We had an exciting weekend a couple weeks ago.  We were out doing a range test and experienced something that we had never seen before.  Our DJI Mavic Air has the latest firmware updates and we have experienced no issues up to that point.  Half-way into the flight we experienced some sort of interference that caused a loss of control of the drone and an almost crash.  The drone was jinking in small increments and we experienced a total loss of GPS signal, only to have it return after moving away from a certain area.  There was no explaination for the behavior of the drone.  Take a look at the video (all range test videos are long, apologies) and let us know your thoughts.  Can we send our fight data recorder information to DJI for analysis?



Thank you for any advice or insight.

Keep flying!

Living In Sin (city)

Flying too close to Area 51? Area 51   

Ok, not sure if its the same thing but your video seemed to be very similar to what I experienced earlier today. I was up in the mountains northeast of Phoenix and  almost as soon as I launched straight up about 150’ the image on my display appears very similar to yours. Unfortunately I couldn’t even start videoing as i was struggling just to gain control of my  MP. Thought i was going to have to crash it myself. I was thinking it might have been the winds aloft were more than what thought they were. I came to the forum tonight to do some homework on todays event. If you find anything out, please post.

Forgot to mention that i have a Mavic 2 Pro and everything is upto date

Forgot to mention that I have a Mavic 2 Pro and everything is up to date

Forgot to mention that I have a Mavic 2 Pro and everything is up to date

upload your log here of this occurrence,  we can see far more data in a log file and compare it to the video..

not hacked, GPS was probably jammed, prob passing car had a jammer and caused the drone to flip to Atti mode.. means basically uncontrolable if you lack experiance in atti mode.    it shoudl have been more stable but i dont know the conditions..  however IMU and VPS should have held the aircraft in more stable psotioning...
techncially you lose GPS diversity....count didnt drop to 0, but rather bars dropped to zero ..bbars dont indicate signal level but rather quality of diversity.. how spread apart those GPS signals are..more spread the more accurate..  typically GPS jammers will mess up diversity first... prob not close enough to the jammer to lost GPS signal

not to be offensive but it looks like ameture hour..

1.  mavic air doesnt support mutliple controller connections, the connection is encrypted and hidden, yes it is wifi so it is fundementally less secure than something dedicated like Occusync or Lightbridge however, if controller/drone was hacked..you wouldnt be conencted anymore.. however it is unlikely given to even pair another controller system requires the aircraft to be grounded first without props spinning and in pairing mode...

2.  Atti mode = full manual control, and it will feel uncontrolable if your not expoecting it or have no experiance flying Atti or manual drones..  

3. Lost GPS and expected controls to remain stable.. no if GPS is lose and posiotning is seriously compromised such as no visual positioning or anything and your experiance flying atti mode drones is zero your level of control is goign to be zero.

4. RTH is a normal proceedure at said battery level.. but you were ypu over 300 feet over take off location, if RTH was initiated, it doesnt reduce its alt at all it just fly at that alt.... however THANK YOU FOR NOT INITATING IT AND CANCLING IT... your flight was comprimised once... not wirth the risk of flying back and risking another GPS issue

5.  UPLOAD YOUR LOGS!

You need to post your flight log, use link below, I find it strange that your drone might have been hacked, it looked very much like when you turned for home your gimbal completely went out of whack for some reason and aircraft went into Atti mode, your flight log may show a bit more, yes your altitude seemed to change but not by much, you altitude is measured from the point where you took off from.

When aircraft goes to Atti mode and you are out of VLOS it can do strange things and every stick movement made by you will be accentuated and it looks like this may have happened.

Why did you loose gps, all I can see from your video is you were coming close to a mountain which could have affected your signal, and contrary to what poster above says, the most important indication of gps is in your signal bars and you had none so were in full Atti mode for a short time. Whether you knew it or not you managed to get some control until gps returned once this happened.

Canceling RTH makes no difference once it goes to Atti mode RTH will be cancelled and will not work without gps.

Think about it, your out in the desert how many people go out to that part of the desert to fly drones, do you really believe that someone will go out to the desert and wait for a MavAir or any other drone just to disable it, I don’t think so and reading myths on the internet won’t help in your quest to find out 2hat went wrong, as long as I’ve been flying drones I have never heard of one taken out like you think yours might have, I think if such a simple device was freely available and in use, we would not have seen the gatwick debacle. Link for log below.

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

hallmark007 Posted at 2-10 03:59
You need to post your flight log, use link below, I find it strange that your drone might have been hacked, it looked very much like when you turned for home your gimbal completely went out of whack for some reason and aircraft went into Atti mode, your flight log may show a bit more, yes your altitude seemed to change but not by much, you altitude is measured from the point where you took off from.

When aircraft goes to Atti mode and you are out of VLOS it can do strange things and every stick movement made by you will be accentuated and it looks like this may have happened.

think you misunderstood me..in dji go 4, sat bars = quality of diversity in signal ToF between each link..   doesn't mean signal strength or even specific quality of signal but rather how wide diversity is..  bars are most important, technically you can have full diversity with 3 meters horizontal accuracy with as little as 3 sat links.. hell technically you can have full 3D accuracy within 3 meters with as few as 3 sat links at full diversity..

GPS jammers attack both, usually attack diversity by messing with ToF between LV1 and LV2 signals of each link, for reason im not entirely sure of, jammers seem to throw this out a lot further than the jamming of the specific link frequencies...  i know, cuz i own a few and have messed with them but given how illegal they are here, my testing is been very limited in scope.  

as for him being near hill/rock/mountains... not sure he was close enough to lost full diversity suddenly, partial diversity maybe...still have decent horizon in the sky so i cant see any natural reasons for full diversity loss in this specific case..

HereForTheBeer Posted at 2-10 01:20
upload your log here of this occurrence,  we can see far more data in a log file and compare it to the video..

not hacked, GPS was probably jammed, prob passing car had a jammer and caused the drone to flip to Atti mode.. means basically uncontrolable if you lack experiance in atti mode.    it shoudl have been more stable but i dont know the conditions..  however IMU and VPS should have held the aircraft in more stable psotioning...

No offense taken, I come humbly to figure out what happened.  I will post the log ASAP.

Ok folks, thank you to hallmark007 for pointing me in the direction of that log viewer (and instructions).  File is located here: Nevada Flight Log File

Flying_Nevada Posted at 2-10 09:27
Ok folks, thank you to hallmark007 for pointing me in the direction of that log viewer (and instructions).  File is located here: Nevada Flight Log File

Seems to be no content there.

hallmark007 Posted at 2-10 10:26
Seems to be no content there.

Please try again; you should be getting a prompt to download the log file in .zip format.

Flying_Nevada Posted at 2-10 11:31
Please try again; you should be getting a prompt to download the log file in .zip format.

[view_image]

I have the file but when I try to open it , nothing no content, it might be on my side but I’ve opened many of these files with no problem, maybe someone will come along a get a better result than I .

here you are! ;)

https://www.phantomhelp.com/LogViewer/CH129A9O8SUS8ZTFTPI7/

hallmark007 Posted at 2-10 11:38
I have the file but when I try to open it , nothing no content, it might be on my side but I’ve opened many of these files with no problem, maybe someone will come along a get a better result than I .

Apologies, the .zip file might not be the best format for all devices.  The raw text file is located here: Raw flight log text file

It may open in your browser when clicked on and you might have to right-click on the link to download locally.

Sorry for all the trouble.

Interesting lol

Hacked?  Yeah, no.

Interference from nearby cell phones? Probably, but that'd have more effect on controllability than GPS.   Put me in the camp that thinks you got caught in a combination of GPS signal defilade and the movement of the satellites.

There are far more likely conclusions to jump to than hacking.

Also, I agree that the canceling of the RTH wasn't the right move.  Every time I've been above my RTH altitude when its activated, it stayed at the higher altitude.

I've also prepared myself mentally that if I ever find myself in ATTI mode due to GPS loss, I'm going straight up to minimize  risk of hitting obstacles.  Fortunately I've flown non-GPS/non-optical flow drones more than anything else.

Lastly, could you see it from your vantage point?  I'm pretty sure VLOS means that you can maintain visual contact.

I'm not trying to be an a*****e here, but when we push the envelope....

Boffin Posted at 2-9 20:49
Flying too close to Area 51? Area 51

Wouldn't have to be Area 51.  There are several military play areas, along with DOD & DOE stuff.

Then there is top-secret submarine base at Lake Mead.  Teens have spent years late at night watching for periscopes out there.  

edit:  i got it to work reconverting it.. im not even more confused, lower litterally at same time you lost GPS and tons of errors..2594 errors most of which arent specifically defined...but some include things like not enough power and refering to motors.... maybe related to few times started going max motor speed..?  not sure but that tilt just before full atti mode was interesting...i wonder if something got caught in the motor..? like moment RTH initiated first time it just started flying stupid...

uploading a few images need a bit more looking into, maybe JJB can come and help out this is an interesting case...
whatever it was, looks like it was more than just purely a loss of GPS diversity.. wasnt a hack but definatly wassomething there.... if you lost GPS positiioning which you did, you lost all diversity but kept full GPS count...yes it flips to litterally uncontrolable Atti mode if your not experianced... however looking at how your aircrat was behaving in video matching it to the logs it looks like somethign more happenened and it managed to recover itself.  you lost altitude without input for reducing it...so thats wild.

Skeeter Honey Posted at 2-10 12:53
Also, I agree that the canceling of the RTH wasn't the right move.  Every time I've been above my RTH altitude when its activated, it stayed at the higher altitude.

I've also prepared myself mentally that if I ever find myself in ATTI mode due to GPS loss, I'm going straight up to minimize  risk of hitting obstacles.  Fortunately I've flown non-GPS/non-optical flow drones more than anything else.

I take zero offense to any opinion, statement of fact, or advice; I appreciate the time it takes to examine the data and come to a more educated conclusion than I can, in my newbie state.  

We were in a very good position for line of sight with zero obstructions.  My first inclination after ATTI mode was to give it elevation, but as you can see, we were pushing the 400 ft ceiling, as per the launch point and ceiling restriction I had on it at the time; in retrospect I should have known there was a slight incline and I should have configured my drone ahead of time to account for the rise in elevation.  A newbie lesson almost learned the hard way.  
I'll have to take your word for it that the drone would have maintained elevation rather than going down far enough to activate the proximity sensors, I'm just a nervous Nellie at this time.

Thanks again for taking a look; don't worry about offending me as I'm a humble person.

:-)

Flying_Nevada Posted at 2-10 18:04
I take zero offense to any opinion, statement of fact, or advice; I appreciate the time it takes to examine the data and come to a more educated conclusion than I can, in my newbie state.  

We were in a very good position for line of sight with zero obstructions.  My first inclination after ATTI mode was to give it elevation, but as you can see, we were pushing the 400 ft ceiling, as per the launch point and ceiling restriction I had on it at the time; in retrospect I should have known there was a slight incline and I should have configured my drone ahead of time to account for the rise in elevation.  A newbie lesson almost learned the hard way.  

no its 400 feet AGL, not ceiling...  AGL = Above ground level.. means you could have flown an additional 400 feet up above that rising ground.   

HereForTheBeer Posted at 2-10 18:15
no its 400 feet AGL, not ceiling...  AGL = Above ground level.. means you could have flown an additional 400 feet up above that rising ground.

Agreed, but the drone seems to track AGL from where it took off from, not where it may currently be at (or at least that has been what I've seen).  Is that correct, or am I off base?

Flying_Nevada Posted at 2-10 19:43
Agreed, but the drone seems to track AGL from where it took off from, not where it may currently be at (or at least that has been what I've seen).  Is that correct, or am I off base?

Somebody correct me if im wrong, but the way i understand how the altude works is that it records a home point and can also read based off of the bottom sensors (?). In my experience its always based off of home point, which is why when I was flying up a cliff face, it thought I was 400 feet up when in reality I was less than 100 feet from the cliff face and only about 200 feet from the bottom of the cliff. Had to adjust max altitude to 550 feet to get 30 feet above the cliff face.

HereForTheBeer Posted at 2-10 17:56
edit:  i got it to work reconverting it.. im not even more confused, lower litterally at same time you lost GPS and tons of errors..2594 errors most of which arent specifically defined...but some include things like not enough power and refering to motors.... maybe related to few times started going max motor speed..?  not sure but that tilt just before full atti mode was interesting...i wonder if something got caught in the motor..? like moment RTH initiated first time it just started flying stupid...

uploading a few images need a bit more looking into, maybe JJB can come and help out this is an interesting case...

This is awesome information!  Thank you for deciphering; I'm glad I can bring an interesting case to the group.

:-)

Flying_Nevada Posted at 2-10 19:43
Agreed, but the drone seems to track AGL from where it took off from, not where it may currently be at (or at least that has been what I've seen).  Is that correct, or am I off base?

drone only tracks altitude from take off position..take off is considered 0 feet.  

lets say you are flying at 400 feet AGL up from take off position and a 100 foot hill you want to climb.. well up can be 400 feet above that hill because that is considered ground level means you are 500 feet above take off position and still perfectly legal 99% of the time.. 1% being if you are in restricted airspace may limit you even lower.

HereForTheBeer Posted at 2-10 19:57
drone only tracks altitude from take off position..take off is considered 0 feet.  

lets say you are flying at 400 feet AGL up from take off position and a 100 foot hill you want to climb.. well up can be 400 feet above that hill because that is considered ground level means you are 500 feet above take off position and still perfectly legal 99% of the time.. 1% being if you are in restricted airspace may limit you even lower.

We are on the same page; in this situation would you have set your AGL limiter higher to accommodate for the destination ground level altitude?  If that's the case, I'm silly for not carefully considering the elevation change and upping the limit to fly 400 AGL at the destination.  I'm not quick enough to change the setting on the fly.

Flying_Nevada Posted at 2-10 20:06
We are on the same page; in this situation would you have set your AGL limiter higher to accommodate for the destination ground level altitude?  If that's the case, I'm silly for not carefully considering the elevation change and upping the limit to fly 400 AGL at the destination.  I'm not quick enough to change the setting on the fly.

i just max mine out to 500 meter limit in app..  then i fly with common sense, then i dont need to change my limit every new location...   for me its 500 meters alt limit and 75 meters for RTH..

AlphaFlightNW Posted at 2-10 19:48
Somebody correct me if im wrong, but the way i understand how the altude works is that it records a home point and can also read based off of the bottom sensors (?). In my experience its always based off of home point, which is why when I was flying up a cliff face, it thought I was 400 feet up when in reality I was less than 100 feet from the cliff face and only about 200 feet from the bottom of the cliff. Had to adjust max altitude to 550 feet to get 30 feet above the cliff face.

Think about it. If it wasn't 400' AGL then people in the higher elevations would never be able to take off.  The FAA rules pertain to AGL because if it didn't then no one would be able to fly in most of Colorado or Arizona (for example), which simply is not the case.

Hi,

Wow, you have been so lucky !  at 9m 7 seconds in he video ; how height above the ground ??  just enough to keep flying.

Only thing i can see is the GPS reception went to zero, so ATTI.
GPS signals were ok, all LatLon postions were recorded in a normal way and values.

And landed with 7 % batt level.....you could have made little more distance out....

But Just before the ATTI it dropped 100 feet, error NotEnoughPower. And no RC input to bring it down. So it could have crashed beacuse of this height dropping!

cheers
JJB

JJB* Posted at 2-11 07:43
Hi,

Wow, you have been so lucky !  at 9m 7 seconds in he video ; how height above the ground ??  just enough to keep flying.

ya, weirdest case i seen.    something really freaked the drone out massively to the point not even it understood.  you see all the errors..?  ton of error marked "other"   

...i lost GPS before mid flight, but never had all this happen... usually atti mode is still fairly stable, and altitude is still held, so i wonder what really happened...  my only guess is someone had a GPS jammer in their car driving by, which wiped out GPS diversity for a split second and happened right when the drone was initiating its initial RTH routine caused it to fly basically out of control.. as for "notenoughpower" that a wild one, happens right as it drops alt...

HereForTheBeer Posted at 2-11 08:07
ya, weirdest case i seen.    something really freaked the drone out massively to the point not even it understood.  you see all the errors..?  ton of error marked "other"   

...i lost GPS before mid flight, but never had all this happen... usually atti mode is still fairly stable, and altitude is still held, so i wonder what really happened...  my only guess is someone had a GPS jammer in their car driving by, which wiped out GPS diversity for a split second and happened right when the drone was initiating its initial RTH routine caused it to fly basically out of control.. as for "notenoughpower" that a wild one, happens right as it drops alt...

oke but the altitude was dropped before going into ATTI.

So my question: is there a relation between height drop due to less power and zero GPS reception.
There is a voltage drop (so high current draw) at the same time,  less power left on the circuit board for GPS module thus dropping reception thus ATTI?

When dropping height the speed calculation goes funny, normally this indicates a IMU  / GPS data error.  Not the normal IMU errors are in the file, OTHERs yes.  Mayby OTHERS will kill the GPS reception to zero thus ATTI.

Guess we will never know....

cheers
JJB

I am still confused of launchpoint elevation zero mark.
So launch at top of cliff will not allow submersion downwards into negative feet 2700? The deep canyon on either side of cliff for flight exploration by ma?

Flying_Nevada Posted at 2-10 11:51
Apologies, the .zip file might not be the best format for all devices.  The raw text file is located here: Raw flight log text file

It may open in your browser when clicked on and you might have to right-click on the link to download locally.

Hi Flying -Nevada , having seen your flight log and looking at what others have reported, it looks just at the moment you got warning for RTH aircraft went a bit haywire, I think it may have been one of your motors was disrupted which would cause your craft to almost go sidewards, you can see this clearly in your video where your props on left hand side come into view and craft is no longer holding correct attitude which in turn is giving off IMU warnings and eventually craft going into Atti mode for a brief moment, I’m inclined to think that one of your motors was blocked briefly, could this have been blocked by collecting dust or sand from the air or from where you launched, it could have been picked up at launch but not shown until you turned craft for home, then blocked one of the motors causing reaction from craft before it was cleared and stable flight resumed.

I don’t think there was anything wrong with IMU gps or compass all those warnings were as a result of motor or motors been blocked or locked briefly, when this happens the craft is designed to try to recover the situation so we see IMU errors until it regains the correct attitude for the craft , I believe once motor was cleared.

With regards to your set altitude 400 ft in the US is correct and you should set it at this height as you did, your only problem with that is you flew the craft out of VISION line of sight and if you are opting to do this then check that you will have enough clearance so you don’t run into obstacles that you cannot visually see.

Yes I think you were lucky, but also craft acted as it should, you got back home safe, you have garnered much knowledge and will live to FLY another day.
Good Luck.

cutis Posted at 2-11 08:55
I am still confused of launchpoint elevation zero mark.
So launch at top of cliff will not allow submersion downwards into negative feet 2700? The deep canyon on either side of cliff for flight exploration by ma?

It will allow, but will continue to show a minus altitude as to the height it took off from, it will also use the parameters you have set for RTH , ie it will climb to RTH set height above take off height.

hallmark007 Posted at 2-11 10:50
It will allow, but will continue to show a minus altitude as to the height it took off from, it will also use the parameters you have set for RTH , ie it will climb to RTH set height above take off height.

Understood.
There was chatter forum of this but craft confusion resulted going in the hole resultant casualty i recall, lemme search for it.

Furthered, it was related to losing gps going down into the depression although can not understand why gps flux not there since wide open depression

Hi Flying -Nevada,

Always good to read what other users think what happend, good to learn from each other. Great forum this is!

In your flight no engine block warning or IMU errors etc, so i rule out a blocking motor or IMU errors.

See the gimbal heading differs from the aircraft heading, just after the RTH turn was ended.So what is seen on the video is not exactly the movement of your MA.

cheers
JJB

i was initially thinking the drone got clothes lined as well before i got the logs and only had video...   with logs, its even more interesting than that...what triggered it... that the mystery. if dirt or debris caused it we have a motor current error/motor obstruction error/NotEnoughForce error or something like that appear to indicate propulsion system issues, even if flight is recovered/maintained..  interesting thing to note in the errors is "NotEnoughPower" error..but no voltage or current error to indicate why..

Although some may say there were no problems with IMU , it is a fact that unless your IMU is holding a good level and correct attitude then it is having a problem doing what it should be doing, it may not show in the log graphs above it will almost certainly show in your .dat file that there were complications with IMU not working as it should.

As I pointed out in my first post that you can clearly see that gimbal was completely out of whack with aircraft heading.

Would be interesting to see if there was any debris collected in motors at launch or in the air.