So after all those years with Dji drones with flyaways and near flyaways, i finally got all the proof i need to document that this is still happening, and not because of a user error.
Now its time for Dji to come clean about this, and tell us what the real problem here is, so we all can prepare ourself the best way before, or when it happens.

Today i literally fought to get it back, the log also shows how erratic it was, and that the drone does everything it can to get away, and fight back against my stick command.
Ive had tendencies like this before, i have had incidents, where the drone barely reacted to the stick input, and hitting the end point of the sticks is what barely makes it respond to get it back.

I had just started a Hyperlapse waypoint session, i noticed how wiggly it was in the sky, and thought to myself, it looked like a drunk seagull. Then next i was watching the screen (iphone x) and suddenly noticed the right front leg came into the picture, next thing i know was that i looked up, and caught it slipping sideways in the opposite way, picking up speed.
Luckily i had some ice in my stumack, and kept my head cool, disengaged the hyperlapse, and started to fight it.
The Mavic 2 Pro barely responded to my sticks, that were all the way to the end point.
Finally back over my head, i still had to fight the sticks to have it land safely, veeery twichy with every input.
Safely to say i was a bit exhausted, but relieved.

Notice another interesting thing, how erratic the flight back was all the way, like the drone wanted to get away, while i tried to fight back in the opposite direction.
You can see it on the snake shaped tracking trail.

I was way above water and away from shore, so theres no chance of any magnetic disturbances, also gps shows max satellites.
IF there was a disturbance in the air, it would only start drifting, not accelerating on its own.
Nothing just starts to fly away at full speed, even if the gps or compass lose orientation, then it will just float around.
The sudden speed and heading is coming from a mishap inside the flight controller and software.
You dont have to be a rocket scientist to understand that ;)

So know we deserve to know what this REALLY is about.
Ive had the exact same on 3 different drones and models from Dji, and the problem have been existing for 5, 6, 7 years now?

Im not going to send it back, just to get a refurbished one in return 4 weeks later, and with no further explanation.
I wanna know WHAT the problem is.

Thank you for understanding.

PS, please spare me for the user error comments, i didnt touch anything before it happened, and the log shows nothing wrong.

Ill try to provide as much documentation here as i can, for research purposes.

Since its a PIA to upload attachments here, ill post the link to my facebook post here as well.

https://www.facebook.com/picturized/videos/10155709841095899/?multi_permalinks=2052358831526420&comment_id=2052381591524144¬if_id=1547418394030073¬if_t=feedback_reaction_generic

Flight record here

https://www.phantomhelp.com/logviewer/GOAWE6W0OS1Y03JEWWR0/#

Do you think someone was messing with you and jamming your drone on purpose? I have ran into this problem before, but mine was totally pilot error I entered an NFZ when I flew over 398 ft. and I was able to bring it back home when I lowered the altitude.

Lucas775 Posted at 1-13 14:56
Do you think someone was messing with you and jamming your drone on purpose? I have ran into this problem before, but mine was totally pilot error I entered an NFZ when I flew over 398 ft. and I was able to bring it back home when I lowered the altitude.

While i think its theoretically possible, i would think its highly unlikely, also because i was standing on a remote location away from public.
No, this is the exact same thing ive experience the past years.

i have experienced this sort of thing before back when i owned a spark.  and luckily for me and for you it sounds like we both able to regain control of our aircraft just enough to reel it in and save it.. but yes not all flyaways are pilot error, i always preached this its true...   pilot errors and crashes go hand in hand pilot error and flyaways less likely than people believe.

HereForTheBeer Posted at 1-13 15:25
i have experienced this sort of thing before back when i owned a spark.  and luckily for me and for you it sounds like we both able to regain control of our aircraft just enough to reel it in and save it.. but yes not all flyaways are pilot error, i always preached this its true...   pilot errors and crashes go hand in hand pilot error and flyaways less likely than people believe.

I have no doubt, at all, that most of these incidents are pilot errors, especially from people who never flew a drone before, or never knew that there was a user manual, but this incident is 100% for sure not a user error.
Ive been flying rc helis with cameras and sensors, long before drones came, so i happen to know a great deal about how this works, and what happens if you get a misreading.
Back in the days when we used Tx on the FM band, you could sometimes meet someone else with the same frequency in the air, then it would start to respond to the other guys stick input, but not get stuck infinite like this...

CineView Media Posted at 1-13 14:34
Flight record here

https://www.phantomhelp.com/logviewer/GOAWE6W0OS1Y03JEWWR0/#

What was wind like, I have noticed when flying hyperlapse and changing flyt path if there is any kind of wind it will cause aircraft to move, you were travelling at between 0.9mph and 3mph it wouldn’t take much wind to affect craft, I have noticed this kind of wobble in my footage using hyperlapse in wind.

Looking at other stick movements when craft was in gps it seemed to react according to your stick movements, these can be matched on your CSV log.
What was wind like at 50metres above take off point ?

hallmark007 Posted at 1-14 05:37
What was wind like, I have noticed when flying hyperlapse and changing flyt path if there is any kind of wind it will cause aircraft to move, you were travelling at between 0.9mph and 3mph it wouldn’t take much wind to affect craft, I have noticed this kind of wobble in my footage using hyperlapse in wind.

Looking at other stick movements when craft was in gps it seemed to react according to your stick movements, these can be matched on your CSV log.

It was dead calm.
Dont know if you can see it, but the unstable forward flight were with much larger movements before it took off to the left, just like when you have a weak gps signal and it starts to float more and keeps repositioning.
The camera also started to pan on its own.
This must be a software messup making things go haywire like this.
If it was going in all directions up and down, i would have believed it was getting signals from something else, but not like this.
I think its strange that Dji after all these years havent able to shield the controllers or add a filter or delay in the software that prevents this when abnormal activity starts to occur.

If you have a friend who is a radio operator and has access to a spectrum analyzer, it might be good to go and look at the signal interference in the area.  By your written description I thought it would be a remote area (over water).   I see in the photo, a shoreline of a large population center.  That and numerous ships and boats.   When we fly these toy products, we rely on a relatively weak signal from a low powered transmitter.  In order to increase range (which beyond visual line of sight is questionable need) they make the receivers as sensitive as possible.  Nothing is exclusive about your use of that frequency and many many transmitters will be attempting to share it most times.  I don't think it is inconceivable that our drones have erratic behavior when "swamped" by a strong wi fi signal or other type of transmitter.  I am not familiar with the equivalent of FCC spectrum allocations for your country.  Maybe you could reply and educate me.  I am wondering on the software if you are relying on the "auto" setting for frequency choice, or do you set it according to the interference shown on the display?
In summary, I see your frustration and it leads you to blame a software fault in the code.  This may indeed be the case.  Companies are very hesitant to ever admit bugs, but how often do we get an "update" for the apps on our phone.... that fix the latest bugs?   My thoughts though.... after decades of work with radio signals, is to look for "competition" on the control frequencies in use.  It is not like the RC days when you had a crystal for 75.6 mhz on your transmitter and a colored flag to alert fellow pilots that you were using the channel.   So, I hope you can pin down the fault and not lose your drone.  In the meantime, I would avoid flights over that harbor for sure.  Regards, flightlogic

Sorry to know that mate

Hiya,

Interesting failure, but grrr to be on control of this flight.

Need to have a better look but your CAM went crazy, iy looks like the MP followed your commands. But closer look....not at the end of this flight
See how cam and drone heading split!

Do i know what happend, nope. See the yaw changes before and just after ID 5100. Looks like my MA yaw glitch! (wich i have occasionally)
cheers
JJB

flightlogicRV9 Posted at 1-14 06:26
If you have a friend who is a radio operator and has access to a spectrum analyzer, it might be good to go and look at the signal interference in the area.  By your written description I thought it would be a remote area (over water).   I see in the photo, a shoreline of a large population center.  That and numerous ships and boats.   When we fly these toy products, we rely on a relatively weak signal from a low powered transmitter.  In order to increase range (which beyond visual line of sight is questionable need) they make the receivers as sensitive as possible.  Nothing is exclusive about your use of that frequency and many many transmitters will be attempting to share it most times.  I don't think it is inconceivable that our drones have erratic behavior when "swamped" by a strong wi fi signal or other type of transmitter.  I am not familiar with the equivalent of FCC spectrum allocations for your country.  Maybe you could reply and educate me.  I am wondering on the software if you are relying on the "auto" setting for frequency choice, or do you set it according to the interference shown on the display?
In summary, I see your frustration and it leads you to blame a software fault in the code.  This may indeed be the case.  Companies are very hesitant to ever admit bugs, but how often do we get an "update" for the apps on our phone.... that fix the latest bugs?   My thoughts though.... after decades of work with radio signals, is to look for "competition" on the control frequencies in use.  It is not like the RC days when you had a crystal for 75.6 mhz on your transmitter and a colored flag to alert fellow pilots that you were using the channel.   So, I hope you can pin down the fault and not lose your drone.  In the meantime, I would avoid flights over that harbor for sure.  Regards, flightlogic

I hear you, but i think youre overanalysing this one, and pay to much attention to radio issues.
These incidents have been happening to Dji drones for many years now, and regardless of the area, populated or not.
Ive even experienced it high up in the mountains where there arent even cellphone towers or houses within miles radius.
Blaming it on local interference, is the easiest to blame and make that an natural explanation.
Im certain these are glitches in the controller software, data that gets mixed up and create weird outputs in the other end.
I was 50m above ground above water, and 5-600m away from nearest cell tower, there shouldnt be any cause for overtaking signals like these.
Also, after all these years with all these flyaway incidents, Dji should have managed to be on top of it by creating filters, delays and other safety measurements that would react by entering some sort of fail safe mode that makes the drone halt and reattempt to reconnect with the original source.

Even my 15 year old Futaba radio has a hopping system that changes frequencies every second to protect it from external sources, never once had a glitch with that.
Im sure youre familiar with this yourself, from what youre writing...

JJBspark Posted at 1-14 07:52
Hiya,

Interesting failure, but grrr to be on control of this flight.

Yes, everything went haywire...
Which app did you use for the log?

CineView Media Posted at 1-14 08:15
Yes, everything went haywire...
Which app did you use for the log?

Hi,

This one  >  https://forum.dji.com/thread-145649-1-1.html

cheers
JJB

JJBspark Posted at 1-14 08:42
Hi,

This one  >  https://forum.dji.com/thread-145649-1-1.html

Ah, im on Mac...

Sorry to hear that.

Over any water there is the chance of radio signal scatter, local wifi and your remote signals maybe reflected up off the water and cause wave interference.

SPIKE_151 Posted at 1-15 02:49
Over any water there is the chance of radio signal scatter, local wifi and your remote signals maybe reflected up off the water and cause wave interference.

In theory yes, but todays controllers should be good enough to filter and block out those signals.
Futaba and other large RC companies managed to do this a decade ago buy using a hopping system.

CineView Media Posted at 1-15 08:21
In theory yes, but todays controllers should be good enough to filter and block out those signals.
Futaba and other large RC companies managed to do this a decade ago buy using a hopping system.

"Hopping system" as in Spread Spectrum technology?

It is pretty clear you already have your mind made up.  So, venting here about DJI software code bugs is pretty pointless.  They are not going to undergo engineering efforts on your individual behalf.  Not when their sales are so strong.  You are operating an expensive toy.  Not a flying machine that undergoes Level A or Level C software review before flight.   If it indeed turns out to be the code, then you best have a savings account for when your drone flies away.
Going back to spectrum sharing once more, which is a realistic factor, you only mention cell towers.   I see in your photo, commercial shipping.   That means weather radar sweeps.   Some ships leave them running in port for long periods of time.  Just watch the antennas turning while walking the docks.  Not to mention radar in aircraft.  Not to mention higher power wi fi systems in business aircraft.   I work on law enforcement helicopters as an instructor for a living.  I often see the N numbers of airplanes on wi fi scans.... even though they are a mile away, on the ground.
The 2.4 ghz spectrum mostly used by DJI (now with some limited use at 5 ghz) is SHARED.   Not assigned to your drone.   And the transmitters operate in Miliwatts.  Not significant power in any sense.  Therefore, the receiver has to be quite sensitive.  So, on a shared spectrum, your bird can easily get swamped by a strong blast of RF energy.
But, as I said... you appear mostly to be just venting at phantoms...    (filters..... delays etc.---- not real solutions)
So many tens of thousands of DJI drones fly world wide, and without flying off un-directed by the operator, that I lean more towards radio interference than software bugs.   The statistics just don't bear out your case.   Anyway, be careful down in that harbor area.  It would be a caution area for me indeed, if my drone had fought me so hard.... to get it back to dry land.  Best of luck and enjoy your flights sir.

CineView Media Posted at 1-14 08:13
I hear you, but i think youre overanalysing this one, and pay to much attention to radio issues.
These incidents have been happening to Dji drones for many years now, and regardless of the area, populated or not.
Ive even experienced it high up in the mountains where there arent even cellphone towers or houses within miles radius.

Also, after all these years with all these flyaway incidents, Dji should have managed to be on top of it by creating filters, delays and other safety measurements ...


What comes to mind is, communication between Drone and RC should involve signed packets.  Interference or garbled packets would be rejected.  Only valid packets going between Drone and RC (which are initially matched) should be accepted, and in case of Drone receiving a command from it's matched RC acted upon.

flightlogicRV9 Posted at 1-15 10:59
It is pretty clear you already have your mind made up.  So, venting here about DJI software code bugs is pretty pointless.  They are not going to undergo engineering efforts on your individual behalf.  Not when their sales are so strong.  You are operating an expensive toy.  Not a flying machine that undergoes Level A or Level C software review before flight.   If it indeed turns out to be the code, then you best have a savings account for when your drone flies away.
Going back to spectrum sharing once more, which is a realistic factor, you only mention cell towers.   I see in your photo, commercial shipping.   That means weather radar sweeps.   Some ships leave them running in port for long periods of time.  Just watch the antennas turning while walking the docks.  Not to mention radar in aircraft.  Not to mention higher power wi fi systems in business aircraft.   I work on law enforcement helicopters as an instructor for a living.  I often see the N numbers of airplanes on wi fi scans.... even though they are a mile away, on the ground.
The 2.4 ghz spectrum mostly used by DJI (now with some limited use at 5 ghz) is SHARED.   Not assigned to your drone.   And the transmitters operate in Miliwatts.  Not significant power in any sense.  Therefore, the receiver has to be quite sensitive.  So, on a shared spectrum, your bird can easily get swamped by a strong blast of RF energy.

I appreciate you sharing your knowledge with us, but that attitude you have, by saying its pointless to come out with our opinions about this, is the very reasons Dji choose to do nothing about it.
There are way too many people with that attitude that prevents things from moving forward, cause why on earth would anyone spend money on extra r&d if they dont need to, right? Its just that simple, a company dont spend more money than they have to, in this case, as you mention, the sales are great anyway.
Once people starts to shut their mouth, youll get a dictating system, where single people are totally ignored.
If EVERYONE with these issues had spoken up about this, it would have affected the sales, and THEN they would have spent the extra money on this.

I dont know how long you been into RC, but if you flew rc before drones came, you would certainly be familiar with Futaba radios, and their shielded system.
Other 2.4GHz systems hold firm to one or two frequencies, increasing the potential for interference. The frequency of Futaba 2.4GHz FASST shifts hundreds of times per second, so there are no signal conflicts or interruptions – and no need for a frequency pin!

Futaba even improved this old system further now, which is called FASSTest.
So yes there ARE a solution this problem, and one of them, is that more people speak up, Sir! ;)

HedgeTrimmer Posted at 1-15 21:27
Also, after all these years with all these flyaway incidents, Dji should have managed to be on top of it by creating filters, delays and other safety measurements ...

Exactly, that Futaba system im talking about...
Other 2.4GHz systems hold firm to one or two frequencies, increasing the potential for interference. The frequency of Futaba 2.4GHz FASST shifts hundreds of times per second, so there are no signal conflicts or interruptions – and no need for a frequency pin!

They have no developed this system further, and called it FASSTest.
Therefore ill say it again, its about time Dji do something about this, after so many years of ignoring this major issue.

CineView Media Posted at 1-16 00:23
I appreciate you sharing your knowledge with us, but that attitude you have, by saying its pointless to come out with our opinions about this, is the very reasons Dji choose to do nothing about it.
There are way too many people with that attitude that prevents things from moving forward, cause why on earth would anyone spend money on extra r&d if they dont need to, right? Its just that simple, a company dont spend more money than they have to, in this case, as you mention, the sales are great anyway.
Once people starts to shut their mouth, youll get a dictating system, where single people are totally ignored.

I think if you are serious about doing something about your problem and you believe it is a problem with the drone rather than envoirment , then to give dji the best chance to figure out what’s wrong and how they can fix it, then it beholds you to send it in, there are many who are new to flying drones who believe that everything is a problem when sometimes it’s only a matter of inexperience, those who are experienced and know there is a problem only by registering that problem can show something needs to be done.
I don’t expect that engineers can see a problem without something to show there is a problem, you have said that you won’t return because you might get a refurbished drone, the truth about that is we have seen very few problems with refurbished drones, apart from it doesn’t sound right.

You are choosing to hold on to a drone that you believe is defective , which is not improving your position, surely if you feel it’s defective and you cannot fix it then you would be better off shipping it back trying to find out what happened properly highlighting what happened and also correctly registering the problem with dji.

I mean are you better with a defective unit or a refurbished or new replacement, what I’m saying in summary is to correctly highlight this problem it’s not enough to just be passing comment while you know you have a defective unit.
I think shipping for examination will be helpful for all.

hallmark007 Posted at 1-16 03:08
I think if you are serious about doing something about your problem and you believe it is a problem with the drone rather than envoirment , then to give dji the best chance to figure out what’s wrong and how they can fix it, then it beholds you to send it in, there are many who are new to flying drones who believe that everything is a problem when sometimes it’s only a matter of inexperience, those who are experienced and know there is a problem only by registering that problem can show something needs to be done.
I don’t expect that engineers can see a problem without something to show there is a problem, you have said that you won’t return because you might get a refurbished drone, the truth about that is we have seen very few problems with refurbished drones, apart from it doesn’t sound right.

I understand what you are saying, and agree with allot of it, however, sending it back will unfortunately not solve anything either, i mean, no one ever got their drone back with a note saying "you were right, we found that there was a problem with your drone, and admit there was a weakness from our side, so here you have a new one"
No, they have never ever once admitted this problem, instead they are so "generous" that they send you a "new one" despite it was nothing wrong with it. Or even worse, youll have to pay a fortune to buy yourself a protection against this, STILL with the same poor terms.
But the problem still remains unsolved, and they keep going on making the same, by not developing a better transmission system, like the rest of the rc business have proven to do ages ago.

Another problem, is that allot of us use it for work, and ship it away, and wait 4-6 weeks before you get it back, is another problem on top of the other.
I dont find it reasonable to buy a second unit just to keep one as a backup when the other one is at service for a month, its still allot of money and it simply doesnt justify the cost, but maybe thats their intention, finding the balance between how much the customers are willing to accept, and then speculate and hope they do just that, buy a second unit for backup, cause most of the customers accept, just that.
In that case, we as customers are shooting ourselves in the leg by doing so.
Why would we ever accept poor service or poor quality, IF we can chose the opposite?
That doesnt make  any sense to any normal person.
If we accept poor quality, by paying a fortune, we will of course continue to get it, if not, they will be forced to do otherwise.

There are of course many many user errors ut there, but im not referring to those, but to the real issues. like this one, which in fact there are still way to many incidents to just burry it and not care...

CineView Media Posted at 1-16 04:57
I understand what you are saying, and agree with allot of it, however, sending it back will unfortunately not solve anything either, i mean, no one ever got their drone back with a note saying "you were right, we found that there was a problem with your drone, and admit there was a weakness from our side, so here you have a new one"
No, they have never ever once admitted this problem, instead they are so "generous" that they send you a "new one" despite it was nothing wrong with it. Or even worse, youll have to pay a fortune to buy yourself a protection against this, STILL with the same poor terms.
But the problem still remains unsolved, and they keep going on making the same, by not developing a better transmission system, like the rest of the rc business have proven to do ages ago.

You know the drone is yours to do whatever you want, but first to use a defective drone for professional purposes is not the best business practice, having used repair service twice waiting time from door to door was 10 working days with no problems, on both occasions I asked and was given explanation for what occurred , I know we read many around here about problems with service but I think this is because if you have a problem it’s here you come, but service has improved greatly from years gone by.

As regards transmission we have seen much improved transmission with the introduction of Ocusync 2 and generally improvements with stability in dji craft, I don’t believe they are any worse than other drones out there, they’re not perfect but are improving.

If you break down your problem read what other users say here on your thread, you will notice many different opinions, so does this mean we don’t know what happened to your craft or that there are many different problems, in order to fix something we must know what’s broke.

I think it’s fair to say that we read of very few flyaways compass problems etc on this forum for Mavic2, comparing with the problems of orignal Mavic these forums were lit up with all sorts of problems then, you may notice that big improvements have been made with problem craft and in particular the M2, so I think that tech and awareness have helped greatly over the last two years. Yes there are some who have problems and yes they should come here to make dji and other members aware of the problems, but you will notice from this thread that there are many differing opinions, which sometimes can only confuse things.
I find myself if I have a problem it’s much easier to try get the opinions from dji repair engineers and relay the information to whom ever is looking for it.

I hope the problem desists and your craft won’t have any more problems, but if it does then I for one would send in and get the information and analysis of what the problem is.

CineView Media Posted at 1-16 04:57
I understand what you are saying, and agree with allot of it, however, sending it back will unfortunately not solve anything either, i mean, no one ever got their drone back with a note saying "you were right, we found that there was a problem with your drone, and admit there was a weakness from our side, so here you have a new one"
No, they have never ever once admitted this problem, instead they are so "generous" that they send you a "new one" despite it was nothing wrong with it. Or even worse, youll have to pay a fortune to buy yourself a protection against this, STILL with the same poor terms.
But the problem still remains unsolved, and they keep going on making the same, by not developing a better transmission system, like the rest of the rc business have proven to do ages ago.

... however, sending it back will unfortunately not solve anything either, i mean, no one ever got their drone back with a note saying "you were right, we found that there was a problem with your drone, and admit there was a weakness from our side, so here you have a new one"

Can't speak for others, but I tend to agree with you.  I have never got any note back.  Only DJI's work bill.


Despite having returned:
1) CrystalSky Ultra Brite - three times

2) Mavic Pro Platinum - two times

3) Remote Control for MPp - two times
Some of returns were new or refurbished product vs. repaired.

{{{{{{So yes there ARE a solution this problem, and one of them, is that more people speak up, Sir! ;)}}}}
Well, first officer..... good luck with that.   Be sure and share the news flash when DJI writes you personally in Norway and says, "thanks to you.... we fixed our faulty software."

CineView Media Posted at 1-16 04:57
I understand what you are saying, and agree with allot of it, however, sending it back will unfortunately not solve anything either, i mean, no one ever got their drone back with a note saying "you were right, we found that there was a problem with your drone, and admit there was a weakness from our side, so here you have a new one"
No, they have never ever once admitted this problem, instead they are so "generous" that they send you a "new one" despite it was nothing wrong with it. Or even worse, youll have to pay a fortune to buy yourself a protection against this, STILL with the same poor terms.
But the problem still remains unsolved, and they keep going on making the same, by not developing a better transmission system, like the rest of the rc business have proven to do ages ago.

I dont find it reasonable to buy a second unit just to keep one as a backup when the other one is at service for a month, its still allot of money and it simply doesnt justify the cost,

If you were to buy a second unit, at very least I would suggest a different model, for example: Phantom-4.  Reason being...

If you have two of exact same models:
1) should you run into a flight problem due to design, or

2) should DJI roll out firmware that causes problems (and can't be rolled back),
then both are grounded.

If I were flying drones as a business, I would have two different model drones.  Quite possibly from different manufactures.  Risk of flawed firmware crossing brands is very unlikely.