“Compass error” has claimed the life of yet another poor Sparky…my own, and now my heart is a little bit broken. =(

So I had my little Spark ripped out of my life by the cruel hands of fate.
I was flying from one of my usual spots, situated a little bit outside of the central parts of town. It’s a quite open area not too far from the river docks. On a Sunday like today, not a soul but me was in that area.

The first flight went perfectly smooth without any error-messages at all. I had calibrated the compass before taking-off and I did it on a fairly large grass field with no building within 40-50m or so. I calibrate before every single flight since I’ve had a few fly-aways due to compass-error previously (In spite of having calibrated the compass pre-flight in open areas in all but one instance)

I changed the battery and took off again from the exact same spot and the first half minute everything seemed normal. The home point got recorded and it was perfectly stable in GPS mode. But 30 seconds into the flight, after climbing to around 15m and flying away about 20m in distance, the compass went haywire. It went completely out of control and stopped responding to controls properly, and shortly after, not at all. The screen is full of error messages and the picture is frozen.

  

My spark blasted off straight away from me, and then did a 180 and came straight at me for a while. I almost thought that it was the RTH finally kicking in, but then it did yet another 180 degree turn and shot off into the distance towards the river, and that’s the last I’ve seen of my little buddy =(

Spent the rest of the day looking for him on both sides of the river and in the bay area, but Sparky is gone and most likely on the bottom of the estuary where the river meets the Baltic sea. May Triton and Poseidon take good care of his sparkling soul where he rests in his watery grave =(

I’ve uploaded the flightlog here:

  

http://app.airdata.com/main?share=HNPHlr

  

What would be the best way to proceed from here? What should I send to DJI support? Only the log file, or do they want a copy of the original receipt along as well?
(I do not have DJI care.)

What kind of response is to be expected from DJI to a Spark fly-away incident?
I really hope that I can get a new one, because the money I had I spent on buying the now gone Spark. =/

Thankful for any advice I can get =)

My kindest regards!

Hey man sorry to hear that.
Your procedure should be to sync your flight record to dji cloud
http://www.dji.com/support , contact support and get a case number, then try to make contact with a moderator here to see if they can escalate your case, Care refresh is no good without AC.

Those airdata logs are not really any good and tell very little. If you want a better understanding of what happened you should upload your log to link below and follow instructions come back here and post link.
I will take a look for you.

http://www.phantomhelp.com/LogViewer/Upload/

Geez....I'm almost afraid to fly mine now.  I had a few really good flights where I was in complete control ...maybe that gives you a false sense of security.  These situations where the Spark seems to lose its mind and take off on its own are pretty disturbing, seems like you don't even have a Fighting Chance to get the little guy to safety.  :-(

Cheers...

Todd in Chicago

I'm sorry for your lost, hope everything gets resolved out soon in a fair way

So sorry for your lost.... I had a bad situation 2 weeks ago too. Spark nearly fly away and landed on a root. Fortunately last second it reconnected and RTH. It's all my fault. I did a stupid mod on the cover. I sprayed paint painted it and it's caused some sort of magnetic interferences. It's all well now with a Red Upper shell. It's scared the hell of me.

Lunis, feel bad about this accident. And please contact our support and start a case: http://www.dji.com/support. You can export your flight data and upload to dropbox, then post the link here.

Very sadden to hear your lost. Now I’m worry my Spark. Would really want to know the cause of it.

小仙女的星巴克 Posted at 2017-11-5 12:26
So sorry for your lost.... I had a bad situation 2 weeks ago too. Spark nearly fly away and landed on a root. Fortunately last second it reconnected and RTH. It's all my fault. I did a stupid mod on the cover. I sprayed paint painted it and it's caused some sort of magnetic interferences. It's all well now with a Red Upper shell. It's scared the hell of me.

Thank you =)

I'm glad to hear that you were able to reconnect! ^^

I really wonder about the actual physics behind these errors. What exactly is it that triggers them, and how can one know beforehand? Would be awesome if DJI somehow could fix this with a firmware update.

hallmark007 Posted at 2017-11-5 11:47
Hey man sorry to hear that.
Your procedure should be to sync your flight record to dji cloud
http://www.dji.com/support , contact support and get a case number, then try to make contact with a moderator here to see if they can escalate your case, Care refresh is no good without AC.

Hi there and thanks a lot for giving me advice! I will do as instructed and report back with the uploaded log =)

DJI Elektra Posted at 2017-11-5 18:43
Lunis, feel bad about this accident. And please contact our support and start a case: http://www.dji.com/support. You can export your flight data and upload to dropbox, then post the link here.
https://www.youtube.com/watch?v=ngEHdAlAh9A

Thank you very much for your concern Elektra. I have uploaded the flightlog here. Please let me know if I was supposed to upload it in another manner, and I will do so.

http://www.phantomhelp.com/LogViewer/2R0NYK88X2UCX9TZIZHR/

Kind regards

hallmark007 Posted at 2017-11-5 11:47
Hey man sorry to hear that.
Your procedure should be to sync your flight record to dji cloud
http://www.dji.com/support , contact support and get a case number, then try to make contact with a moderator here to see if they can escalate your case, Care refresh is no good without AC.

Here is the link with the uploaded flightlog Captain

http://www.phantomhelp.com/LogViewer/2R0NYK88X2UCX9TZIZHR/

/Lunis

It looks as though as your little Sparky has landed on atop of the roof of the large building next to ED Bygg Sverige AB Sjötullsgatan 9 602 28 Norrköping Sweden and not in the river as you initialy thought.

Perhaps you could go and ask the kind owner of the building if they can retrieve it for you for a case of your finest Swedish beer

Lunis Posted at 2017-11-6 01:52
Thank you very much for your concern Elektra. I have uploaded the flightlog here. Please let me know if I was supposed to upload it in another manner, and I will do so.

http://www.phantomhelp.com/LogViewer/2R0NYK88X2UCX9TZIZHR/

Please upload your flight log to dropbox instead. Thanks for your support.

My best guess:
Look onto the roof, where your Spark went into auto landing mode.
maybe you are lucky!

Lunis Posted at 2017-11-6 01:54
Here is the link with the uploaded flightlog Captain

http://www.phantomhelp.com/LogViewer/2R0NYK88X2UCX9TZIZHR/

58.59392352        16.20255528       

Hi Lunis
Above is your last coordinates, your AC didn’t make the river but landed in auto land on a roof, it may well have dropped of the roof but taking a close look at it on google earth that roof could well have held it up there, there are a few raised windows on the roof that it could have got caught up on.
Your best bet is to contact the owner to see if you can get access I’m almost certain it is on that property.

Why it happened well that could be many things but you picked up mag interference somewhere along your route whether on the ground or flying over buildings it’s impossible to say.
You picked it up early in your flight and then regained gps that I suppose was the time to land safely.

Good luck I hope you retrieve it I’m certain you can if you can permission.

Lunis Posted at 2017-11-6 01:52
Thank you very much for your concern Elektra. I have uploaded the flightlog here. Please let me know if I was supposed to upload it in another manner, and I will do so.

http://www.phantomhelp.com/LogViewer/2R0NYK88X2UCX9TZIZHR/

this is your roof

hallmark007 Posted at 2017-11-6 03:02
58.59392352        16.20255528       

Hi Lunis

Thank you for your reply and your research.

I too got those coordinates out from the flightlog and actually checked the roof top yesterday twice after being allowed to use the emergency stairs on the side of the building.

But even then I was pretty sure that it wasn't there since I had seen my Spark drift over that very building in a fixed attitude (no more changing direction or spinning around) at around 100m above the ground.

Now, I did not see were it went AFTER flying over the building. But in another flyaway I had, the same thing happened. That time I found my Spark 1.9km from my take-off point and 800m from the point where it initiated autoland according to the log. So that time it drifted for 800m and ended up on a porch. I was lucky to have a lady with dogs witness the last parts of the flight. She said it had been gradually lowering it's altitude while flying straight at  around 20km/h.

This time I'd say that it drifted away from me at around 15km/h, and I could keep up with it for a little while as I tried to reconnect. The speed was greater during it's irratic movments a little bit earlier. But it was motionless in pitch, roll and yaw. I was not able to tell if the attitude of the Spark was completely level or not though. At any rate it seemed to simply be drifting in the prevailing winds like during the previous flyaway.

Still checked the roof in case it had come back to land there somehow., but nothing...=(
It was in RTH mode almost during all of this ordeal, so I kinda hoped that it somehow would find it's way home once contact was broken, but it went in the opposite direction of the home point.

This might sound a bit naive, but I would love to find some sort of tool/device that can show magnetic interference in an area beforehand. Maybe that is impossible, at least for non-constant sources of interference. But with a static source maybe? I don't know if such a device is impossible, or if it already exists and is common knowledge. But if it indeed exists, maybe it can help you know when to ground your Spark and not take off at all?

Lunis Posted at 2017-11-6 05:10
Thank you for your reply and your research.

I too got those coordinates out from the flightlog and actually checked the roof top yesterday twice after being allowed to use the emergency stairs on the side of the building.

You can get an app electro magnetic detector don’t know how it works it’s free in App Store.

Last movement from your Aircraft was auto land and having looked at it again I can see it starting to drift fast as it was trying to land so you could be right so you could be right it could have landed in the river and battery was getting low and you were in Atti mode.
Reason it didn’t RTH was simply no gps so it wasn’t sure where home was it just seemed to hover for a period.
You should contact dji I have seen in some cases warranty given and in worse case scenario 30% discount.
Good luck.

Lunis Posted at 2017-11-6 05:10
Thank you for your reply and your research.

I too got those coordinates out from the flightlog and actually checked the roof top yesterday twice after being allowed to use the emergency stairs on the side of the building.

Electromagnetic Detector EMF by Jorge Gregorio Martin Bello
It may not be what your looking for it’s an app but I think you can get an array of measurement tools to the job!


https://itunes.apple.com/ie/app/ ... mf/id836603095?mt=8

hallmark007 Posted at 2017-11-6 03:02
58.59392352        16.20255528        

Hi Lunis

I don't think its quite as simple as the location at your GPS coordinates.  As of that last coordinate, it was still at a height (IMU Altitude) of 369.8ft and moving at a rate of 8 miles per hour.

In the last 1.4 seconds in the flight record, it had descended 13.4 ft.  That is a descending rate of 9.57 ft per second.  At that rate of descent, it would have taken approximately 38.6 seconds for it to get back to ground level.  

Moving at a rate of 8mph is approximately 11.7 ft per second.  11.7ft/sec times the 38.6 seconds would be a distance traveled of about 452 ft beyond the last GPS coordinate to get back to "ground level".

I doubt the rates quoted above stayed constant.  Also, I'm sure there is a change in elevation from where the aircraft took off from (appears to be from top floor of a parking garage) and wherever its final location ends up being.

Unfortunately, if my calculations are correct and the rates stayed consistent, I'm afraid it is very likely the aircraft had a good chance of ending up in the river.  Just eyeballing the trajectory of the flight makes this seem like a possible conclusion as well.  If it cleared the river, it likely wasn't by much and probably ended up in the contruction/gravel yard just on the other side.


Also, I try not to be judgemental, but in my opinion the flying environment I see from maps doesn't really match up with the environment described in the original post.  This is not an open area at all.  It is very challenging industrial environment with big buildings with metal structures and roofs and mulitiple large towers.  

Kloo Gee Posted at 2017-11-6 06:57
I don't think its quite as simple as the location at your GPS coordinates.  As of that last coordinate, it was still at a height (IMU Altitude) of 369.8ft and moving at a rate of 8 miles per hour.

In the last 1.4 seconds in the flight record, it had descended 13.4 ft.  That is a descending rate of 9.57 ft per second.  At that rate of descent, it would have taken approximately 38.6 seconds for it to get back to ground level.  

I did have a another look at it and came to the same conclusion as yourself,  #19 , it was indeed drifting at increasing speed when it was trying to auto land. One thing I did notice is when it was in Atti mode it still continues to record coordinates much the same way as a phantom switched to Atti mode will do the same, but if you fly without gps coordinates will not be recorded, I’ve seen this in cases where user took of with no gps.
I was thinking back to another case I looked at on Saturday symptoms were exactly the same AC trying to rth went to auto landing but lucky enough this user managed to find his up a tree in the location of his last coordinates.


https://forum.dji.com/forum.php? ... &fromuid=260008

Regarding environment I fully agree sometimes when you contradict or suggest you end up coming out the worse for mentioning it.

小仙女的星巴克 Posted at 2017-11-5 12:26
So sorry for your lost.... I had a bad situation 2 weeks ago too. Spark nearly fly away and landed on a root. Fortunately last second it reconnected and RTH. It's all my fault. I did a stupid mod on the cover. I sprayed paint painted it and it's caused some sort of magnetic interferences. It's all well now with a Red Upper shell. It's scared the hell of me.

Little  Starbucks fairy?  Interesting name! Sry, I get distracted!
;-)
Was it a metallic  spraypaint?

hallmark007 Posted at 2017-11-6 07:35
I did have a another look at it and came to the same conclusion as yourself,  #19 , it was indeed drifting at increasing speed when it was trying to auto land. One thing I did notice is when it was in Atti mode it still continues to record coordinates much the same way as a phantom switched to Atti mode will do the same, but if you fly without gps coordinates will not be recorded, I’ve seen this in cases where user took of with no gps.
I was thinking back to another case I looked at on Saturday symptoms were exactly the same AC trying to rth went to auto landing but lucky enough this user managed to find his up a tree in the location of his last coordinates.

Ok, I did not realize it was over 100m high, when it started to autoland.
If you know the wind direction, my "second best" guess then is, to look into that direction.
Autolanding is about 1/2 m/s, isn't it?
Which gives you raffly a flight time from max . 3+ minutes, about.
Which should be covered by the 22% left battery at that time.
Ok, if some trees in the way, or buildings, ....
But all is now about one thing:
Do you know, which way it drifted away?
And how quick was the wind/drift speed aproxymate?

How does a compass error cause the radio control connection to shut down?  I am not worried about navigation aids faulting out since I can fly RC models (that have no aids).  But the idea that it drops the radio connection then hauls but to a random GPS location sounds odd.  Those should be unrelated faults that would almost never happen together.

Gunship9 Posted at 2017-11-6 07:51
How does a compass error cause the radio control connection to shut down?  I am not worried about navigation aids faulting out since I can fly RC models (that have no aids).  But the idea that it drops the radio connection then hauls but to a random GPS location sounds odd.  Those should be unrelated faults that would almost never happen together.

The compass error won’t cause radio signal to disconnect, his downlink disconnected IMO disconnected because of probable interference in the area he was flying in, you can see in his log it disconnected for some 18 seconds but this is not as a result of compass or gps problems.

S-e-ven Posted at 2017-11-6 07:49
Ok, I did not realize it was over 100m high, when it started to autoland.
If you know the wind direction, my "second best" guess then is, to look into that direction.
Autolanding is about 1/2 m/s, isn't it?

It will continue to drift in the same direction it was heading in Atti mode speed tends to increase with the amount of propulsion which can increase if pilot is pushing up on the elevation stick , which he was although his AC was in Atti mode he was still trying to fly forward when it went into landing mode.

To check how far it may have drifted you could get weather for the area speed of wind and direction of wind and tie this in with the amount of battery power and get a better idea of where it might end up.

hallmark007 Posted at 2017-11-6 07:35
I did have a another look at it and came to the same conclusion as yourself,  #19 , it was indeed drifting at increasing speed when it was trying to auto land. One thing I did notice is when it was in Atti mode it still continues to record coordinates much the same way as a phantom switched to Atti mode will do the same, but if you fly without gps coordinates will not be recorded, I’ve seen this in cases where user took of with no gps.
I was thinking back to another case I looked at on Saturday symptoms were exactly the same AC trying to rth went to auto landing but lucky enough this user managed to find his up a tree in the location of his last coordinates.

Ah, I hadn't seen your follow up post when I typed up my comments.  Glad to see we are on the same page there.  

Regarding the other thread, good job on helping them to find it.  On that one, it looks like they had flown up the side of a hill and so even though the altitude was still showing 200+ft and some speed, it looks like it ran into the trees on the hill near the coordinates it last transmitted.  Great for that person.  

I find it fascinating to look at all these logs from different flights to try to figure out what has happened.  For the flight from this particular thread, it is fairly easy to see looking at the maps that magnetic interference could very likely be a significant contributor to the behavior seen with the Spark.  However, for the 2 different flights in the other thread you mentioned, there isn't anything reasonably identifiable just from looking at the maps that would have contributed to their compass issues.  However, its hard tell with any degree of accuracy from a satellite image likely taken a number of years ago.  

Kloo Gee Posted at 2017-11-6 06:57
I don't think its quite as simple as the location at your GPS coordinates.  As of that last coordinate, it was still at a height (IMU Altitude) of 369.8ft and moving at a rate of 8 miles per hour.

In the last 1.4 seconds in the flight record, it had descended 13.4 ft.  That is a descending rate of 9.57 ft per second.  At that rate of descent, it would have taken approximately 38.6 seconds for it to get back to ground level.  

I don't know how you made it. but that map with the 3d flight path on it is awesome.

I had my first flight this evening since the update to both the firmware and the app and was very apprehensive after reading this and a few other posts.  After I flew around for a while with almost zero problems, I gained more confidence, even though I must say my heart skipped a beat when I saw a "Transmission Lost" message.  I'm curious to take it out again, as one thing I seemed to notice, and it could just be me or the conditions, but hovering off the landing pad was immediately off, meaning drifting 2-3 feet off center. Additionally, when flying and doing a slow rotation to the right, when I stopped the rotation...it seemed to back up a few degrees....meaning...it rotated the other way a few degrees.  Seemed really odd.  I'll have to see if this continues or just some weird "blip".  I do plan on flying again this weekend, and maybe before then if I can swing it.  But tonight has built up my confidence again.

Here is my first AEB photo, nothing to get too excited about but it was fun for me!

Cheers..

Todd in Chicago

Todd in Chicago Posted at 2017-11-6 22:14
I had my first flight this evening since the update to both the firmware and the app and was very apprehensive after reading this and a few other posts.  After I flew around for a while with almost zero problems, I gained more confidence, even though I must say my heart skipped a beat when I saw a "Transmission Lost" message.  I'm curious to take it out again, as one thing I seemed to notice, and it could just be me or the conditions, but hovering off the landing pad was immediately off, meaning drifting 2-3 feet off center. Additionally, when flying and doing a slow rotation to the right, when I stopped the rotation...it seemed to back up a few degrees....meaning...it rotated the other way a few degrees.  Seemed really odd.  I'll have to see if this continues or just some weird "blip".  I do plan on flying again this weekend, and maybe before then if I can swing it.  But tonight has built up my confidence again.

Here is my first AEB photo, nothing to get too excited about but it was fun for me!

When you talk about rotating, is it similar to what is shown/discussed in these 2 threads below?

https://forum.dji.com/thread-108182-1-1.html
https://forum.dji.com/thread-109376-1-1.html

Kloo Gee Posted at 2017-11-7 01:07
When you talk about rotating, is it similar to what is shown/discussed in these 2 threads below?

https://forum.dji.com/thread-108182-1-1.html

Kloo Gee....

Interesting....thanks for those.  I will do a deep dive on the threads after work today, but at least initially, mine is a little different.

Mine is almost like when you fast forward on your DVR and you press play, the DVR backs up a second or two compensating for when you pressed the play button and when the DVR actually started playing.

And by that I mean, I yaw and then stop, and then the drone slowly yaws the opposite direction from which I was going.

I'll see if I can post a quick vid clip tonight, as I didn't really notice it until I watched my video.  One thing I did catch in those 2 links you sent was low light conditions, and this was indeed a low light condition.

I'll post vid clip tonight sir, thanks for your input!

Cheers...

Todd in Chicago

Todd in Chicago Posted at 2017-11-7 05:19
Kloo Gee....

Interesting....thanks for those.  I will do a deep dive on the threads after work today, but at least initially, mine is a little different.

For the videos in the other thread, to me it seems like it is the optical image stabilization re-centering itself after trying to stabilize a rotation/yaw movement.  I'm interested to see the behavior you are seeing.

Kloo Gee Posted at 2017-11-7 01:07
When you talk about rotating, is it similar to what is shown/discussed in these 2 threads below?

https://forum.dji.com/thread-108182-1-1.html

Kloo Gee.....

So, this may not be a *real* problem at all.  I reviewed the whole 8 min clip, and I see that it actually only happened once at the beginning, so it may have just been a momentary lapse of reason, or wind, or my fatso fingers slightly touching the stick.  Here is a link to the edited version of the clip.  I've put 2 or 3 yaw movements in there....and you can see it backtracks on the first one.  The other yaw movements when stopped, appear to be rock solid.  I think I will keep an eye on it and see if it repeats.



Let me know what you think.

Cheers...

Todd in Chicago

S-e-ven Posted at 2017-11-6 07:41
Little  Starbucks fairy?  Interesting name! Sry, I get distracted!
;-)
Was it a metallic  spraypaint?

its not metallic

小仙女的星巴克 Posted at 2017-11-8 06:43
https://photos.app.goo.gl/oOH5zEgE8I9Qs3Gg1

its not metallic

But how could the paint then do this?
"I sprayed painted it and it's caused some sort of magnetic interferences."
Another phenomenon with a DJI Spark?

S-e-ven Posted at 2017-11-8 06:47
But how could the paint then do this?
"I sprayed painted it and it's caused some sort of magnetic interferences."
Another phenomenon with a DJI Spark?

check it out....its alot of errors!

http://app.airdata.com/main?share=jGSJBl

小仙女的星巴克 Posted at 2017-11-8 06:52
check it out....its alot of errors!

http://app.airdata.com/main?share=jGSJBl

Can you upload your logs to the PhantomHelp logviewer?  It is much easier to read and give much more information for people on the forum to be able to try to assist you.

http://www.phantomhelp.com/LogViewer/Upload/

Todd in Chicago Posted at 2017-11-7 19:42
Kloo Gee.....

So, this may not be a *real* problem at all.  I reviewed the whole 8 min clip, and I see that it actually only happened once at the beginning, so it may have just been a momentary lapse of reason, or wind, or my fatso fingers slightly touching the stick.  Here is a link to the edited version of the clip.  I've put 2 or 3 yaw movements in there....and you can see it backtracks on the first one.  The other yaw movements when stopped, appear to be rock solid.  I think I will keep an eye on it and see if it repeats.

To me the behavior seems nearly identical to the behavior reported in the other 2 threads I mentioned and I've seen as well.

小仙女的星巴克 Posted at 2017-11-8 06:52
check it out....its alot of errors!

http://app.airdata.com/main?share=jGSJBl

And all of these errors dissapeared with changing to a new cover, plastic only?
Did you fly that location again?