Hi All,

I've made this thread hoping to get feedback from the community and also raising awareness to what I think to be some hardware issue with the Spark.

About 2 months ago I was flying my Spark for a few short flights when on my 4th flight it suddenly fell from the sky shortly after take off ~00:32 seconds according to the flight record and about 20m height.

Now a few details about the incident:
- On the fourth flight, swapped the battery as previous one was getting low; this had a full charge as flight record also verifies
- Spark was in sport mode
- Not much movement before the suspected power loss, I had only flew it straight up and did a 360 view before heading forwards

I've done a quick search on google and it seems to confirm at least a few other cases of Sparks reporting sudden loss of power. Right now I have a case open with DJI support and depressingly I'm in the process of disputing their assessment of the crash. According to their assessment the Spark was in Sport mode and probably hit a tree as sensors are off..... I have video from the unit and also in the cache from DJI Go app which shows the logging suddenly cutting off which I want to present as counter evidence. Unfortunately I was not aware of the DJI Assistant app available to download more detailed logs myself before sending the unit in for repair.

Has anyone else experienced or seen this issue elsewhere? This has been a frustrating experience to have to defend what happened. Luckily I did buy the care package before using the Spark, but no way do I want to spend more money to fix something I believe was a hardware fault...

Thanks for reading my rant

Sorry to hear about the crash, could you please provide me with your case number so I can look into your case?

DJI Diana Posted at 2017-11-28 19:01
Sorry to hear about the crash, could you please provide me with your case number so I can look into your case?

Thanks for the quick response, my case number is: CAS-1090881-R9S3Q2

I also have some email thread with the support team ongoing..

Evergreen Posted at 2017-11-28 19:10
Thanks for the quick response, my case number is: CAS-1090881-R9S3Q2

I also have some email thread with the support team ongoing..

Hello, Sir, per the data analysis, the drone hit the tree during the flight.
1. Unit was flying in SPORTS mode and was responsive to RC commands.
2. At t=64s, relative height =34.6m, unit hit on a tree in the front and fell under SPORTS mode.
3. User commanded full pitch forward before impact, no obstacle avoidance on sports mode.

You mentioned you have the video, could you please provide to us for further analysis? Thank you.

DJI Mindy Posted at 2017-11-30 05:15
Hello, Sir, per the data analysis, the drone hit the tree during the flight.
1. Unit was flying in SPORTS mode and was responsive to RC commands.
2. At t=64s, relative height =34.6m, unit hit on a tree in the front and fell under SPORTS mode.

Without the video, how can you tell it hit a tree?

djiuser_Kf4iPA3 Posted at 2017-11-30 05:59
Without the video, how can you tell it hit a tree?

The flightdata contains a lot of information.

What information in the flight data can tell it hit a tree?

djiuser_Kf4iPA3 Posted at 2017-11-30 06:36
What information in the flight data can tell it hit a tree?

Well you can tell the pitch roll Yaw you can tell what stick movements Pilot is making you can tell if drone is upside down, you can use terrain maps to show position of aircraft , and a heck of a lot more besides.

That is not to say dji engineers cannot make a mistake and I’m sure if OP has the evidence to show by video or logs to prove this then dji should reverse their decision.

I still don't see how they can tell it hit a tree without anything visual.
They may be able to say it probably hit something if it had momentum and it abruptly stopped or was deflected, though. A sudden loss of power would have had the drone continue its horizontal path simply by inertia.

djiuser_Kf4iPA3 Posted at 2017-11-30 07:43
I still don't see how they can tell it hit a tree without anything visual.
They may be able to say it probably hit something if it had momentum and it abruptly stopped or was deflected, though. A sudden loss of power would have had the drone continue its horizontal path simply by inertia.

There is also recording on internal sd card

hallmark007 Posted at 2017-11-30 08:47
There is also recording on internal sd card

of what? the video? but DJI doesn't have his video, from whatever source.

djiuser_Kf4iPA3 Posted at 2017-11-30 07:43
I still don't see how they can tell it hit a tree without anything visual.
They may be able to say it probably hit something if it had momentum and it abruptly stopped or was deflected, though. A sudden loss of power would have had the drone continue its horizontal path simply by inertia.

I invite you to send me your last flightlog without any other information and I'll report back what information that file alone is containing ;-) And in the case we are discussing here DJI has had the entire blackbox data for analysis in addition to that.

djiuser_Kf4iPA3 Posted at 2017-11-30 08:49
of what? the video? but DJI doesn't have his video, from whatever source.

Internal sd card records all fpv and he said he sent his drone in for assessment.

Wachtberger:
Well, I'm not the one who crashed, so my files would not contain anything too interesting or challenging for you...

hallmark007 Posted at 2017-11-30 08:53
Internal sd card records all fpv and he said he sent his drone in for assessment.

Well, in this very thread DJI has asked him to provide them with the video, so apparently they don't have access to it. Looks like he turned his drone in but kept the SD card.

djiuser_Kf4iPA3 Posted at 2017-11-30 08:53
Well, I'm not the one who crashed, so my files would not contain anything too interesting or challenging for you...

It would demonstrate you the level of detail of the information contained in that single file alone. And DJI has had a multitude of more data for analysis. But you do not seem to believe that.

djiuser_Kf4iPA3 Posted at 2017-11-30 08:58
Well, in this very thread DJI has asked him to provide them with the video, so apparently they don't have access to it. Looks like he turned his drone in but kept the SD card.

There are 2 sad cards one internal not removable unless yo take spark apart and one that is removable that’s the one he has, so if his is different than their one it’s this that dji Mindy just asked for.

Wachtberger Posted at 2017-11-30 09:01
It would demonstrate you the level of detail of the information contained in that single file alone. And DJI has had a multitude of more data for analysis. But you do not seem to believe that.

And how do you figure that?
Actually i said that from flight data they can ascertain with some likelihood that the drone had hit something, even if only from the sudden loss of horizontal momentum.
How they can definitively tell it hit a tree though... that's a different matter.

djiuser_Kf4iPA3 Posted at 2017-11-30 09:10
And how do you figure that?
Actually i said that from flight data they can ascertain with some likelihood that the drone had hit something, even if only from the sudden loss of horizontal momentum.
How they can definitively tell it hit a tree though... that's a different matter.

Sorry, but I give up on this discussion. It's leading nowhere apparently. Hallmark007 has explained it all to you but you simply do not want to believe him. For me there is no point in arguing about it.

hallmark007 Posted at 2017-11-30 09:05
There are 2 sad cards one internal not removable unless yo take spark apart and one that is removable that’s the one he has, so if his is different than their one it’s this that dji Mindy just asked for.

There's an internal memory (I doubt it's in SD card form), but why would video footage be kept on it? it's not meant for that.
And DJI's debugging equipment  would most likely have access to the internal memory without having to take the drone apart anyway e.g. via USB debugging, like most devices have today.

besides, what is the argument about? DJI has specifically asked him to provide them the video from the incident, so apparently they don't have it and don't have access to it.

djiuser_Kf4iPA3 Posted at 2017-11-30 09:15
There's an internal memory (I doubt it's in SD card form), but why would video footage be kept on it? it's not meant for that.
And DJI's debugging equipment  would most likely have access to the internal memory without having to take the drone apart anyway e.g. via USB debugging, like most devices have today.

I think you will find that dji Mindy asked for his video to try help him , and to see if his video differs from engineers findings, but as you have probably read he says power cut out on him so I’m not to sure how this can help him , but from dji mindys post she seems to be saying that dji have the proof of the end of the flight.
So maybe it’s best to wait to see what the outcome will be .

Wow, woke up and didn't expect to see this many responses and debate in my thread... thanks all for helping bring some attention with this discussion.

To update everyone, I've continued to talk to DJI support through email and did provide them the last video from my SD card and also that was cached from the App for review. Unfortunately they just came back with this response:
"Upon checking and attaching the video you have provided. I have found out that your case is now being quoted of the repair service fee."

I'm following up to ask for their reasoning based on the video and hope they can explain why. Also, as mentioned in my OP, I was never aware of the ability to download additional detailed logs from the Spark before sending it out... I've asked the support team to provide those logs they are using and explain how they came to their conclusions.. no update on that request so far.

I'm trying to figure out how to upload the video here as well for the community to take a look, also the flight record(?) that I can see in the app to help correlate my evidence if that helps

Decided to upload it to youtube for easy viewing, link below:

Evergreen Posted at 2017-11-30 11:16
Decided to upload it to youtube for easy viewing, link below:

Thank you very much for sharing this video. Assuming that this is all video that you could retrieve (without cutting anything), I can now fully understand your questions and consider them to be very reasonable. But why have you waited two months before raising this issue if I may ask?

Wachtberger Posted at 2017-11-30 13:33
Thank you very much for sharing this video. Assuming that this is all video that you could retrieve (without cutting anything), I can now fully understand your questions and consider them to be very reasonable. But why have you waited two months before raising this issue if I may ask?

Unfortunately I was out of town when the incident actually happened, then some shipping mistakes on my end caused the delay until now

I raised the request in my ticket to get a hold of the flight logs used for assessment, and hopefully they can be correlated to the video and refute the crash.

Open question:
Would the black box logging continue even if there was sudden power loss on the device?
My best guess is MAYBE it lost power and started free falling forwards and hence indicating it collided front first? But then if this were the case I would have expected the data to also show the gradual free falling momentum

Evergreen Posted at 2017-11-30 13:49
Unfortunately I was out of town when the incident actually happened, then some shipping mistakes on my end caused the delay until now

I raised the request in my ticket to get a hold of the flight logs used for assessment, and hopefully they can be correlated to the video and refute the crash.

I must admit that I do not have sufficient experience to answer to all this. But that any data recording could continue after the loss of power I doubt very much (unless there would be a second (internal) battery that I have never heard or read about, but I don't believe there is one).

Wachtberger Posted at 2017-11-30 13:57
I must admit that I do not have sufficient experience to answer to all this. But that any data recording could continue after the loss of power I doubt very much (unless there would be a second (internal) battery that I have never heard or read about, but I don't believe there is one).

It may take some seconds for video to be fully transferred to sd card but black box may have recorded those extra seconds before crash/or power loss, so once aircraft crashed this halted any more video being downloaded to sd card. It’s just a hunch but usually the dat file will record more data than say phantomhelp or airdata.

hallmark007 Posted at 2017-11-30 14:59
It may take some seconds for video to be fully transferred to sd card but black box may have recorded those extra seconds before crash/or power loss, so once aircraft crashed this halted any more video being downloaded to sd card. It’s just a hunch but usually the dat file will record more data than say phantomhelp or airdata.

Thank you very much for this clarification/explanation. Sounds very logic.

Update:
My response from DJI is that they are looking at the DJI flight logs that I've synced as a basis for their conclusion.

I've attached the last log which I believe is the correct one to this post. It's in TXT format but I can't find an application that can make it legible. Does anyone have any recommendation to make it readable for analysis?

Evergreen Posted at 2017-12-6 13:42
Update:
My response from DJI is that they are looking at the DJI flight logs that I've synced as a basis for their conclusion.

Sigh attachments don't seem to work, uploaded to my google drive instead:

https://drive.google.com/open?id ... SpHxa6O2quNiWxdxU0Q

Same link

http://www.phantomhelp.com/logviewer/upload/

Evergreen Posted at 2017-12-6 13:47
Sigh attachments don't seem to work, uploaded to my google drive instead:

https://drive.google.com/open?id=1VQPWebMwLZAq9SpHxa6O2quNiWxdxU0Q

Okay figured it out and here is a parsed log from this site:

http://www.phantomhelp.com/LogViewer/3LQRZEYDYWKHY4J3JT7V/

What looks strange to me is the VPS Altitude seemed to be working for a while and went to zero at: 0m 22.1s

However the other columns appear to continue tracking to my video before it cuts off. I don't understand how a conclusion to hitting a tree or any obstacle can be reached?

Hopefully you get everything squared away.

So this will be my final update on this thread so there's closure for anyone interested...

After my continued requests for their team to review the data plus my video (linked earlier in thread), I was eventually arranged a phone call with one of their account specialists to discuss the situation. It was nice that the case was escalated for me to speak to someone live, but that it wasn't very productive as you can imagine; I just received some reassurances that the team will review once again for me. I had asked why my case was any different from a few others I had found here on the forums, but no comment except its case by case review.

SPARK LOSE BATTERY in AIR???
https://forum.dji.com/forum.php? ... D516%26typeid%3D516
DJI Spark fell from the sky
https://forum.dji.com/forum.php? ... D516%26typeid%3D516
Spark lost power with full charge, fell smashing to pieces
https://forum.dji.com/thread-105996-1-1.html

Shortly after the call I received an email update noting that "the final determination is that my unit hit a tree and that caused the crash, which is considered non warranty".

Here's the final email update below that I received for this matter. Their graph was not helpful nor conclusive IMO, the data appears to cut off and is consistent of my theory of sudden power loss. Also, the GPS coordinates show an area with a lot of trees, which is consistent with my video, except IMO I was already well in the clear before anything happened...

FLY 043
1. unit was flying in SPORTS mode and was responsive to RC commands
2. at t=64s, relative height =34.6m, unit hit on a tree in the front and fell under SPORTS mode.
3. user commanded full pitch forward before impact, no obstacle avoidance on sports mode
GPS:43.8033874 15.9639825




Anyway, since the language was quite clear, I decided not to bother anymore and just pay for the repair cost. While I'm happy with the responsiveness of the support team, their refusal to acknowledge the evidence I presented and continued defense that sport mode==crash has been extremely disappointing.

I'll continue to enjoy the Spark as it's a great device, and hopefully this won't happen again for myself nor others!

The last alert on your remote will indicate if the battery or power failed during flight.

I have long single beeps and pauses, NO ONE HERE WILL CONFIRM this because they are confirming a glicth or malfunction which means they have to replace the product so instead they sent emails suggesting 30% off because they found nothing in their records indicating the cause to be either the pilot or the aircraft., TOTAL BS, i spent time with this before because i got a staff of 65 employees on a amusement park excited about the idea of opening the first drone area in a amusement park so people can fly small drones with goggles, this plan is now totally trashed as i was humiliated bny DJIs handling of this case..a calm perfect day with 60+% battery 45m from me and 100 up high i get a signal from my remote (according to others) indicating the drone is taking over anding mode or power failure.
When this happened with the GoPro Karma they faced the music, here we are learning its the case with the Spark, how many more users has this happened to is the question.

So you all know, in France and Europe the manfacture has to prove the end user caused an incident to be delt with under warranty and if they have flight records yet pretending not to know what the alerts from the remote mean it is clear intent to abandond their responibility. ..get ready for class action.!

djiuser_Kf4iPA3 Posted at 2017-11-30 05:59
Without the video, how can you tell it hit a tree?

I would like to see the video.  I’m inclined to agree with it’s the assessment.    Proof is all he needs… Waiting…