Mavic 3 flyaway

GadgetGuy · Jul 18, 2022

slup said:
No...according to the log it started to drift sideways.

Again... not according to the log.

Unfortunately very typical in cases like this...

The reason for this incident was very likely a yaw error coming from powering on the drone in a magnetically interfered spot... that caused the IMU yaw to be wrongly initialized, plain and simple, when the drone took off it didn't have a true heading direction (look at the attached picture)

This made the flight controller to use the wrong motors to keep the position when the drone started to be pushed by the wind... it looks like the yaw error was about 90 degrees which usually cause a curved flyaway path like what we see here.

This isn't a hardware error...it's a pilot error. In the past it's been a bit of a hit & miss from DJI in cases like this, you might be lucky.

(Click on the picture to make it larger)
View attachment 152014

It's been so long since seeing one of these dreaded J hooks that the P3P was famous for. Had one happen myself! The preventive solution is always verify the drone heading on the map corresponds to the actual heading on the map before launching, to eliminate a bad magnetic calibration at boot up. Inevitably, it is caused by hidden metal underneath the drone at the launch site. The compass often detects it, but it is clearly not a guarantee! Pro Tip: Verify your drone heading, as displayed on the map, is accurate before taking off! This one is clearly off, for a drone aimed at and facing the waterfall!

AKAurora · Jul 18, 2022

GadgetGuy said:
It's been so long since seeing one of these dreaded J hooks that the P3P was famous for. Had one happen myself! The preventive solution is always verify the drone heading on the map corresponds to the actual heading on the map before launching, to eliminate a bad magnetic calibration at boot up. Inevitably, it is caused by hidden metal underneath the drone at the launch site. The compass often detects it, but it is clearly not a guarantee! Pro Tip: Verify your drone heading, as displayed on the map, is accurate before taking off! This one is clearly off, for a drone aimed at and facing the waterfall!

Excellent advice. I hadn't even considered that! I'll add it to my checklist and preflight

GadgetGuy · Jul 18, 2022

AKAurora said:
Excellent advice. I hadn't even considered that! I'll add it to my checklist and preflight

What surface did you launch from that might have had hidden metal underneath? You still can try blaming it on DJI for not detecting the metal and allowing it to set an incorrect compass direction. Good luck!

AKAurora · Jul 18, 2022

GadgetGuy said:
What surface did you launch from that might have had hidden metal underneath? You still can try blaming it on DJI for not detecting the metal and allowing it to set an incorrect compass direction. Good luck!

I launched from a foldout pad on the parking lot/area. I'm not blaming DJI, but want to understand [Phrase removed by ADMIN] happened. I already replaced the drone. Not happy about it, but now I have care refresh. Expensive lessons!

FLDave · Jul 18, 2022

I still contend that the data indicates high gusty wind at altitude. Cliffs can do strange things to wind no felt 400 feet below.
Have you ever watched the sudden attitude changes in gust when the drone is clearly visible at 50 feet above? It's incredible to see the Air 2s tilt just to maintain position. The data indicates that happened here. I don't think the pilot could visually see that at 500 feet away and 200-400 feet up. I wouldn't be able to visually determine the heading at those distances.
If the attitude changes were not needed then the drone would have flow an erratic path based on the speed from the tilts. Yet the GPS track is smooth indicating the attitude changes were the compensate for wind to maintain stability.
I'm guessing that the plan was to fly directly to the falls yet the GPS data shows a curve; that leads me to believe it was a wind issue.
One can choose to believe the data or not. If you are not going to believe the data, that what do you trust?

FLDave · Jul 18, 2022

AKAurora said:
I launched from a foldout pad on the parking lot/area. I'm not blaming DJI, but want to understand wtf happened. I already replaced the drone. Not happy about it, but now I have care refresh. Expensive lessons!

I had a pilot error crash with my Air 2S; it wasn't a lost drone so the CareRefresh wasn't so expensive, but well worth it. It was still a lesson learned about flying into trees.

AKAurora · Jul 18, 2022

FLDave said:
I still contend that the data indicates high gusty wind at altitude. Cliffs can do strange things to wind no felt 400 feet below.
Have you ever watched the sudden attitude changes in gust when the drone is clearly visible at 50 feet above? It's incredible to see the Air 2s tilt just to maintain position. The data indicates that happened here. I don't think the pilot could visually see that at 500 feet away and 200-400 feet up. I wouldn't be able to visually determine the heading at those distances.
If the attitude changes were not needed then the drone would have flow an erratic path based on the speed from the tilts. Yet the GPS track is smooth indicating the attitude changes were the compensate for wind to maintain stability.
I'm guessing that the plan was to fly directly to the falls yet the GPS data shows a curve; that leads me to believe it was a wind issue.
One can choose to believe the data or not. If you are not going to believe the data, that what do you trust?

It wasn't 500 feet away from me. It was straight above me at 300ft agl. The fact that a guest hits it and I get a wind warning at home but as soon as this started going totally bonkers and 0 warnings or response (even to command to decend), tells me it was not wind. Uav forecast had wind up to 1k feet agl at 6mph to 9mph. And after the drone dipped down it continued to accelerate in a j hook motion. Even after it dropped down to near tree level where I was.

Phantomrain.org · Jul 18, 2022

AKAurora said:
I launched from a foldout pad on the parking lot/area. I'm not blaming DJI, but want to understand wtf happened. I already replaced the drone. Not happy about it, but now I have care refresh. Expensive lessons!

So magnetic interference can happen really at any time, from taking off around Metal , Concrete < Power Wires , and was very Prevalent with the Phantom 3 and Phantom 4 Drones, I had two Phantom 4 Replaced by DJI from Fly always do to Magnetic interference . They offered full replacement coupons.

Now the Mavic drones have been so much better but not immune. , and I pretty confident that they will work with you on this situation as well.

Two things help with negotiating this . One you know already and that is
making sure the Compass is pointing the right direction that your facing,
but also trying to minimize Magnetic Interference by taking off in the grass ,and not like from a Pier or Parking lot although we do get away with a lot.

If the drone goes into atti mode , DJI will look to see if you tried to use the stick commands to get it back,
other times they will understand that stick movements had no effect on the drone and that is when your likely to get a full replacement from DJi, its going to take a lot of emails and they may ask you for the video as well.

I think this will work in your favor

Phantomrain.org
Gear to fly in the Rain. Land on the Water ,

AKAurora · Jul 18, 2022

Phantomrain.org said:
So magnetic interference can happen really at any time, from taking off around Metal , Concrete < Power Wires , and was very Prevalent with the Phantom 3 and Phantom 4 Drones, I had two Phantom 4 Replaced by DJI from Fly always do to Magnetic interference . They offered full replacement coupons.

Now the Mavic drones have been so much better but not immune. , and I pretty confident that they will work with you on this situation as well.

Two things help with negotiating this . One you know already and that is
making sure the Compass is pointing the right direction ,
but also trying to minimize by taking off in the grass ,and not like from a Pier or Parking lot although we do get away with a lot.

If the drone goes into atti mode , DJI will look to see if you tried to use the stick commands to get it back,
other times they will understand that stick movements had no effect on the drone and that is when your likely to get a full replacement from DJi, its going to take a lot of emails and they may ask you for the video as well.

I think this will work in your favor

Phantomrain.org
Gear to fly in the Rain. Land on the Water ,

Here is another head scratcher. I have no video. My cache is a 36m file that is corrupt and I can't get it to view or recover anywhere (I'm a systems engineer, so I have and use the tools). Another one to make you go "hmmm"

Phantomrain.org · Jul 18, 2022

AKAurora said:
Here is another head scratcher. I have no video. My cache is a 36m file that is corrupt and I can't get it to view or recover anywhere (I'm a systems engineer, so I have and use the tools). Another one to make you go "hmmm"

Are you sure its the right file. ? Have you posted the video before from the Cashe. ?

AKAurora · Jul 18, 2022

Phantomrain.org said:
Are you sure its the right file. ? Have you posted the video before from the Cashe. ?

Absolutely positive. Location of file and date/time match (only flight of the day). Didn't even show in the app, I had to dig it out manually

Phantomrain.org · Jul 18, 2022

AKAurora said:
Absolutely positive. Location of file and date/time match (only flight of the day). Didn't even show in the app, I had to dig it out manually

That is what is bothering me that it did not show on the App, is is possible you just did not record ?

AKAurora · Jul 18, 2022

Phantomrain.org said:
That is what is bothering me that it did not show on the App, is is possible you just did not record ?

No, I'm certain it was recording. Otherwise a file wouldn't have even been created in the cache. That's what I mean, there is a lot of weirdness going on

Phantomrain.org · Jul 18, 2022

AKAurora said:
No, I'm certain it was recording. Otherwise a file wouldn't have even been created in the cache. That's what I mean, there is a lot of weirdness going on

Upload the file to google drive and There is a Fixer we can put it thru, that we have used in the past, It goes thru frame by frame and pull them out. Time consuming yes but it works, Lets see what on that file

Phanmtomrain.org
Gear to fly in the Rain.

BigAl07 · Jul 18, 2022

AKAurora said:
No, I'm certain it was recording. Otherwise a file wouldn't have even been created in the cache. That's what I mean, there is a lot of weirdness going on

Could it be the file wasn't "Closed" properly? That's common when a sudden power failure occurs etc and the file is not correctly finalized. There's a thread here for a free and easy tool that fixes this many many times.

AKAurora · Jul 18, 2022

BigAl07 said:
Could it be the file wasn't "Closed" properly? That's common when a sudden power failure occurs etc and the file is not correctly finalized. There's a thread here for a free and easy tool that fixes this many many times.

I've tried it. To no avail.

2022_07_16_17_08_10_Cache.mp4

drive.google.com

There it is, if anyone wants to take a crack at it, I'd be thrilled if anything came of it.

MS Coast · Jul 18, 2022

FLDave said:
I still contend that the data indicates high gusty wind at altitude. Cliffs can do strange things to wind no felt 400 feet below.
Have you ever watched the sudden attitude changes in gust when the drone is clearly visible at 50 feet above? It's incredible to see the Air 2s tilt just to maintain position. The data indicates that happened here. I don't think the pilot could visually see that at 500 feet away and 200-400 feet up. I wouldn't be able to visually determine the heading at those distances.
If the attitude changes were not needed then the drone would have flow an erratic path based on the speed from the tilts. Yet the GPS track is smooth indicating the attitude changes were the compensate for wind to maintain stability.
I'm guessing that the plan was to fly directly to the falls yet the GPS data shows a curve; that leads me to believe it was a wind issue.
One can choose to believe the data or not. If you are not going to believe the data, that what do you trust?

Where are you finding data on wind speeds and gusts to believe?

Have you read post #40? The behavior of the aircraft can be explained by factors other than wind.

GadgetGuy · Jul 18, 2022

AKAurora said:
I launched from a foldout pad on the parking lot/area. I'm not blaming DJI, but want to understand [Phrase removed by ADMIN] happened. I already replaced the drone. Not happy about it, but now I have care refresh. Expensive lessons!

Don't give up still pursuing replacement of the original M3 by DJI. Open a warranty replacement case immediately. The log file that you have and @slup 's analysis of it with the screenshot in Post #40 above is sufficient to establish that undetected magnetic interference caused the drone to set a bad compass calibration before launch. The drone is supposed to detect that, and display red notification screen, preventing your launching. It didn't. That is your argument. Simple and straight forward. This is the classic J hook from a bad compass calibration at launch, that DJI did not detect. Keep escalating through managers as necessary to get them to accept responsibility. Worst case, they still deny warranty coverage. Ideally, they replace the drone. Keep us posted.

Phantomrain.org · Jul 18, 2022

AKAurora said:
I've tried it. To no avail.

2022_07_16_17_08_10_Cache.mp4

drive.google.com

There it is, if anyone wants to take a crack at it, I'd be thrilled if anything came of it.