At 105ft after 129 flights over 120,000ft travelesd.   Lost GPS went into ATTI mode then drifted in wind into a tree and crash to ground.  I could not react fast enough to regain control.

Poor little Drone broken now and off to DJi Care Refresh for help!

-Fun with Drones-

One more to the List. =/
hope Dji fix this problem as soon as possible, i dont take off untill theres a fix.

Just about considering a spark now have reconsider

Sorry to hear that, it may help to find out what happened. You can upload your logs to link below.

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

We are sorry for the accident, please sync the flight records and then contact support to send in the drone, we will have data analysis team try to analyze the reason of the accident.
The drone will be covered by warranty if it is malfunction.

Welcome to the club.

hmm , after 129 flights i can tell you did learn how to fly it so a user error is less possible here.. WTF DJI.  

DJI Mindy Posted at 2017-8-3 19:15
We are sorry for the accident, please sync the flight records and then contact support to send in the drone, we will have data analysis team try to analyze the reason of the accident.
The drone will be covered by warranty if it is malfunction.

Drone has been sent, should be received today.

hallmark007 Posted at 2017-8-3 14:57
Sorry to hear that, it may help to find out what happened. You can upload your logs to link below.

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

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

Went P-GPS to ATTI 8m into flight while altitude decline at 107ft then the following errors and crash. No control.

Message:
-Yaw Error
-Compass Error. Exit P-GPS Mode; Yaw Error
-In Flight,workingIMUencountersheading exception,please switch to atti mode if craft behave abnormally; -Yaw Error
-GPS Position NoMatch

All well bird had 17-18 Satellites even when into ATTI Mode

-Brett

Was using Lasted firmware .500?

URGENT REQUEST
Dji please come with statement and solution regarding this issue, before a fallen spark causes serious damage or harm . There have been already too many crashes and warnings . Enough is enough

HiWEB Posted at 2017-8-3 22:49
Via http://www.phantomhelp.com/LogViewer/Upload/

Went P-GPS to ATTI 8m into flight while altitude decline at 107ft then the following errors and crash. No control.

You need to put your link up, that link is only link to phantomhelp page.

Why would a compass error disable GPS?
Orientation (compass data) and position (GPS data) are different but related.
When the Spark is moving and getting position information from the GPS that tells the Spark that it is travelling in a different direction from what it's getting from compass data, The Sparks programming can't work with the data conflict.
The solution is to drop data from one source and since the Spark can't fly without a compass but can without GPS, it's the GPS data that gets dropped.

There have been some compass problems with Sparks lately or your issue may have been related to something you did.
I'd be curious about when and where you calibrated it and where you launched from as these may be relevant factors.
You can also see the flight data to learn more and perhaps solve the mystery of your incident.

When you are trying to control your spark compass is pulling in different direction causing conflict with your commands to spark and basically confusing the heading you want to go thus IMU encounters a problem.

I have seen dji offer warranty for similar situations, so hopefully they will for you. Good luck.

Regarding the compass calibration I have been reading a lot of misleading assumptions and interpretations in this forum already.

I have even read several posts people suggesting to calibrate everytime you change location or even before every flight!

I just wanted to clarify that I only did compass calibration when being asked by the app to do so. This is what the manual says.

My guess is that something is not right with the compass here: does the Spark actually have 2 compasses or just one?

ciesnik Posted at 2017-8-4 01:25
Regarding the compass calibration I have been reading a lot of misleading assumptions and interpretations in this forum already.

I have even read several posts people suggesting to calibrate everytime you change location or even before every flight!

Spark has only 1 compass unlike Mavic which has 2, but that said there have been very similar problems with Mavic over its lifetime, I suppose there is so much packed into such a tight space, that yes something could be effecting compass if wiring is net sealed properly this could be a problem.

There are also many outside factors that can cause problems for your compass , calibration to often increases the chances of doing it wrong and then having problems, taking off from rebar concrete balconies metal objects, flying over metal roofs electrical cables etc.

It's very simple to check the heading on your compass relative to the heading on your spark before you fly, by checking small red triangle on your map. If both headings match the only result you will get by calibration is the same result you already have or worse, so hence no need to calibrate.

There are some who get warning to move or calibrate and choose to calibrate on the spot where they get the warning Not Good, always move to another area and try again.

I have 4 AC and I don't think in the last two years I have calibrated compass more than 3 times and I've had numerous times travelled 2/3 hundred miles and never had a problem I have never calibrated my spark and don't intend until it shows me it needs calibration.

HiWEB Posted at 2017-8-3 22:43
Drone has been sent, should be received today.

Then no worries, local repair team will take good care of it, please wait patiently.

Szwedu Posted at 2017-8-3 21:38
hmm , after 129 flights i can tell you did learn how to fly it so a user error is less possible here.. WTF DJI.

is this a serious argument ? Human factor errors are (unfortunately) much more complex than just gaining experience with time. Unfortunately again, they are also very difficult to analyze, put into equations, predict and reduce risks.

What is a bit scary with this case hapenning with large number of flights is that it more or less goes against the idea of a bad batch of Sparks units.
Was it with a new battery or something new ?

You checked the kp index?
The above kp index 4 may be gps problems.

Rafal R. Posted at 2017-8-4 03:28
You checked the kp index?
The above kp index 4 may be gps problems.

Special space weather events such as large solar eruptions can have impact on GNSS especially when using non differential GNSS such as the one in lost cost UAV's.
However I've never seen any effect at practical level. Even during large eruptions that occureds few years ago (Its not the same at all for other kind of radio links where I did experienced heavy impact)
As for "regular" ionosphere activity this is quite unlikely that it will raise up to some issue level if he was flying in the US :

http://www.septentrio.com/insigh ... -positioning-course

The main effects are through modification of TEC (total electron count) and scintillation effects. Those affects the time GNSS signal takes between satellites and receiver, and in the end can end up with quite wrong positions.

website http://www.trimble.com/GNSSPlanningOnline/#/IonoMap

allows to plot amplitude (TEC) and phase disturbances of the ionosphere for a given date and time. I checks and 3rd of August was pretty quiet everywhere !

It also allows to check how many satellits should be visible at a given place/date/time which can be helpfull to check (however, in a open environment with GPS+GLONASS, I can't imagine how you could lack satellties nowadays)

At a practical level, the main issues I experienced with GNSS (I have strong background in navigation, and I use high end GNSS receivers daily in professionnal environment) are :

- self or inter receiver GNSS jamming : a GNSS antenna + receiver is not  a "passive" system. There is a Low Noise Amplifier somewhere, either in the antenna itself or the receiver. If not properly shielded/designed, this LNA can emit signal that will re-enter the antenna and mess-up all measurements.
I already experienced it with low grade receivers such as the one in smartphones (and most of time so far in UAVs). Either from the receiver to itself, or from a low grade receiver to another one if they are too close. (I never had issues with high end receivers being the root cause of issues, on the opposite side of good things, they are more sensitive to jamming...)

- USB3 to GNSS jamming. I experienced it twice with active USB 3 extension cables : if the plug was too close to GNSS antenna, the receiver was not able to get any measurements. Also UsB 3 and GNSS are in different radio band, I guess that harmonics can be too high.

That would be nice to know if the Spark GNSS receiver is a "code only" receiver or if it do have some phase measurements capabilities (e.g. for applying some phase measurement smoothing techniques). Phase measurement enabled receivers are more accurate, but also more sensitive to external environment and jamming.

Matioupi Posted at 2017-8-4 04:40
Special space weather events such as large solar eruptions can have impact on GNSS especially when using non differential GNSS such as the one in lost cost UAV's.
However I've never seen any effect at practical level. Even during large eruptions that occureds few years ago (Its not the same at all for other kind of radio links where I did experienced heavy impact)
As for "regular" ionosphere activity this is quite unlikely that it will raise up to some issue level if he was flying in the US :

I do not think that actually "losing" GPS signal or satellites is what happens in the fly away cases. It seems rather obvious to me that there is either an issue with the compass in the spark causing the software to switch from P-GPS to ATTI mode, or there is simply a software bug which causes the switching to ATTI mode.
Because, in all cases the GPS signal was strong and at least 12 or more satellites have been seen by the AC right before switching to ATTI mode.

Yeah, what he said.

ciesnik Posted at 2017-8-4 05:39
I do not think that actually "losing" GPS signal or satellites is what happens in the fly away cases. It seems rather obvious to me that there is either an issue with the compass in the spark causing the software to switch from P-GPS to ATTI mode, or there is simply a software bug which causes the switching to ATTI mode.
Because, in all cases the GPS signal was strong and at least 12 or more satellites have been seen by the AC right before switching to ATTI mode.

It's not a software bug , it's the way Spark is programmed, if there is a problem with compass then spark is supposed to go to Atti, if you read my post number 12 you might understand . This is the same with all new dji AC.

hallmark007 Posted at 2017-8-4 06:12
It's not a software bug , it's the way Spark is programmed, if there is a problem with compass then spark is supposed to go to Atti, if you read my post number 12 you might understand . This is the same with all new dji AC.

You can not rule out, that there is not a software bug. I understand that the algorithm behaves like you described it. But, what if there is nothing wrong with the compass? And yet, the software behaves like if there is an issue with the signals from the compass and switches to ATTI mode.

Because, it is rather unlikely that there are such strong magnetic interferences occuring minutes after flying smoothly. There is either a hardware design issue (too small - no redundancies - more likely to fall for interferences) or a software bug causing the algorithm to switch to ATTI mode (although the compass/hardware may work fine).

We can not know what the exact issue is. But, there are bunch of possible scenarious here.

ciesnik Posted at 2017-8-4 06:21
You can not rule out, that there is not a software bug. I understand that the algorithm behaves like you described it. But, what if there is nothing wrong with the compass? And yet, the software behaves like if there is an issue with the signals from the compass and switches to ATTI mode.

Because, it is rather unlikely that there are such strong magnetic interferences occuring minutes after flying smoothly. There is either a hardware design issue (too small - no redundancies - more likely to fall for interferences) or a software bug causing the algorithm to switch to ATTI mode (although the compass/hardware may work fine).

If the craft loses GPS and compass, will it still respond to sticks?  The reason I ask is that I had one episode with my P3S where I was flying w/o GPS and without compass.  The P3S did react to my controller, but response was extremely unpredictable.  I was able to land, but it was a tough flight.  I'm wondering if the Spark will behave similarly (it can still be "controlled" with the RC).  If so, then that puts a great deal of emphasis on flying LOS.

hallmark007 Posted at 2017-8-4 01:15
You need to put your link up, that link is only link to phantomhelp page.

Why would a compass error disable GPS?

UAV can fly without a compass. What you don'y have without a compass is your plaform orientation with respect to North at low speeds.

What is needed to "fly" i.e. to maintain the Spark attitude is the accelero + gyro part of the IMU which is enough to get roll,and pitch and a quicky drifting heading (due to Z gyro bias)
When you have a 9 or 10 DOF IMU it's
3 DOF for gyros
3 DOF for acceleros those first measured values are absolutely needed for flying and mainting the Spark levelled. If Spark does no goes upside down in flight, this mean that this part of the IMU is still working ok... (and probably the compass is embedded with it)

3 DOF for compass i.e. Earth magnetic field measurement
1 DOF for barometer to have a non or very slowly drifting altitude reference

If there is a mismatch between Compass and GNSS track on ground (the other heading reference) the Spark should of cause sort it out, but choosing Compass over GNSS is a questionnable choice...
GNSS gives good course over ground if there is not too much wind and if speed is high enought.
Plus GPS allows the user to know where the Spark is on the map (for people flying too far..)

I'd be curious to know if Spark takes into account magnetic declination that can reach a few degrees. When comparing some GNSS "heading" (which also embedd wind drift) and compass heading, the magnetic declination should be considered too. If incident occurs on places where this value is quite high, it may reduce the margin to consider that GNSS track and compass are saying the same thing.

Still, it should be possible to design a flight procedure to enter a degraded flight mode between P-GPS and ATTI.

Of course if your flying in middle of tree (where GPS is maybe not that good) such a mode will probably not help 100% of cases.

ciesnik Posted at 2017-8-4 06:21
You can not rule out, that there is not a software bug. I understand that the algorithm behaves like you described it. But, what if there is nothing wrong with the compass? And yet, the software behaves like if there is an issue with the signals from the compass and switches to ATTI mode.

Because, it is rather unlikely that there are such strong magnetic interferences occuring minutes after flying smoothly. There is either a hardware design issue (too small - no redundancies - more likely to fall for interferences) or a software bug causing the algorithm to switch to ATTI mode (although the compass/hardware may work fine).

I think Charles ADAMS has answered your question , you must remember your the pilot and your in charge of your AC, while we don't know if this is a software or hardware issue, the behaviour is exactly the same as Mavic phantom and inspire.
Interference is something that can come slowly or quickly, that's why many pick up interference directly from the ground.
On thing that is noticeable from all these cases is that what happens is exactly what has been happening with all dji craft when there is a compass problem, the fact that it goes into Atti mode is correct and pilot should be able to deal with this. Many more pilots have dealt with this occurrence correctly than have not.

Matioupi Posted at 2017-8-4 06:48
UAV can fly without a compass. What you don'y have without a compass is your plaform orientation with respect to North at low speeds.

What is needed to "fly" i.e. to maintain the Spark attitude is the accelero + gyro part of the IMU which is enough to get roll,and pitch and a quicky drifting heading (due to Z gyro bias)

The problem is your discussing something that is possible, but not how Spark and all new dji aircraft are set up, that's something to be discussed with dji engineers, I'm explaining what happens with spark as it is set up now, you are also assuming that when dji aircraft go into Atti mode they are uncontrollable but it is quite the opposite of that, I have had compass problems with Mavic it is very controllable in Atti mode, and I for one think the way it's set up is working fine.

It seems strange that dji would be adding a second compass to all of their new bigger aircraft if there was something else out there that was better or safer.

Matioupi Posted at 2017-8-4 03:07
What is a bit scary with this case hapenning with large number of flights is that it more or less goes against the idea of a bad batch of Sparks units.
Was it with a new battery or something new ?

Nothing new same bird flying..... Just error then abnormal flight and crash.

+1  I had this error pop up and ATTN mode. disable P-GPS mode. then it flies to forward. and it seems 3D vision wasn't active as well. it kept going forward and crashed on stalk. however luckily min didn't get any crash on ground or scratch a lot. After that I was fine to fly again.

ciesnik Posted at 2017-8-4 05:39
I do not think that actually "losing" GPS signal or satellites is what happens in the fly away cases. It seems rather obvious to me that there is either an issue with the compass in the spark causing the software to switch from P-GPS to ATTI mode, or there is simply a software bug which causes the switching to ATTI mode.
Because, in all cases the GPS signal was strong and at least 12 or more satellites have been seen by the AC right before switching to ATTI mode.

agree with this comment.  sounds more like a software glitch and poor software decision making.

hallmark007 Posted at 2017-8-4 07:10
The problem is your discussing something that is possible, but not how Spark and all new dji aircraft are set up, that's something to be discussed with dji engineers, I'm explaining what happens with spark as it is set up now, you are also assuming that when dji aircraft go into Atti mode they are uncontrollable but it is quite the opposite of that, I have had compass problems with Mavic it is very controllable in Atti mode, and I for one think the way it's set up is working fine.

It seems strange that dji would be adding a second compass to all of their new bigger aircraft if there was something else out there that was better or safer.

I don't see why this is a problem, I know that this is not the way it is coded right now and I also know the AC can be flown in ATTI.
Altough I believe that the typical Spark user is not always prepared to do so (and DJI should know that)

To go in your sense (from other posts from you I've been reading) I even think that many of those issues are related may be related to bad user calibrations (but we may also say that calibration QC by DJI is maybe not tight enough. I have not recalibrated since delivery, but if I do so, I would prefer that Spark tells me that my calibration is KO 10 times in a row because QC control thresholds are tight rather than letting me fly with a bad calib. e.g. it should not let do the calib inside... which from videos I've seen on Youtube looks possible. I would always do the calibration outside, far from any metallic part (that can be hidden like in concrete slabs or walls)

I hope that DJI engineers are reading all those post and sorting rubbish from good points (that would be great to have more feedback from them, to have a know issues list, to have notes about issues they are working on....)

As for compass issue, my point is that for an AC such as Spark which is intended to be flown without AC the choice of selecting the compass over GPS when they mismatch seems questionable to me.
Without RC range is very short and ATTI can only be flown "in control". Even with RC , range is still short in some places (EU) and in addition some users report range issue (which is not my case)
I would find much safer that the flight control laws navigation stack give priority to "staying withing an area around the home point / last position" would that result in "brownian flight" because GNSS heading is noisy.
At least pilot who have safely choosen a wide open area and a reasonnable max distance for flying will be on the safe side from that first choice as fly alway would be "bounded".
There may be other reasons to favor compass over GPS, but I'm not really seing good ones.

Aside from that, the other strange thing is that for many users, this switch to ATTI seem "permanent" I mean, it does not recover to GNSS at some point and that is pretty strange. I've experienced some very short ATTI mode messages, but it always returned to P-GPS almost instantly.

Matioupi Posted at 2017-8-4 09:59
I don't see why this is a problem, I know that this is not the way it is coded right now and I also know the AC can be flown in ATTI.
Altough I believe that the typical Spark user is not always prepared to do so (and DJI should know that)

Hi Matioupi, we are not far apart in the way we are thinking, and I agree with a lot of what you say.

First we must be responsible for how we take care of our AC and how and where we fly our AC. However if there is a Quality Control issue then we are basically pis*ing into the wind, if this is the case with some spark AC then I feel it's a small number, but can become somewhat exaggerated when other s crash and see there problem the same as those with genuine problem.

I do think dji take feedback from these forums they would be fools not to, but changes coming direct from feedback will take some time.

Regarding which would be best dropping gps or dropping compass , I'm sure it's worth debating and some testing I'm sure could be done,  but I'm sure if dji were testing such different methods and better safety redundancies then this would be kept very secret.
I'm sure dji are always looking at improving safety and redundancy in their AC, we can see this with other new AC like Mavic and phantom which now have redundant Imu and compass.

I agree flying with Phone virtual sticks make it very difficult to fly in any other modes except gps, and some oversight on dji's thinking leaves little to be desired, I don't see the point with such capability that Spark has why they would offer the option of using phone wifi to control and I believe it was a mistake to sell Spark without RC, if they had have sold with RC only they then could have offered using phone but only allowing it to fly 20 m x 15 m, and this would be a much safer option.

As regards to brownian flight, I bow to your superior knowledge, I have heard of it once in flight school but only in passing , I must read up on this soon.

Recovering in Atti , it's very rare to recover if compass is the problem, I suppose the best chance maybe to fly higher and maybe God will help, loosing gps on its own is usually recovered by simply finding open sky and increasing altitude, being at higher altitude in VLOS is always much easier to manage Atti.

I think in the near future we will see many changes in this technology and safety and redundancy will be improving all the time, let's hope it's sooner rather than later.
Good luck.

hallmark007 Posted at 2017-8-5 04:17
Hi Matioupi, we are not far apart in the way we are thinking, and I agree with a lot of what you say.

First we must be responsible for how we take care of our AC and how and where we fly our AC. However if there is a Quality Control issue then we are basically pis*ing into the wind, if this is the case with some spark AC then I feel it's a small number, but can become somewhat exaggerated when other s crash and see there problem the same as those with genuine problem.

brownian motion is the motion of gas molecules (for example) and their trajectories looks random. What I meant is that flying in GPS only may result in a lot of direction changes because heading may be noisier than with a compass (especially at low speeds when GPS track is even undefined)