Hi all - feeling a little overwhelmed after reading so many posts/flight logs on spark fly away situations.  There are so many variables I'm not exactly sure what to include. I wish I was posting cool aerial footage.

The main items:
- Got my spark about a month ago.  17 flights for 65k ft travel in about 2.5hrs.  I feel like I have a pretty good handle on the drone.
- Yesterday took my drone out in a new location, near Lake Michigan.  Re-calibrated compass in the middle of a beach as I've seen suggested you should do when substantially changing locations.  I'm on vacation until Tuesday.
- Had 2 successful flights (1 battery each) the same day before the one where the drone was lost.
- Last flight was over some sand dunes in low to moderate wind.  Had the following errors (on the DJI Go app screen) just before I never saw it again:
     - Signal lost. Aircraft is returning home.
     - Magnetic Filed Interference. Exit P-GPS Mode;
     - Weak GPS signal. Positioning accuracy may be compromised. Please fly with caution.
-  Note that I was flying pretty far (~6,500ft) but it still had 67% battery left and 18 GPS satellites so from my view there shouldn't have been any issue getting home.
- Walked toward the drone and the controller never reconnected. I was using the controller obviously connected via OTG to a pixel 2.
- Spent hours looking at the coordinates indicated in "Find My Drone" and the last point where I had video.
- Canvassed about 1/2 mile of the dunes (starting at the take off location) with a few people.  No trees, but lots of tall grass.
- I've experienced plenty of lost signal errors at various distances.  Every time the drone properly starts to return to home and shortly after I can regain control.


Additional Details:
- https://www.phantomhelp.com/LogViewer/940D6X8Z4FU34NSPQCHH/#
- DJI support request # (1537563)




Any suggestions?  According to the flight log the last thing I see is that the Spark was in Go Home mode.  My only thoughts are that it switched back to ATTI and continued drifting inland or that it somehow ran out of battery along the RTH path.  I find it hard to believe it ran out of battery though as it had 67% charge and it wasn't so far that it couldn't get home at the default 10m/s.

My worse nightmare, you lost connection to the RC and it went into ATTI mode, so no return home, just drifting away like a dumb thing.  It was very far away and out of sight. You would probably be better off using a GPS tracker or another type of tracker if you are planning to send your drone 2km away.

Hi thanks for the reply. I actually just read through your thread about the incident at your friends house. Learned about the csv viewer program that I'll try tomorrow.

As for my own last flight, wasn't really planning to take it out so far, but I've taken it out quite a ways on a few occasions now and didn't realize that so many people are having issues much closer. The whole time I had plenty of gps satellites and I had solid video until the errors happened.

At this point I just want to know how the drone malfunctioned and if possible find out how it behaved after the errors and maybe get a better idea of where it is.

Researching the issue indicates this seems to happen. Lot.

the CsvView is very interesting, but I used the .dat file from the black box, not the flight log. It showed that my Spark came close to disconnect from the remote twice and then disconnected for a few seconds on a third time. It is very informative. It is sad that you lost GPS and RC connection at the same time. If you would have just lost RC, the Spark would have come home on its own. I think when that happen, it is not the pilot's fault because there is nothing you could have done. No RC means it should return home and it did not, it just lost GPS. I would contact DJI to get a replacement as this is not normal. It had plenty of satellites, plenty of battery so it should have returned home. Was it in Sport mode?

It's interesting. Your spark was coming back home fast and in sport mode, then switched to ATTI mode due to yawn error. It seams that ATTI mode often switch when in sport mode especially.

It went 300 feet + in few seconds in ATTI mode.

At some point you were in sport mode and got this message: High wind velocity. Fly with caution and land in a safe place ASAP. Maybe the high winds with the sport mode combined made it unstable, not sure. I am new to all this and hopefully someone with good experience in reading those log files will be able to tell you exactly what happened and why.

Hi ajmckay,

Sorry for your loss, too bad that this happens a lot with a Spark.

After 4m38s your spark lost RC control, so it went into a RTH.
During this RTH yaw errors and Speed errors, at the end magnetic Field Interference, so exit position mode and into ATTI.
At the end again no connection with the RC for 63 secs, when connected again Spark had GPS so again in RTH mode.
Flightrecords end there, so no connection, Spark should RTH but i think that it enters ATTI again.
In ATTI with no connection Spark starts to land, question is where!?

My guess is the Orange marking where it has landed.


cheers
JJB

Daniella3d Posted at 12-30 22:46
It's interesting. Your spark was coming back home fast and in sport mode, then switched to ATTI mode due to yawn error. It seams that ATTI mode often switch when in sport mode especially.

It went 300 feet + in few seconds in ATTI mode.

Oh maybe I misunderstood on csvview. You had your drone so you could access the "black box". In my case all I have to work with is the log file. Well hopefully something can still be figured out the amount of detail in the"verbose" csv download from phantomhelp is pretty good it seems - I just haven't figured what means what yet lol.

Ive read some other accounts (I think) that referenced sport mode and wind as a possible contributor to some errors. The problem is sport mode is so much fun! I really only use it as a decent altitude though because it doesn't stop very quickly so it could be easy to crash into something.  

As for the RTH speed, i thought the speed for the spark was 10m/s or 22mph.  Either way there was lots of battery left.

JJBspark Posted at 12-31 08:13
Hi ajmckay,

Sorry for your loss, too bad that this happens a lot with a Spark.
[Image]

Thanks a ton JJB!  This is great info.

In particular I'm interested in your estimation that the drone went into ATTI again after the disconnect.  Since the last status was RTH I assumed that it would fly back home. But if the protocol is to shift to ATTI and land then your estimation is the little orange marker with the black diamond shape in it?  Given the trajectory from the 63 second blackout it would make sense that if it started to land immediately it would follow along that path.

The timing of this is ideal as well because it's supposed to snow tonight.  It's only a few hundred ft off that small road so I'll go check it out!

Thanks again all of you!

Flying more than a 1000 meters from the controller is taking a lot of risks, at this distance the only hope is that RTH will work.

ajmckay Posted at 12-31 10:43
Thanks a ton JJB!  This is great info.

In particular I'm interested in your estimation that the drone went into ATTI again after the disconnect.  Since the last status was RTH I assumed that it would fly back home. But if the protocol is to shift to ATTI and land then your estimation is the little orange marker with the black diamond shape in it?  Given the trajectory from the 63 second blackout it would make sense that if it started to land immediately it would follow along that path.

It is 100% sure that it went back into ATTI after the second disconnect because if it would have remained in GPS mode, it would have returned home. It had enough battery to do so. Since it did not return home, it probably went again into ATTI and then was lost.

And you can probably calculate it's distance in ATTI mode because on the fist disconnect it moved about 300 feet. Another thing that is interesting is that when it regained GPS, it slowed a lot probably to return home in the opposite direction and the distance to home was again going closer to home, but then another disconnect at the end and nothing with what? 67% battery. I wonder if it actually landed in ATTI and without connection to the remote.

very much wonder how it is able to control itself enough to land? when it is barely able to stay in the air! Anyone ever saw an Autoland in ATTI mode??? I would love to see that.

ssylca44 Posted at 12-31 12:57
Flying more than a 1000 meters from the controller is taking a lot of risks, at this distance the only hope is that RTH will work.

Thanks for the comment!  So far I've taken it to the 1,000m mark several times and if it drops signal the RTH has always performed.  The weird part is sometimes I lose connection at 500ft and other times I can go for a long time without losing connection. In this case I was probably emboldened some given it was a long stretch of beach. Didn't fly over trees much and there aren't any power lines or sources of interference. I had 18gps satellites at the time of disconnect so why wouldn't it RTH properly?  

The thing that's kinda freaking me out after doing some reading on the forums is the fact that many of us have to "hope" that the rth actually works consistently on these drones.

Daniella3d Posted at 12-31 14:02
It is 100% sure that it went back into ATTI after the second disconnect because if it would have remained in GPS mode, it would have returned home. It had enough battery to do so. Since it did not return home, it probably went again into ATTI and then was lost.

And you can probably calculate it's distance in ATTI mode because on the fist disconnect it moved about 300 feet. Another thing that is interesting is that when it regained GPS, it slowed a lot probably to return home in the opposite direction and the distance to home was again going closer to home, but then another disconnect at the end and nothing with what? 67% battery. I wonder if it actually landed in ATTI and without connection to the remote.

Good point on it switching to ATTI...

So I went out again and spent about 1.5hrs searching approximately the area in blue. The yellow indicates where I searched yesterday. Unfortunately I didn't find the spark.  Also this was probably my last real opportunity to search since it's supposed to snow 5" tonight and I leave tomorrow.  I do have my name and phone # on it.

When I searched today I searched under the assumption of no obstical avoidance. I assumed that the spark started to land at the point where the last known coordinates were - while drifting along the same line created between the location before and after the first disconnect. This was JJB's extrapolation and it seems to make sense. It either crashed or got lucky and drifted between the trees.  Are there any other options?  

Could the spark switch to ATTI mode at the point of disconnect and basically maintained altitude at about 100ft and just drifted with the wind until the battery got so low that a landing was forced?

ajmckay Posted at 12-31 17:13
Good point on it switching to ATTI...

So I went out again and spent about 1.5hrs searching approximately the area in blue. The yellow indicates where I searched yesterday. Unfortunately I didn't find the spark.  Also this was probably my last real opportunity to search since it's supposed to snow 5" tonight and I leave tomorrow.  I do have my name and phone # on it.

Here's the image I forgot to attach

Daniella3d Posted at 12-31 14:02
It is 100% sure that it went back into ATTI after the second disconnect because if it would have remained in GPS mode, it would have returned home. It had enough battery to do so. Since it did not return home, it probably went again into ATTI and then was lost.

And you can probably calculate it's distance in ATTI mode because on the fist disconnect it moved about 300 feet. Another thing that is interesting is that when it regained GPS, it slowed a lot probably to return home in the opposite direction and the distance to home was again going closer to home, but then another disconnect at the end and nothing with what? 67% battery. I wonder if it actually landed in ATTI and without connection to the remote.

Here you go.  This guy goes into ATTI mode, selects RTH which won't work with ATTI mode, and then pushes the sticks in the wrong direction repeatedly.  Wrong direction because he make the tilt and drift worse.  He should have moved the stick left to correct for the drone sliding away to the right unstabilized.

As soon as remote link is lost, the Spark starts landing where it is (can't RTH in ATTI mode) along with winds and drift.  DJI doesn't let Sparks fly by themselves across the nation's airspace.  They either RTH when in GPS-P stabilized, or land where they are when in ATTI mode.  This one recovers GPS lock just before touchdown.  It then RTH.

Gunship9 Posted at 12-31 17:39
Here you go.  This guy goes into ATTI mode, selects RTH which won't work with ATTI mode, and then pushes the sticks in the wrong direction repeatedly.  Wrong direction because he make the tilt and drift worse.  He should have moved the stick left to correct for the drone sliding away to the right unstabilized.

As soon as remote link is lost, the Spark starts landing where it is (can't RTH in ATTI mode) along with winds and drift.  DJI doesn't let Sparks fly by themselves across the nation's airspace.  They either RTH when in GPS-P stabilized, or land where they are when in ATTI mode.  This one recovers GPS lock just before touchdown.  It then RTH.

Many thanks for the video - though I feel even more confused now lol.

So in my case I had no video, just a bunch of jumbled pixels on my screen. I didn't know which way the spark was aiming and the drone initiated RTH. Shortly after it went into ATTI and drifted for a little over a minute. I dont think I had any control because the drone was at the same altitude when it reconnected for a brief moment about 300ft from where it lost connection.

So obviously I need to somehow train myself on how to better handle ATTI, but I still don't understand what the drone was supposed to be doing.  If it's supposed to land when connection is lost why didn't it land while it was disconnected for over a minute?

Why did it go back to atti mode when the last known status was "go home".  

Finally is it possible that the drone drifted for some time wasting battery, then at some point regained GPS (like the video where it hovered for  bit) and started RTH with insufficient battery?

Thank you everyone

Daniella3d Posted at 2018-12-31 14:02
It is 100% sure that it went back into ATTI after the second disconnect because if it would have remained in GPS mode, it would have returned home. It had enough battery to do so. Since it did not return home, it probably went again into ATTI and then was lost.

And you can probably calculate it's distance in ATTI mode because on the fist disconnect it moved about 300 feet. Another thing that is interesting is that when it regained GPS, it slowed a lot probably to return home in the opposite direction and the distance to home was again going closer to home, but then another disconnect at the end and nothing with what? 67% battery. I wonder if it actually landed in ATTI and without connection to the remote.

yes, i have seen an drone landed in ATTI. there is a video on this forumm but how to find that one. (lol, its in a post before this!)

In ATTI mode and out of control (so no RC inputs gets to the drone) the ATTI drone fly`s level pitch and roll and just drifts with the wind. After 3 seconds in ATTI with no connection it starts to land.
Landing speed Spark 2 m/s i think, so drifing and landing at the same time. Flying at 30 meters height means 15 seconds before touching the ground (or tree or hill or whatever obstacle is in the flight path)

cheers
JJB

ajmckay Posted at 2018-12-31 10:43
Thanks a ton JJB!  This is great info.

In particular I'm interested in your estimation that the drone went into ATTI again after the disconnect.  Since the last status was RTH I assumed that it would fly back home. But if the protocol is to shift to ATTI and land then your estimation is the little orange marker with the black diamond shape in it?  Given the trajectory from the 63 second blackout it would make sense that if it started to land immediately it would follow along that path.

hi,

Yes it was flying RTH but as there is no more data in the flightlog i guess its disconnected again.
If the RTH was succesfull, you had seen in coming home....

So my guess it disconnected again and into ATTI or it flew to home for some time and went into ATTI again due the errors. If the last happend than the search area become huge, more bad luck and hard to find your drone back.

cheers
JJB

ajmckay Posted at 2018-12-31 22:32
Many thanks for the video - though I feel even more confused now lol.

So in my case I had no video, just a bunch of jumbled pixels on my screen. I didn't know which way the spark was aiming and the drone initiated RTH. Shortly after it went into ATTI and drifted for a little over a minute. I dont think I had any control because the drone was at the same altitude when it reconnected for a brief moment about 300ft from where it lost connection.

Hi,

Confusing and questions....

Flying in  ATTI out of sight is not that easy, if connected get some heigth if needed, no stick input bu yaw to turn to home direction, than fwd stick. See data in the app and the attitude indicator/compass.
Check distance decreases! if possible just land!

Spark without errors goes into RTH when RC<>Spark connection fails.
Spark with errors (depending wich errors) goes into ATTI.

In your flight Spark into ATTI because of errors. [5m2.1s], but still connected to the RC.
At 5m7.7s connection lost, this for 63 seconds.
After 63 seconds logging is there so conenction with the RC>device.
In the 63 seconds the yaw errors became less, so GPS data used again. But no connection so Spark went into RTH mode. When connected again to the RC logging starts again writing GoHome....

After 1.5s flying in GoHome logging stops, no connection Spark>RC>device.

If GoGome was succesful (not connected to RC but connected to GPS and no errors) you would have seen Spark landing at the homepoint.

IMO errors popup again, so in ATTI and drifting away, after 3 secs autolanding.

Hope it will clarify your questions.

cheers
JJB

JJBspark Posted at 1-1 04:57
Hi,

Confusing and questions....

A lot of these intermittent connections with the remote are because the panicked pilot is now lifting up the controller to take a closer look at the screen and its info.  Lifting the remote up to your face tilts the antennas so their radiation pattern is straight up instead of at the distant drone.  It gives you intermittent connection to the distant drone.   Distant drones on the edge of connection don't like you moving around the antenna radiation pattern.

Ask me how I discovered that issue,...though in my case it was just the video feed I kept losing.
   

ajmckay Posted at 2018-12-31 22:32
Many thanks for the video - though I feel even more confused now lol.

So in my case I had no video, just a bunch of jumbled pixels on my screen. I didn't know which way the spark was aiming and the drone initiated RTH. Shortly after it went into ATTI and drifted for a little over a minute. I dont think I had any control because the drone was at the same altitude when it reconnected for a brief moment about 300ft from where it lost connection.

By "when connection is lost", do you mean video connection or remote control of the drone?  There are two connections.  One is the remote's command link out to the drone.  The other is the Spark's flight data and video return link back to the remote controller.

People say they lost link but it likely is just the video feed they lost since that is gobs and gobs of data to make a video and it is the first to cut out when the signal starts to go bad.  The lower left ADI with the drone icon and the rest of the drone data is second to go.  Last the big heavy transmitter in the remote controller will cut out.  The drone owner will still have positive control of the drone with the video feed screen blank.

My old Spektrum 7 channel radio controller for model helicopters can maintain radio control at over five miles due to its dual channel transmitter that hops to avoid signal noise on the channels.  But it only sends out control commands (one link) to my models and has no return data from the model.  I fly the models within a couple hundred yards of myself.

Gunship9 Posted at 1-1 09:03
A lot of these intermittent connections with the remote are because the panicked pilot is now lifting up the controller to take a closer look at the screen and its info.  Lifting the remote up to your face tilts the antennas so their radiation pattern is straight up instead of at the distant drone.  It gives you intermittent connection to the distant drone.   Distant drones on the edge of connection don't like you moving around the antenna radiation pattern.

Ask me how I discovered that issue,...though in my case it was just the video feed I kept losing.

Oke, never thought about that possibility.
And yes, a solid connection at 2000 meters away at 30 meters height is not 100% sure.

JJBspark Posted at 1-1 04:20
yes, i have seen an drone landed in ATTI. there is a video on this forumm but how to find that one. (lol, its in a post before this!)

In ATTI mode and out of control (so no RC inputs gets to the drone) the ATTI drone fly`s level pitch and roll and just drifts with the wind. After 3 seconds in ATTI with no connection it starts to land.

Hello! Can you point the source of information about landing in ATTI mode during aborted connection with RC? Where did you see that drone lands after 3 s of absence of connection?

Lysak2003 Posted at 1-1 12:19
Hello! Can you point the source of information about landing in ATTI mode during aborted connection with RC? Where did you see that drone lands after 3 s of absence of connection?

Hi Lysak,

Good question! Its not documented in DJI manuals, just read it somewhere on the www.
Its a safety thing i guess implemenetd by DJI software, just minimize the risk of a drone drifing in ATTI away uncontrollable with 99% battery,....  ;-)  can fly for many minutes this way.

In the video above it can be seen. (post #14)

cheers
JJB

Gunship9 Posted at 1-1 09:03
A lot of these intermittent connections with the remote are because the panicked pilot is now lifting up the controller to take a closer look at the screen and its info.  Lifting the remote up to your face tilts the antennas so their radiation pattern is straight up instead of at the distant drone.  It gives you intermittent connection to the distant drone.   Distant drones on the edge of connection don't like you moving around the antenna radiation pattern.

Ask me how I discovered that issue,...though in my case it was just the video feed I kept losing.

It's very thoughtful remark. We all should remember that. Thank you.

So basically if it disconnect from the RC and goes into ATTI, it just land anywhere and anyhow after 3 seconds? it just lower its altitude until it hits something, trees, hills, ground, anything.

It's funny that it disconnected from the remote for more than 3 seconds and was still at the same altitude, but more than 300 away from the previous position, so drifting.

Maybe I got this wrong but here there seem to be a near one minute gap when it lost RC connection and was in ATTI mode then regained connection and satellite and was doing a return home again, but during that near one minute ATTI and disconnected, it did not lose any altitude, so what about that 3 seconds safe land thing? it's not there.

5m 3.4s - ATTI - 98.4ft altitude and 6,320.2ft from home point

Then nothing for more than one minute and than it briefly reconnect:

6m 7.6s - Go Home - 99.4ft altitude - 6,671.9ft from home point

So in that minute it was out of connection and in ATTI mode, it never lowered its altitude, so no landing after 3 seconds. Also we can see it went another 341ft away from home.


Landing after 3 seconds??? it does not seem like it in this case. It is obvious it was just drifting away at about the same altitude.

Gunship9 Posted at 1-1 09:17
By "when connection is lost", do you mean video connection or remote control of the drone?  There are two connections.  One is the remote's command link out to the drone.  The other is the Spark's flight data and video return link back to the remote controller.

People say they lost link but it likely is just the video feed they lost since that is gobs and gobs of data to make a video and it is the first to cut out when the signal starts to go bad.  The lower left ADI with the drone icon and the rest of the drone data is second to go.  Last the big heavy transmitter in the remote controller will cut out.  The drone owner will still have positive control of the drone with the video feed screen blank.

Happy New Year. So let me say that you all have been very helpful in answering my questions and helping me piece this together.
JJB - I really hope that it didn't start RTH and then ATTI again and land somewhere completely random...  It never gave me any indication that it reestablished connection with the RC.  That is to say that the light never turned green again after I noticed that it had turned red (I don't remember exactly when that was though).  Also I wanted to specifically say thanks for posting those "FRAP" screens.  That looks like a pretty cool program - did you write it?

Your timeline helps - I've spent some time today while in the car thinking about things and trying to better analyze.  My initial conclusion is that I encountered a perfect storm of crap.  Signal lost, not in sight, spark had errors, went into ATTI mode, GPS lost, and also the RC disconnected.  How much worse could it have possibly gotten? Then there's the unknown.  How long did the spark stay in RTH? Did it immediately return to ATTI and drift some more or maybe land (more on that in a minute) or did it fly for some time in RTH and then error out again and drift / land in some new completely unknown trajectory? I canvassed a large portion of the beach near the home point with 3 other people and didn't see anything.

Gunship - I can see that being the case (losing signal due to a reaction that changes antenna position).  I assume then that the antenna emits a signal pattern similar to a wifi router?  Fortunately I use a 3D printed phone mount that holds the phone above the controller and at the perfect angle to my face.  So I'm pretty sure that the controller antenna were at roughly 90 degrees to horizontal the whole time. Still a great point.  When I say "Connection Lost" I mean the video.  I lost video at the point where the spark entered RTH the first time. I don't know if I moved it using the controller in ATTI mode. I don't even remember it even saying "ATTI mode". I definitely don't recall thinking "maybe I should go higher" though...  Looking at the stick positions at the last point the flight log registered it seems like I'm at least going forward in a direction that should result in getting closer not further (not sure that really means anything though at this point).

Daniella - I am also still confused about how the Spark was able to drift in ATTI while disconnected for over a minute.  Not to mention it seems to have disconnected a few other times for more than 3 seconds before that (as indicated by "Partial or complete disconnect").  If it would have self-landed during one of those times it probably would have landed on the beach and it could have been possible to recover.  What I do know is that I spend quite a few hours checking the area approximately indicated by yellow/blue and didn't find anything.  One other thing that I find odd is the battery life...  It's at 67% before and after the extended disconnect.  I think it should have been at least a few % lower at the last moment of connection.

ajmckay Posted at 1-1 21:27
Happy New Year. So let me say that you all have been very helpful in answering my questions and helping me piece this together.
JJB - I really hope that it didn't start RTH and then ATTI again and land somewhere completely random...  It never gave me any indication that it reestablished connection with the RC.  That is to say that the light never turned green again after I noticed that it had turned red (I don't remember exactly when that was though).  Also I wanted to specifically say thanks for posting those "FRAP" screens.  That looks like a pretty cool program - did you write it?

For the battery, maybe since it was just drifting in the wind and not fighting it to stay at the same spot, it is possible that it did not use much battery.  If that's the case, it could have gone for a while.

Good luck with it, hope you find it or DJI send you a new one.

Daniella3d Posted at 1-1 17:25
So basically if it disconnect from the RC and goes into ATTI, it just land anywhere and anyhow after 3 seconds? it just lower its altitude until it hits something, trees, hills, ground, anything.

It's funny that it disconnected from the remote for more than 3 seconds and was still at the same altitude, but more than 300 away from the previous position, so drifting.

Hi Daniella,

Good observation but what we do not know is how long the last ATTI lasted.

63 seconds nothing in the file means no connection but somewhere between 1 and 63 secs the Spark gained GPS connection again (but no connection to the RC)....that`s why the first record after connection is the RTH mode.
When GPS connected and no connection the the RC RTH is activated, first action is gain height, turn towards home and move to home.Height was already OK, heading turned from 223 to 172, as the heading after that is 174 imo Spark turned anti clockwise from 223 to 174, this takes some time....(home heading from that point is 174.3) [ and yes, i have seen in other files this kind of taking the long way around.... ]
If GPS connection was intermittet than this 3 seconds try to land did not got active, or it did !! and into RTH mode height increased again to the RTH height.
We will never know but 3 seconds in ATTI without connection to the RC = autolanding.

Sometimes only the video is lost, RC still connected but the user is not aware of that. Good thing is that the attitude indicator/compass shows actual Drone data, so usefull to steer the right way. Not sure if records are written to the flightlog if that situation.
cheers
JJB

ajmckay Posted at 1-1 21:27
Happy New Year. So let me say that you all have been very helpful in answering my questions and helping me piece this together.
JJB - I really hope that it didn't start RTH and then ATTI again and land somewhere completely random...  It never gave me any indication that it reestablished connection with the RC.  That is to say that the light never turned green again after I noticed that it had turned red (I don't remember exactly when that was though).  Also I wanted to specifically say thanks for posting those "FRAP" screens.  That looks like a pretty cool program - did you write it?

Hi ajmckay,

Thanks, yes FRAP is my own analysis tool.

Next step is to implement CVSview data into FRAP, than all available data can be seen in a nice way.
(i think)

cheers
JJB

That far out, you weren't leaving yourself any margin for something to go wrong. Too far to receive video and see through the camera. Too far to see the drone. Oops, too far for radio control. Oops, lost GPS signal. Lost spark.

Aviators talk about "operating well inside the envelope." You were real close to the edge.

Tentoes Posted at 1-2 05:57
That far out, you weren't leaving yourself any margin for something to go wrong. Too far to receive video and see through the camera. Too far to see the drone. Oops, too far for radio control. Oops, lost GPS signal. Lost spark.

Aviators talk about "operating well inside the envelope." You were real close to the edge.

Hi, thanks for the reply.

After reading on the forums it turns out that you're certainly correct.  At the time though I didn't have any reason to think that the Spark wouldn't return home after it lost signal like it has every other time at various distances.  It was a pretty wide open area with no power lines or obstructions in the flight path.

Regardless the outcome of my spark, I'll be incorporating some additional layers of redundancy and putting the things I've learned on the forum to good use.

Sorry 'bout your spark.

It's likely somewhere near it's last contact, but "somewhere near" might take a lot of searching.

were u in 2ghz or 5ghz?
i find that 2ghz is unstable and got few sudden drops even when near

SAADHERO Posted at 1-2 22:19
were u in 2ghz or 5ghz?
i find that 2ghz is unstable and got few sudden drops even when near

The last I recall seeing as I watched my signal bars was 5ghz.  I went to check in the flight log details and didn't see anything about that though.  Strange!

There surely wasn't any competing wifi signals or anything nearby.  There's only spotty cell service as I was in the middle of a large park.

ajmckay Posted at 1-3 07:38
The last I recall seeing as I watched my signal bars was 5ghz.  I went to check in the flight log details and didn't see anything about that though.  Strange!

There surely wasn't any competing wifi signals or anything nearby.  There's only spotty cell service as I was in the middle of a large park.

well i guess its a malfunction so u should get a replacement hopefully

Your first mistake (and most likely reason it went into ATTI mode and you got the error messages) that I don't see anyone mentioning was calibrating the compass. I know everyone keeps reading on the forums that you should always calibrate the compass in a new place, but this is false info. The only time you should calibrate the compass is when the DJI Go app tells you it needs to be calibrated. Even this is misleading because it says "calibrate compass or move the aircraft". 99% of the time when it shows that warning it means there was some type of metal near where you turned your drone on, wether it be from a car, or metal rebar under cement. This causes the magnetic field interference and calibrating the compass will do nothing. The only solution is to move the aircraft. It is interesting that you still got the compass error when taking off from the middle of the beach. Did you turn your drone on where you launched? The only metal I can think of that would be on the beach, would maybe be beach chairs. But anyways, do not do a compass calibration (even if you are 5000 miles away from where your last flight was). Only do it when the app says to.

SAADHERO Posted at 1-4 12:07
well i guess its a malfunction so u should get a replacement hopefully

DJI support emailed me back and requested all the flight data.  If it's not a malfunction then I don't know what could have caused the fatal loss of signal.

One thing is for sure - I am fairly invested in the hobby.  I bought the Spark as a controller combo in November.  Since then a friend who is a more experienced quad enthusiast got me 3 new batteries and I also just recently got the multi charger.  Add to that some smaller accessories and also DJI care refresh - it's a huge bummer to have this happen.  My plan was to build up my interest and skill in the hobby with the Spark as that's all I could afford now and then in a year or 2 get a Mavic assuming my skills improve with the Spark.  It's tough to convince your significant other to get another drone after you lost one though!

I have a Hubsan X4 (H107) but I really never got that good with it.  I have a small house (noise and not much room indoors) and when outside the tiniest amount of wind would make it too hard to control.

fansb: Thanks for the reply and the advice.  I find conflicting direction on this subject so I'm not surprised that it's coming up.  Some more info:
- For the last flight specifically the drone was turned on over asphalt.  I sat it on a brick pillar (may have had concrete?) for takeoff.  Controls were responding like normal.
- The compass calibration was done in the middle of a beach (no chairs even).
- I had 2 successful flights before the one where it was lost.  
- The prior flight was very close to the last flight where the drone was lost (can see the next takeoff location in the footage!). I did some fairly detailed maneuvers and it was spot on.


If there was an issue with calibration wouldn't it have presented itself immediately? I see plenty of examples where this same/similar error happens within 100m even and the spark goes into ATTI.
I'm pretty sure I've seen flyaway threads both with and without compass calibration.  I get that doing compass calibrations too often can increase the chance that something could get done wrong during the calibration process but other than that what happens that would make the new data less reliable than the old?  I still need to research this to understand why.

Finally - if this does get resolved somehow I'm painting the drone orange and strapping a tracker to it!  Also plan to fly closer/more strategic.  Instead of going really far, walk to the mid point and fly half the distance in either direction for example.   

Another good advise I got is to fly against the wind first, then if the Spark goes into ATTI, there is more chance that it would drift back to you with the wind. If you start your flight with the wind, then it won't drift back in your direction. I think it's a good advise.

Daniella3d Posted at 1-4 22:11
Another good advise I got is to fly against the wind first, then if the Spark goes into ATTI, there is more chance that it would drift back to you with the wind. If you start your flight with the wind, then it won't drift back in your direction. I think it's a good advise.

That's pretty good advice and makes sense to me.

fansb1fe1104 Posted at 1-4 16:45
Your first mistake (and most likely reason it went into ATTI mode and you got the error messages) that I don't see anyone mentioning was calibrating the compass. I know everyone keeps reading on the forums that you should always calibrate the compass in a new place, but this is false info. The only time you should calibrate the compass is when the DJI Go app tells you it needs to be calibrated. Even this is misleading because it says "calibrate compass or move the aircraft". 99% of the time when it shows that warning it means there was some type of metal near where you turned your drone on, wether it be from a car, or metal rebar under cement. This causes the magnetic field interference and calibrating the compass will do nothing. The only solution is to move the aircraft. It is interesting that you still got the compass error when taking off from the middle of the beach. Did you turn your drone on where you launched? The only metal I can think of that would be on the beach, would maybe be beach chairs. But anyways, do not do a compass calibration (even if you are 5000 miles away from where your last flight was). Only do it when the app says to.

There is too many conflicting information about compass calibration. Too many people saying that it should be done each flight to avoid problems, too many saying it should be done at a new location, and too many saying it should not be done unless asked for.

The one thing that gives a bit of light on it is checking if the triangle in the compass point at the same direction as the Spark.