Hey Guys,

I've been reading more and more instances of Sparks getting disconnected mid flight. I wish I saw these more before I purchased mine!

Anyone have thoughts on this?



This happened shortly after I updated the firmware. As you can see, I'm not very far from my location, and the drone was in RTH mode. It was out of sight, though just barely. I also had ample battery life and signal when my Spark suddenly disconnected from my remote. Immediately, my wife and I searched the area it was last said to be, also using the Find my Drone's last coordinates, but while searching the area and surroundings for two hours, it couldn't be found.

When I contacted DJI about it, they offered 30% off a new Spark, but I feel like given the circumstances (and that this appears to be an issue with the Spark drones) that they should offer a full refund or replacement. I've gone in circles with DJI Customer Service between all platforms (phone, chat, email) and all the while being told a manager will call me, but they have not, only responding through email.

The reason's they've listed after several emails (initially telling me that they could not determine what caused this) don't seem to pertain to the situation, and when I mentioned that I feel like this might need legal action, the response was as follows:



"There are many possibilities which will lead to the interruption of the flight record. For example, if there was obstacles or interference blocking the remote signal, disconnection may happen; the record also will be interrupted if the USB cable between your mobile device and the remote controller loose; or it also has something to do with the mobile device itself, and there are many other reasons which I will not list here. But all of these doesn’t indicate the aircraft itself malfunctioned. Due to the interruption of the record, the cause was not recorded, but it can be verified from the record that there was no malfunction found before the record was cut off.

We are taking your case seriously. And we could answer your inquiries for this case. But the policy is the policy and the fact is the fact. We could only offer the related solution based on the result of analysis and our after-sales policy.

We respect your decision if you want to legal action, although we don't expect it."

I wasn't using a cable at all, and it was in a relatively clear area very close to me when the drone disconnected. I don't see how this is acceptable, especially with an issue that is becoming more and more common. I'm curious to hear more thoughts on the matter.

The altitude is not good for such a distance. Also while you post such cases, kindly tell us where you were flying the drone. for instances over building top, trees, forests , near DJI classified zones,beach ,industrial areas etc

The drone went to that altitude during the RTH flight. It was in an area with a few trees, and it was not in any restricted classified areas at all. Can I ask why the altitude was not good for the distance while I was using the remote, which is supposed to support further distances?

To add, here are the specs for the Spark with Remote distances:

        2.412 - 2.462 GHz (unobstructed, free of interference)
FCC: 1.2 mi (2 km); CE: 0.3 mi (500 m)
SRRC: 0.3 mi (500 m); MIC: 0.3 mi (500 m)
5.745 - 5.825 GHz (unobstructed, free of interference)
FCC: 1.2 mi (2 km); CE: 0.18 mi (300 m)
SRRC: 0.7 mi (1.2 km); MIC: -

This is taken right from the DJI Spark specs page. It was well within those ranges.

The Spark is a good drone, but everyone does strange things and a lot of fly-aways caused by pilot errors:
1. disconnection due to unstable apps (use version what shown on dji site release note (not >4.1.22)
2. agressive flying in sport mode caused yaw errors/gps lost and went in atti mode.
3. fly further than LOS.
4. fly with too much wind
5. use un-supported devices
6. fly in an area with much interference

1. I was using the most updated version of the app.
2. I was not flying in sport mode at all, and the drone was in RTH on it's way back to me when it disconnected
3. It was not further than LOS, as you can see on my screenshot of my report when the drone was lost.
4. There wasn't too much wind that day; there were no warnings of such and there wasn't even a breeze in the air.
5. I had double checked that my devices were supported.

I can really understand that drones can be lost because of pilot errors, but I was incredibly careful while flying to make sure that my phone was compatible, everything was up to date, and I was within a small distance of the drone. I didn't fly very far with it at all. And finally, it was on the RTH flight, so shouldn't it have been programmed to return to me even if the signal was lost? If so, why didn't this happen? And if not, why isn't that the protocol for the drone? I paid $600 for the drone bundle and for this to happen is absurd.

Jos A Posted at 2018-5-24 06:33
The Spark is a good drone, but everyone does strange things and a lot of fly-aways caused by pilot errors:
1. disconnection due to unstable apps (use version what shown on dji site release note (not >4.1.22)
2. agressive flying in sport mode caused yaw errors/gps lost and went in atti mode.

You are absolutely right! It's almost always the pilot's error. Spark is a very good drone. I have also lost the signal about 2 times but in both cases the RTH has worked as indicated in the manual and it has brought me back my Spark.

Pond Posted at 2018-5-24 06:42
1. I was using the most updated version of the app.
2. I was not flying in sport mode at all, and the drone was in RTH on it's way back to me when it disconnected
3. It was not further than LOS, as you can see on my screenshot of my report when the drone was lost.

The best solution would be to upload the flight data. So you can see all the details of the flight, the messages of the app and the possible cause of the loss of your Spark.

Sorry to hear you lost your drone.

It would be helpful for someone to assist you if you could provide more information about the location and flight data.

Taking legal action - You have to proof that you have done everything as per manufactures guidelines. In your case you haven’t (you lost sight of the AC at some point as per your statement above. This also breaks FAA’s UAS rule).

Hope it helps!

Ram-UK Posted at 2018-5-24 07:18
Sorry to hear you lost your drone.

It would be helpful for someone to assist you if you could provide more information about the location and flight data.

Hello!

Here is a copy of the flight log:

https://app.airdata.com/main?share=jUqiYD

I had to upload it to airdata because the forums will not allow for me to attach a .txt file for some reason. As you can see, the signal strength was great throughout.

The location was near a river and a farm. There are a few surrounding trees, but nothing significant; a lot of open areas.

Still risky to fly between 5 -20m in that distances.
Load the log also up here, please:
http://www.phantomhelp.com/logviewer/upload/

Since the home point was recorded and RTH succesfully initiated, the return should have finished fully automated, so I feel anything happening after a succesful initiation of RTH is kind of product fault to me. Though there was a high wind warning which may or may not contributed to this - but probably this will be enough for DJI to reject your claim...

It could have simply crashed into a tree. At this distance in RTH mode, obstacle avoidance was disabled, and I guess there are trees as high as 100 ft, aren't there? You should have set the RTH altitude to a higher value, to be on the safe side.

Hello,

The default RTH att. is 30 meters.
When it is returning to home the obstacle avoidance is off, so it could fly into a tree.
I saw on you flying spot shadows of very big trees.
I would go back to the coordinates where your lost your signal with the drone and walk back to the home point.
Look carefully into the trees of that height of 30 meters (binoculars).
I think it will be there hanging in a tree.
The spark is very light and small, so on 30 meter height it is already difficult to see.

I hope you will find it back.

Regards.

Thank you all for your suggestions and help; you have been far more helpful than DJI Support.

That being said, the Spark was above the tree line before it was lost. I had been recording at the time and had gotten this before the Spark lost a signal, and as you can see on it's flight home that it was above the tree line so I don't think it was lost in a tree.

https://www.dropbox.com/s/335oe9vsgcu94pk/2018_05_13_13_57_52.mp4?dl=0

Pond Posted at 2018-5-24 09:29
Thank you all for your suggestions and help; you have been far more helpful than DJI Support.

That being said, the Spark was above the tree line before it was lost. I had been recording at the time and had gotten this before the Spark lost a signal, and as you can see on it's flight home that it was above the tree line so I don't think it was lost in a tree.

I think you're right, you should not lose the connection at this point.
You were well above the trees.
Were you at this white house?

Regards.

It's heart breaking to lose your AC, especially if you feel like you did nothing wrong.
I hope you find it or get a replacement.

Pond Posted at 2018-5-24 09:29
Thank you all for your suggestions and help; you have been far more helpful than DJI Support.

That being said, the Spark was above the tree line before it was lost. I had been recording at the time and had gotten this before the Spark lost a signal, and as you can see on it's flight home that it was above the tree line so I don't think it was lost in a tree.

At the end of the video (where I assume the connection was lost) one can see the propellers. I'm not sure of this but I thought Spark can film it's own propellers only in Sport mode (I might be terribly wrong on this). You say it wasn't in sport mode, and if the fight record also says it wasn't, maybe there's something into it so you can press the matter with DJI.

Antipaxi Posted at 2018-5-24 10:31
At the end of the video (where I assume the connection was lost) one can see the propellers. I'm not sure of this but I thought Spark can film it's own propellers only in Sport mode (I might be terribly wrong on this). You say it wasn't in sport mode, and if the fight record also says it wasn't, maybe there's something into it so you can press the matter with DJI.

When RTH is activated the drone automatically enters in sport mode and returns to home. That's why the obstacle avoidance sensor is dezactivated

Antipaxi Posted at 2018-5-24 10:31
At the end of the video (where I assume the connection was lost) one can see the propellers. I'm not sure of this but I thought Spark can film it's own propellers only in Sport mode (I might be terribly wrong on this). You say it wasn't in sport mode, and if the fight record also says it wasn't, maybe there's something into it so you can press the matter with DJI.

Which version do you use of the Go 4 app ?
There was already a problem rapport with the RTH.
See : "Can not set RTH point to controller anymore in 4.2.16 (IOS)"

I had a few times that the RTH was not updated with start of the flight.

Ones I had a quickshot flight  (a dronie) where the RTH was changed to that last position.
On low battery the RTH kicked in and did fly back to the dronie position instead of the original start position. I could not remember if the RTH was update on the start on that incident.

Now, I carefully check the RTH before I go on flying, do a little up, down, left, right and turning.

The only thing I want to say: not everything on the Spark is perfect.
A big part of this site/forum is about not perfect.

Pond Posted at 2018-5-24 09:29
Thank you all for your suggestions and help; you have been far more helpful than DJI Support.

That being said, the Spark was above the tree line before it was lost. I had been recording at the time and had gotten this before the Spark lost a signal, and as you can see on it's flight home that it was above the tree line so I don't think it was lost in a tree.

Yes, it’s above the trees. The only thing that comes to mind is that the battery could have fallen off. It has happened before with other Sparks, and it would explain the sudden loss of connection.

Pond, I feel sorry for your loss, I understand that it must be frustrating. I managed your case number and the data analysis result is on below.
1. The aircraft was working in GPS and Sport mode and responded to RC command well.
2. At t=04:49 h=30.4m d=112.5m, the flight record ended.
The flight record ended without any sign of abnormality, we're not able to verify what happened afterwards. Considering this special situation, we provided you with the corresponding offer based on after-sales policy. If you have more details and data about the accident, we hope you can send them back to us, we will to forward to the engineers to review again.
Regarding your case, it has been escalated, will have the manager contact you as soon as possible.
Sorry again for the difficulties.

Or could be an attack of a bird also you use the latest app(mavic air stuff in it), on dji site under spark the last release is 4.1.22!!!
I said it many times , don't upgrade always the newest version without knowing it if it is for the spark!

Pond Posted at 2018-5-24 06:26
The drone went to that altitude during the RTH flight. It was in an area with a few trees, and it was not in any restricted classified areas at all. Can I ask why the altitude was not good for the distance while I was using the remote, which is supposed to support further distances?

In theory, the SPARK must fly at very low altitudes. But practically due to geographical reasons , there are so many interference  caused by ground earth while using Wi Fi  and therefore it is proven that in such cases you must fly these drones at a height > 50 meters to avoid such errors and fly away mishaps especially using the SPORTS mode

Yup I agree the drone is a good little drone I’ve had mine for 4 days now I flown it at high altitude over houses and cables it lost connection and it automatically went into RTH and come back to me which I was very impressed with. I hope I never have a a issue where it won’t come back. Feel sorry for you pal losing it hope you get the outcome you need from dji

I've been using the DJI GO 4 app on Android, not IOS. It's version 4.2.16

DJI Thor: what other information can I send you to further help this situation? I've also been told several times over the past week that a manager would specifically call me (from online chat support and through the phone support) and I have not received any calls from DJI at all.

Hey, it just so happens I'm a meteorologist.  I decided to run a trajectory model showing the winds on that date/time (13/1800 UTC).  I am thinking if for whatever reason the drone went into ATTI mode, the drift would have taken it along a path roughly 342° at 10 mph (pictured below), allowing it to go maybe a mile before it gives out.  If it were me, I would search along the axis pictured below.  The ground rises about 60 feet with the treetops even higher, so I'm suspecting there's a possibility it crashed as the terrain rose up to meet it.  This would be along the treeline north of the farm up the road from you (pictured) or, less likely, among the buildings at that farm.




Of course it just as well might have had the battery pack come loose and gone down in the river, or zoomed off toward another heading, etc, but I'm just throwing out one possibility.  Barring a power problem, obviously GPS wasn't working or it would have made it home, and I'm guessing OPTI mode would have put it in the river with battery depletion, but I'm not sure it would kick in that high.  So ATTI mode might have brought it along the track shown, unless uneven motor power causes it to move off the track.  Does anyone know if they stay fairly fixed when embedded in the wind?

On that Airdata UAV site I see they do have a wind tab.  It's clear from all the data on that page that it comes comes from historical observations at the nearest federal station, probably Lewiston about 15 miles away, and is just one sample of data.  It seems to agree but I'd be more inclined to believe the trajectory plot in the image above as it comes from a fine-grained mesoscale model that accounts for terrain.  Also it appears the Airdata UAV has a wind calculation as a premium service.  It's not clear to me where they're getting the data, but if it's calculated from the GPS and drift inside the Spark, that would potentially be the most valuable data of all in determining drift location.

Anyway at least it's something to work from as a "likely" scenario.

Sorry about the double images... I clicked "edit" and saw no double image.  Can't figure out what the deal is with that.

rolypolyman Posted at 2018-5-25 07:13
Hey, it just so happens I'm a meteorologist.  I decided to run a trajectory model showing the winds on that date/time (13/1800 UTC).  I am thinking if for whatever reason the drone went into ATTI mode, the drift would have taken it along a path roughly 342° at 10 mph (pictured below), allowing it to go maybe a mile before it gives out.  If it were me, I would search along the axis pictured below.  The ground rises about 60 feet with the treetops even higher, so I'm suspecting there's a possibility it crashed as the terrain rose up to meet it.  This would be along the treeline north of the farm up the road from you (pictured) or, less likely, among the buildings at that farm.
[view_image]

Nice analysis, I really enjoyed it.

Thanks.  Here is the KMZ file in case anyone wants to see it.  There's an east-west offset I can't do anything about, so you have to factor it in.

https://www57.zippyshare.com/v/cg8lrvuJ/file.html

We should form a little NTSB bureau for lost Sparks.   We could have an office in Denver somewhere, and a little pad out back where we can take drones for a spin.

Pond Posted at 2018-5-25 05:45
I've been using the DJI GO 4 app on Android, not IOS. It's version 4.2.16

DJI Thor: what other information can I send you to further help this situation? I've also been told several times over the past week that a manager would specifically call me (from online chat support and through the phone support) and I have not received any calls from DJI at all.

Pond, due to the flight record interrupted suddenly, we cannot speculate what happened later, so it is inconclusive. If you have more data or proof to show it is the malfunction, please share with us for double check, thanks in advance!

Jos A Posted at 2018-5-24 06:33
The Spark is a good drone, but everyone does strange things and a lot of fly-aways caused by pilot errors:
1. disconnection due to unstable apps (use version what shown on dji site release note (not >4.1.22)
2. agressive flying in sport mode caused yaw errors/gps lost and went in atti mode.

To name just a few:
https://forum.dji.com/thread-149701-1-1.html
https://forum.dji.com/thread-146808-1-1.html
https://forum.dji.com/thread-149695-1-1.html
https://forum.dji.com/thread-103792-1-1.html
https://forum.dji.com/thread-145379-1-1.html
https://forum.dji.com/thread-142793-1-1.html
Enough of the good advice ... !!!

"It's not like you've been flying too far, because if the connection is broken, it should come back 10 seconds later. (or do not I have to be? I'm wrong?)
- Strong winds can not cause a compass error, nor can GPS errors or interruption cause any interruption (just cause it to blow something)
- No aggressive flight can be made to it just for what it was designed (if this was a bad idea why it was "sport" mode)
- The system automatically switches between the channels to be optimized. In order to be a fault, all channels and both frequencies must be fully occupied, which is unusual even in the city, but outside is impossible in nature. Of course the interference the reason for everything ...

When do you understand that there is a relationship between these cases? A sudden mistake in the Spark navigation system, which even surprised the experienced driver, is surprised by the less experienced ...

Believe me, the pilot knows (he feels) if he's lost his wire due to his fault. These cases are NOT pilot errors (this is also supported by logs ....)

You do not need to post comments that blame the operator and do not realize the real reasons ...

S.J Posted at 2018-5-24 22:42
In theory, the SPARK must fly at very low altitudes. But practically due to geographical reasons , there are so many interference  caused by ground earth while using Wi Fi  and therefore it is proven that in such cases you must fly these drones at a height > 50 meters to avoid such errors and fly away mishaps especially using the SPORTS mode

We agree with a very strong electromagnetic space to disrupt all systems (GPS receiver, compass, air pressure sensor, RC link). Eg. high voltage cable, radio tower. etc

But in most fly away cases this is what happens.
- GPS error
- compass error
- Yaw error
- connection is interrupted

This can not be caused by strong interference.
Interference can cause disturbances in control because it is the WI-FI channel used by everyone. In this case (after 10 seconds) the RTH and the drone will return home.

It is impossible for all systems to be disturbed outside the city, some interference.

So: If you fly away far, the connection may be interrupted for other WI-FI or other radio broadcasts, but the compass and GPS still need to work !!!!!!!!!
Then Spark must come home, not fly away.

If this is not the case, this is the drone bug.

Jakab Gipsz Posted at 2018-5-26 15:03
We agree with a very strong electromagnetic space to disrupt all systems (GPS receiver, compass, air pressure sensor, RC link). Eg. high voltage cable, radio tower. etc

But in most fly away cases this is what happens.

But you are still missing the fact that low altitude is the cause for these issues !!

S.J Posted at 2018-5-27 06:08
But you are still missing the fact that low altitude is the cause for these issues !!

There is nothing missing, I have written everything in the past few days, logically and factually.

1. It is also apparent to the price and size that it was destined for amateurs
2. Amateurs are not qualified pilots
3. designers have built up sensors and protection (visual sensors, RTH ...)
4. In general, this also works, so you can flee it far too far, if it's okay to come home.
5. It was also prepared for the disturbances: channel, and even frequency, if any, and if there is no connection then come back (RTH).
6. It would have to stop other disturbances (in a confined space or in a crowded room with electrical equipment, the compass does not bother me unless I take it to the top of my computer's metal enclosure.
7. Only strong near-electromagnetic radiation or large metal may be disturbed, but there is also GPS
8. the GPS signal can be shielded only by high-rise partially or enclosed space.
9 outdoors in flight the maximum can only fluctuate in the GPS signal, but it is impossible to missing it.
10. it is impossible for all systems to suffer interference at the same time, because they operate on a completely different frequency and other principle, with a different signal (WiFi, GPS, compass)
A large metal object can disturb the compass near, but not GPS or Wifi. A stronger WIFI can interfere with the connection, but it does not interfere with the compass or the GPS ...

But just ... be in the soup: for some mistake that could not happen, the control crashes.

Therefore it is not legitimate to blame the pilot and not to blame legitimate unknown interference.
These are just excuses not to compensate the owner.

Jakab Gipsz: Thank you for the defense!

DJI had tried to call me once, at a time that I was unable to get to my phone in time. The number that called me sent me back to the DJI main support number, and they had left a message saying they would try again within 24 hours; it has now been 51 hours since then. Still highly unimpressed with the DJI support.

On another note, thank you so much, rolypolyman! That info was very interesting and informative. We ended up going towards that area to search and had no success, but it was worth a shot! Anything to try and get my drone back.

DJI Susan Posted at 2018-5-25 23:59
Pond, due to the flight record interrupted suddenly, we cannot speculate what happened later, so it is inconclusive. If you have more data or proof to show it is the malfunction, please share with us for double check, thanks in advance!

What further information can I give you? What are you looking for to show that this is clearly an error? The RTH was on and functioning, you have the video and the flight record to prove that it cut out while on RTH while 100ft in the air, and the video shows that it clearly was above the tree level in my area.

Pond Posted at 2018-5-30 10:19
What further information can I give you? What are you looking for to show that this is clearly an error? The RTH was on and functioning, you have the video and the flight record to prove that it cut out while on RTH while 100ft in the air, and the video shows that it clearly was above the tree level in my area.

The designated team will keep follow up for you to see if there are any better solutions or suggestions. Thanks for your patience.

Please describe the event briefly in:
https://forum.dji.com/thread-151287-1-1.html
The goal is to find the similarities and to infer the cause of the mistake.

Thanks. G. J.