Mavic Air Fly-away

[Helgegustav]

After batteryOn was it moved from one location to a different location for launch? Or, was it powered on and launched without moving it?

I may have powered it on inside my house and then brought it out, but I am not sure

 

[sar104]

I’m restricted to my iPad for the next couple of days. Try this. The Mavic Air has two IMUs. Specify Engineered then DatDefined and then look at lat long in IMU(1). The units are radians.

There are other discrepancies beyond what you pointed out. GPS:heightMSL does a jump discontinuity at one point. But, it’s the same flight. I don’t think the .DAT is corrupted. It does have gaps though as does the .txt.

I’ll need to take another look but I’m not particularly surprised that gps coords jump. GpsHealth was inadequate at first.

@Helgegustav did the AC sit at a geomatically distorted location before being moved to the launch site? Inside a vehicle? Or on a truck tailgate (my flying buddy and I did this)? This would explain why magYaw was correct but Yaw was compromised at launch.

The IMUs do not agree at all. The recorded lat/long location is clearly derived from IMU_0, but IMU_1 agrees, broadly, with the GPS data.



This plot shows the vicinity of the flight but, as I mentioned above, IMU_0 and the recorded location end up 20 km to the NE - off the scale here, as shown in the zoomed out view below.



I'm really not sure what to make of this. The behavior still looks like the originally proposed compass error - and the txt file, plus initial magnetic heading in the DAT file, completely support that. These DAT file data make little sense to me at all beyond that.

 

[sar104]

I may have powered it on inside my house and then brought it out, but I am not sure

It was 40 seconds from power on to motor start. Integrating the velocity data suggests that it may have moved around 12 m during that time - it says 5 m east and 8 m south, but that could be backwards due to the compass error.

 

[Helgegustav]

It was 40 seconds from power on to motor start. Integrating the velocity data suggests that it may have moved around 12 m during that time - it says 5 m east and 8 m south, but that could be backwards due to the compass error.

OK. I cant really remember to be honest.

Anyway, I've been in contact with DJI and provided them the flight logs, and they've said they'll look into it. It will probably take a few days at best. It will be interesting to see what conclusion they make. Big thanks to you for sorting this out. Is it ok if I show them your graphs and explanations?

 

[sar104]

OK. I cant really remember to be honest.

Anyway, I've been in contact with DJI and provided them the flight logs, and they've said they'll look into it. It will probably take a few days at best. It will be interesting to see what conclusion they make. Big thanks to you for sorting this out. Is it ok if I show them your graphs and explanations?

Good luck. You are welcome to use anything that I've posted.

 

[highcastle]

Wow this thread is the best thing I’ve read here period.

Im pretty sure most of us thought OP was the classic first-time drone owner whose expensive new toy just broke and they want someone to tell them it’s not their fault.

Then the logs come out. And a name is forever etched in our memories. The name of a man who can look at a flight log and tell you to the day what you ate for dinner 3 years ago and where.

sar104.

I hope you work on this kind of stuff for a living. Your analysis and reanalysis of the flightlog data was infinitely more compelling than I would have thought possible.

Hats off to both sar104 and famous OP Helgegustav - sorrrrrry we all thought you were a clueless noob at first - and thanks to the both of you for proving us wrong!

 

[Jameslavis]

I’ll even say sar should start a pinned post on good preflight checks and in air trouble shoot tips. We all spend time researching but come on, we learn tons on posts like these. A nice ongoing guide would be great.

 

[sar104]

Wow this thread is the best thing I’ve read here period.

Im pretty sure most of us thought OP was the classic first-time drone owner whose expensive new toy just broke and they want someone to tell them it’s not their fault.

Then the logs come out. And a name is forever etched in our memories. The name of a man who can look at a flight log and tell you to the day what you ate for dinner 3 years ago and where.

sar104.

I hope you work on this kind of stuff for a living. Your analysis and reanalysis of the flightlog data was infinitely more compelling than I would have thought possible.

Hats off to both sar104 and famous OP Helgegustav - sorrrrrry we all thought you were a clueless noob at first - and thanks to the both of you for proving us wrong!

Well thanks for the endorsement but don't forget that there are others on this forum who do this too - especially @BudWalker, who wrote the CsvView and DatCon software, and @msinger, who runs the PhantomHelp website.

I do this from an interest in the flight control systems and also in forensic puzzles, not for a living, although I use many of the same methods regularly for real work.

 

[Prismatic]

Well thanks for the endorsement but don't forget that there are others on this forum who do this too - especially @BudWalker, who wrote the CsvView and DatCon software, and @msinger, who runs the PhantomHelp website.

I do this from an interest in the flight control systems and also in forensic puzzles, not for a living, although I use many of the same methods regularly for real work.

I just cannot LIKE this enough. This is possibly the most interesting technical read I've had in the many months since I retired as an engineer (but my interests are wide, so I won't commit on that. )

I truly get how honed skills and insatiable curiosity can drive a person to dissect seemingly impossible puzzles, just for the fun of it. But "fun" does not adequately covey the satisfaction of a difficult analysis that results in a clear answer. And when the answer is a surprise even to you, there's a big bonus exponent on "fun"!

Similarly, "kudos" does not adequately convey my appreciation for what I've read here, but it's all I've got. So, *ahem*: Kudos! To you, @sar104, @BudWalker, @msinger, and those whose names I've yet to learn. If ever I find a topic on which my own remotely tangential expertise applies, I can only hope to rise to the standards you set.

 

[Helgegustav]

Wow this thread is the best thing I’ve read here period.

Im pretty sure most of us thought OP was the classic first-time drone owner whose expensive new toy just broke and they want someone to tell them it’s not their fault.

Then the logs come out. And a name is forever etched in our memories. The name of a man who can look at a flight log and tell you to the day what you ate for dinner 3 years ago and where.

sar104.

I hope you work on this kind of stuff for a living. Your analysis and reanalysis of the flightlog data was infinitely more compelling than I would have thought possible.

Hats off to both sar104 and famous OP Helgegustav - sorrrrrry we all thought you were a clueless noob at first - and thanks to the both of you for proving us wrong!

Haha! Spot on. Never expected this kind of expertise from this forum.

At least I am a bit less clueless noob after this thread!

 

[Slowcaptain]

Haha! Spot on. Never expected this kind of expertise from this forum.

At least I am a bit less clueless noob after this thread!

Same here. But on the other hand, forums are exactly the place to find such expertise. You can find lot of other information on websites if you Google but not this. I had similar experience on a few other forums too.

 

[baddog]

Hello!

So my Mavic Air flew away today. I started off with 76% battery, recorded RTH point with gps signal and all, and when I reached 30m altitude it freaked out. It went full speed ahead away from me. activating RTH function had no effect at all and it didnt respond to the controller. ( after some googling I've discovered that sport-mode is your friend in these situations...)

I lost videotransmission at about 250m distance ( 4km range my ***) but somehow it was still transmitting its GPS-position, which made me able to find it about 1km away from starting point.

It had crashed into a pit of sand and gravel. 3 of 4 propellers broken and sand and dust trapped in the motors and air intakes. Amazingly, not a scratch on the lens, and the gimbal seems ok. I brought it home and started to clean it up, and noticed that it sounded like a freaking maracas. This fell out from behind the gimbal after a bit of shaking. I don't dare to start it because of the sand that's trapped in the motors.

I have contacted DJI support and the reseller I bought it from, hoping that some kind of warranty will cover this. I'm hoping their support is not as bad as people say.

Thoughts?

Anyone else who has been in contact with DJI support with something like this?

Never had these flyaway issues with my MP. I just assumed there were pretty much the same when it comes to navigation, gps, and control... Sorry that you crashed. Share your logs these people on this forum are awesome with diagnostics...

 

[Helgegustav]

Never had these flyaway issues with my MP. I just assumed there were pretty much the same when it comes to navigation, gps, and control... Sorry that you crashed. Share your logs these people on this forum are awesome with diagnostics...

You might want to read the whole thread again. Logs are posted and the issue is pretty much solved

 

[Zorrobazz]

I'm Lovin'n it! R[̲̅ə̲̅٨̲̅٥̲̅٦̲̅]UTION Thanks for a fasinating read of Forensics

 

[MavicFlyGuy]

I think it’s a simple case of not doing pre-flight checks properly. For the first 20 seconds you didn’t have any satellites. What I always do is a precision takeoff just to be on the safe side. You need to tick that box. Or just take your time after switching it on for the satellites to login. Taking off with „0“ satellites is asking for trouble. In the end you had 14 which would have been fine from the start. Patience... even after you change your battery. Luckily you didn’t hit any cars or a Tesla. Plus, probably pick a better takeoff location. Thanks for sharing your experience.

 

[raymo]

Well thanks for the endorsement but don't forget that there are others on this forum who do this too - especially @BudWalker, who wrote the CsvView and DatCon software, and @msinger, who runs the PhantomHelp website.

I do this from an interest in the flight control systems and also in forensic puzzles, not for a living, although I use many of the same methods regularly for real work.

Thank you sar104, msinger, Budwalker and others for all you do to sleuth out these puzzles. Over the years of owning Phantoms and now a MP, I have come to depend on indicators in the readouts that are triggers for "Be careful and take precautions".

Where I live, in the middle of winter we get "Chinook Winds" that drive the temperature from -20 or more up to as much as the low teens plus. Everybody gets outside to do their thing, including fly UAVs. I'm no different. What I have noticed over the years of RC flying, and now drone flying is that differences in ground wind speed and wind speed at slight elevation can be drastically different. Ground wind speed almost zero and elevated wind speeds of 30 mph or more. Experienced this early on with the MP as our ground wind speed was almost nil, and once we were up and flying around, were getting wind warnings. Switched to Sport mode and was able to bring it back home, although at a dreadfully slow pace even in Sport mode. Flying lower would have been an alternative if we had not made any headway getting back and having 90% battery or more was a plus. Two of us were flying at the time and we both experienced the same thing.

The other thing is that GPS lock is so critical. "Poor GPS" may mean "no GPS". I have always come to have as part of my checklist to evaluate GPS condition and number of sats. Also, on numbers of occasions, I have gotten IMU calibration warnings on startup in areas where you might least expect them. The time for that warning to go away varies but can be a considerable length of time. Don't ignore it.

Glad our OP got his bird back and we all learned some new stuff on this thread. I too, use forensics of a sort in my work and really appreciate the abilities of others to sort out a sticky mess.

 

[BD0G]

You can always count on "SarLock Holmes"

Any time I see a thread with fly away or crash in it I try and bring it to his attention (as well as Msinger/ Budwalker) as in many cases if this involves a lost and unfound drone then time is of the essence.

 

[joejaxvic123]

Truly amazing.

 

[mallchin]

Sorry your bird took a hit!

+1 for pre-flight checks

 

[Prismatic]

...Where I live, in the middle of winter we get "Chinook Winds" that drive the temperature from -20 or more up to as much as the low teens plus. ...

Montana? Home of the Chinook tribe? In Loma, near Great Falls MT, the temperature once rose from −54 to +49 °F (−47.8 to +9.4 °C) in just 24-hours. That’s what a Chinook can do!

... Glad our OP got his bird back and we all learned some new stuff on this thread. ...

Amen, brother!