in this situation. But dji May see it differently, so you could get a warranty claim.

I decided to lower the drone for the good reason of... going under the tree!

There was enough clearance for the drone to go under the tree, but then it jumped up into the tree without inputs from me. It went up by around 1.8m, which was enough for it to hit the tree, break the propellers and then fall on the ground. It continued spinning and bouncing while on the ground, until I stopped the motors manually (which doesn't sound right either).

There was no "user error": drone hit the tree on its own, I was flying it perfectly fine and very slow.

If the expected behaviour is for that drone to suddenly jump 5m higher because of OA, then it should never been flown under any object or tunnel or tree or roof etc.

TedBob Posted at 2018-10-9 14:02
I decided to lower the drone for the good reason of... going under the tree!

There was enough clearance for the drone to go under the tree, but then it jumped up into the tree without inputs from me. It went up by around 1.8m, which was enough for it to hit the tree, break the propellers and then fall on the ground. It continued spinning and bouncing while on the ground, until I stopped the motors manually (which doesn't sound right either).

Again I’m only telling you what could have happened, you again mention 5 m , it didn’t jump 5m it jumped just less than 5 ft, when your drone sees obstacle underneath that’s what it will do, I’m not saying it will jump 5ft as yours did, but flying so close to an obstacle will cause aircraft to rise.

Hallmark is correct in what he is telling you. Hes not going after you but just stating the facts based off what your flight record shows happened and nothing more.

Hopefully DJI will take the facts given and decide to still replace the drone though.

hallmark007 Posted at 2018-10-9 14:20
Again I’m only telling you what could have happened, you again mention 5 m , it didn’t jump 5m it jumped just less than 5 ft, when your drone sees obstacle underneath that’s what it will do, I’m not saying it will jump 5ft as yours did, but flying so close to an obstacle will cause aircraft to rise.

Yes, documentation mentions that, with OA, drone may go up by 5m to avoid obstacle.

In my case, drone didn't go up 5m, as it crashed into the tree that was around 1.8m higher than the drone. Drone couldn't go up by 5m.
It probably tried to continue going up once in the tree, but the propellers broke...
I understand that the drone is expected to rise when detecting an obstacle, but I would not expect by 5m (as per the documentation would seem to indicate), particularly when its altitude is less than 5m already... If it was at 40m high, it would be different.

I think DJI's logic is flawed, if a drone doubles its altitude because some obstacle is detected at a low altitude. This would create many crashes when people fly inside tunnels, under bridges etc.

TedBob Posted at 2018-10-9 14:45
Yes, documentation mentions that, with OA, drone may go up by 5m to avoid obstacle.

In my case, drone didn't go up 5m, as it crashed into the tree that was around 1.8m higher than the drone. Drone couldn't go up by 5m.

Flying under bridges in tunnels is more likely to cause drone to loose gps and crash, we have seen this many times here on this forum.
I wouldn’t give up on dji yet, let them examine your .dat log and see what they come up with.
Good Luck...

Boffin Posted at 2018-10-8 16:34
OA kicked in and did what it's supposed to do ?

What is OA?  please.  

sorry,  i had a brain fart...   i remember now oa.....    doy!

TedBob Posted at 2018-10-9 14:02
I decided to lower the drone for the good reason of... going under the tree!

There was enough clearance for the drone to go under the tree, but then it jumped up into the tree without inputs from me. It went up by around 1.8m, which was enough for it to hit the tree, break the propellers and then fall on the ground. It continued spinning and bouncing while on the ground, until I stopped the motors manually (which doesn't sound right either).

this is an interesting and enlightening discussion.   

I have seen my drone "shoot up" on a RTH when I lowered it.   However I did not understand why.   Now I have a much better understanding of the OA system.   

I crashed my 1st mavic (pilot error) and so I know the shame and pain it causes.   

Although this looks like an avoidable incident, it is not pilot error as far as "control of an aircraft" but it sounds like pilot inexperience of the way the software for this particular bird works.

thanks for posting and thanks for "most" of the comments here.

Dronin Posted at 2018-10-10 12:54
this is an interesting and enlightening discussion.   

I have seen my drone "shoot up" on a RTH when I lowered it.   However I did not understand why.   Now I have a much better understanding of the OA system.   

Every crash gains a wealth of experience. As long as we have deep enough pockets for repairs, the gained knowledge is priceless.  I force myself to do a written accident report with a lessons learned summery for each crash.

Dronin Posted at 2018-10-10 12:54
this is an interesting and enlightening discussion.   

I have seen my drone "shoot up" on a RTH when I lowered it.   However I did not understand why.   Now I have a much better understanding of the OA system.   

Your exactly right, it’s not pilot error, and experience will always help, it’s a learning curve for all of us.

Fly Safe.

Dronin Posted at 2018-10-10 12:54
Although this looks like an avoidable incident, it is not pilot error as far as "control of an aircraft" but it sounds like pilot inexperience of the way the software for this particular bird works.  

In term of my inexperience in the way this done behaves: yes, that's correct, I don't know yet all the random ways this drone is going to behave, or what I would call software bugs.

However, I am not willing to spend a significant amount of money in repairs to find out those! I don't mind taking responsibility when I do mistakes, but I am quite annoyed by random behaviours creating a lot of damage (gimbal broken, lens damaged, middle shell damaged, front sensors not calibrating etc).

I also have a Mavic Air, and it has never gone up by several meters (at high speed) to avoid obstacles, particularly when in control. It usually stops or slows down. I have done 100 flights with it, and haven't crashed once, or even broken or damaged a propeller.

I have only bought the Mavic Air and Pro in the last 2 months, and those are my first GPS "clever" drones. Before that, I have flown many RC helicopters and quadcopters without built-in assistance. I did have some crashes, but now I can confidently fly those unassisted.

Based on this Mavic Pro crash, I am now doubtful on whether all that assistance and automation really helps!

I have raised a ticket via the app on Monday, and I haven't heard anything from DJI yet (not even an email confirmation). I can't find an online form where warranty claims can be made and flight logs uploaded.

TedBob Posted at 2018-10-10 21:42
In term of my inexperience in the way this done behaves: yes, that's correct, I don't know yet all the random ways this drone is going to behave, or what I would call software bugs.

However, I am not willing to spend a significant amount of money in repairs to find out those! I don't mind taking responsibility when I do mistakes, but I am quite annoyed by random behaviours creating a lot of damage (gimbal broken, lens damaged, middle shell damaged, front sensors not calibrating etc).

Hi Ted, sorry for my late response, with your Forum account, I only found one case number which was created today, may I know how did you raise the support ticket before?

TedBob Posted at 2018-10-9 14:02
I decided to lower the drone for the good reason of... going under the tree!

There was enough clearance for the drone to go under the tree, but then it jumped up into the tree without inputs from me. It went up by around 1.8m, which was enough for it to hit the tree, break the propellers and then fall on the ground. It continued spinning and bouncing while on the ground, until I stopped the motors manually (which doesn't sound right either).

Hi Ted,

Had a look at your flight too.

RTH by low batt, you decided not to cancel this mode but manually fly home using "down stick" to control and lower height. Why not CX the RTH, than much easier to fly at low height and control that height.

At the end AC raises some height, because you did not gave down stick input, so the RTH mode wants to get her fly higher. See my charts.
AC raises height rapidly (thats normal), so after down-stick again it takes some time to react and lower height, too late as you hit the trees.

So IMO not at OA fly up but a human 'error'.

cheers
JJB

JJBspark Posted at 2018-10-12 01:37
Hi Ted,

Had a look at your flight too.

I think your analysis is correct:

RTH engaged, but I was busy controlling the drone under a tree, and had full control. It didn't try to go up while I was controlling it (I was not "fighting" the drone at all).

However, is the drone supposed to go up at very high speed so close to its "home location". I understand that would be the case if it was far away...

DJI Mindy Posted at 2018-10-12 00:12
Hi Ted, sorry for my late response, with your Forum account, I only found one case number which was created today, may I know how did you raise the support ticket before?

Hi Mindy,

I raised a ticket on the Android DJI Go 4 on Monday soon after the crash. I completed the various fields and uploaded the flight log. Then got a confirmation message, but nothing else (no email confirmation).

I raised another ticket yesterday directly on the website.

Thanks

TedBob Posted at 2018-10-12 03:04
Hi Mindy,

I raised a ticket on the Android DJI Go 4 on Monday soon after the crash. I completed the various fields and uploaded the flight log. Then got a confirmation message, but nothing else (no email confirmation).

The app never seems work. I found that out when I used it to open a ticket for my crashed Mavic a month or so ago. After a day of getting no confirmation email I ended up doing it again online and it set up right away. Not sure why the app doesn't take it. (on android at least)