Mavic 2 Pro Flyaway

[BudWalker]

I don't believe that this is the case. It would require enormous storage in the drone. And the field is constantly (Subtly) changing. It would need the GPS coordinates to then define the location on the magnetic map. But if it has the GPS coordinates then it can determine the Declination without a map.

If it was going to determine declination it would have to do it Via the Home point GPS coordinates set procedure or the compass calibration procedure. It needs an event to trigger the calculation. My guess would be that it only does this when a compass calibration is done. It would then be part of an overall compass procedure module in the code. At least that is they way I would write it.

It would require many CPU cycles to do this on the fly. And this would mean that all compass calibrations are meaningless. The CPU is way too busy doing other stuff.

Also the most precious resources in the FC is CPU cycles and fast executable memory. Executable storage is very precious. And CPU cycles generate heat. The developers have to keep their FW code as small as they can get it. All functionality is a decision based on priorities.

Without digging into the code it is pretty much conjecture.

The descriptions I gave relate to the actually technology not necessarily directly related to the implementation in a particular flight controller. My litmus test is what does the Manufacturer recommend.

Cheers.

p.s. I only went to this level of detail in the post reply as the poster I responded to used the same words as an introduction to a post that was very rude and abusive on another forum.

It's mathematically impossible for declination to be determined by the compass calibration procedure.

There is pretty compelling evidence that the DJI drones, P3 and after, compute declination based on GPS coordinates. The computation is actually pretty simple and doesn't consume too many resources. DatCon uses this algorithm http://www.ngdc.noaa.gov/seg/potfld/DoDWMM.html
When the AC is powered up the Yaw value is set to the value seen by the compass (magnetometers). When enough satellites are seen the declination is computed and Yaw is adjusted accordingly.

I normally fly at home where DatCon computes the geoDeclination to be 12.72 degrees. On start up the Yaw is set to a value computed from the magnetometers. Then, when gpsHealth gets to 4 (out of 5) that Yaw value is then adjusted to reflect the geoDeclination. That's happened at time -9.512

Without re-calibrating I flew at another location 1200 miles distant where the GeoDeclination is 8.61. It can be seen that the Yaw adjustment reflects that of the new location.

 

[Meta4]

If there is too much electricity in the air or kp index is above 5, DON'T fly. Prevention better than cure. If you are using outside, non DJI app, switch from P mode on controller to S mode and back to P mode and press Home on controller. Exit the app. Reboot the app. If able, will connect to UAV. Press Home in outside app. If both fail, See the first statement on conditions before flying.

Current Kp-index - Homepage of Michael Theusner

The K-index hype is a myth.
We're still waiting for the first confirmed case of it having affected a drone.
Think about it ... if it was true, there would be thousands of drones going out of control every time the K-index gets up.

Or here's something more detailed from someone that knows what he's talking about when it comes to the K-index:
Solar Flares

 

[Flying Dragon22]

This is great information. Thank you fellow pilots.

 

[sar104]

I don't believe that this is the case. It would require enormous storage in the drone. And the field is constantly (Subtly) changing. It would need the GPS coordinates to then define the location on the magnetic map. But if it has the GPS coordinates then it can determine the Declination without a map.

If it was going to determine declination it would have to do it Via the Home point GPS coordinates set procedure or the compass calibration procedure. It needs an event to trigger the calculation. My guess would be that it only does this when a compass calibration is done. It would then be part of an overall compass procedure module in the code. At least that is they way I would write it.

It would require many CPU cycles to do this on the fly. And this would mean that all compass calibrations are meaningless. The CPU is way too busy doing other stuff.

Also the most precious resources in the FC is CPU cycles and fast executable memory. Executable storage is very precious. And CPU cycles generate heat. The developers have to keep their FW code as small as they can get it. All functionality is a decision based on priorities.

Without digging into the code it is pretty much conjecture.

The descriptions I gave relate to the actually technology not necessarily directly related to the implementation in a particular flight controller. My litmus test is what does the Manufacturer recommend.

Cheers.

p.s. I only went to this level of detail in the post reply as the poster I responded to used the same words as an introduction to a post that was very rude and abusive on another forum.

No - it definitely is the case. And, as pointed out by @BudWalker in post #81, global geomagnetic models are commonplace and do not require much storage at all. There is no significant change in declination during a flight since the drone doesn't fly far enough and the calculation is done once, on startup once it has a GPS location. The process is recorded in the DAT event log.

 

[Static8]

So I couldn't resist the Mavic 2 Pro. It's my first drone.

For the five days I had it, I was impressed by the incredible piece of technology it was and excited about the new photographic possibilities.

I was flying on 9/11. Took off from the front of my house. When I connected the controller, it said there was mag interference, but the message went away. The drone established a home point and I took off, climbed to 150 feet (100 feet above the highest obstacle for about a half mile radius) for about a minute and a half, directly above my position, and looked at the sunset and lights. Everything seemed fine. At about 1:45, I turned and flew forward about 300 feet, and the drone just disappeared. No intermittent signal. It just totally dropped.

I immediately paused control inputs and waited about a minute, but it may not have been that long. Thinking the RC may still be connected i attempted to ascend to see if I could regain connection, and tried to get input on the map view, but it was not responding. Initiated RTH command, but don't believe it was ever received by the drone. The controller was continuously beeping and displaying "connecting..." Thinking the failsafe would kick in and it would come home, I waited about 5 min. Nothing. Walked directly under the last known point, still attempting to connect. Nothing. Took controller and drove in direction of flight path. Nothing.

Someone waited at the launch point in case it returned, but it did not.

Not being that experienced, it feels like a total in flight power failure, but I was hoping for some feedback.

DJI is reviewing records now. I attached the flight log.

Any advice appreciated.

Lincoln

THis is exactly what happened to my Mavic Pro right after new firmware update back in July...since I was passed warranty by a few months, DJI did nothing...I lost my drone in the ocean..and I have been flying for 4 years and all DJI products, they woudln’t give me the time of day...Its a firmware issue I can gurantee you this...more will drop , flyaway until they resolve the issue...Until then, I will not buy another DJI product...its a crap shoot

 

[hherbson]

Hahaha. You are a good person for defending me against me! I’m the OP. A little self-deprecation to admit how dumb it was of me not to put it on there. Oddly, didn’t even cross my mind.

Haha, I should have paid more attention. Oh well, hope this doesn't discourage you from getting another one, and don't be so hard on yourself. Who would have expected that to happen?

 

[Anndybo]

Sar104,

Your trajectory was dead on. Neighbor recovered it in his back yard exactly where you pointed.

Thank you so much.

Not sure if I should continue my warranty claim or use DJI Care Refresh as it seems like it did just fall out of the sky. Totally busted up.

Thanks to everyone else who helped with advice as well.

Great you found it!

 

[METIOR]

So I couldn't resist the Mavic 2 Pro. It's my first drone.

For the five days I had it, I was impressed by the incredible piece of technology it was and excited about the new photographic possibilities.

I was flying on 9/11. Took off from the front of my house. When I connected the controller, it said there was mag interference, but the message went away. The drone established a home point and I took off, climbed to 150 feet (100 feet above the highest obstacle for about a half mile radius) for about a minute and a half, directly above my position, and looked at the sunset and lights. Everything seemed fine. At about 1:45, I turned and flew forward about 300 feet, and the drone just disappeared. No intermittent signal. It just totally dropped.

I immediately paused control inputs and waited about a minute, but it may not have been that long. Thinking the RC may still be connected i attempted to ascend to see if I could regain connection, and tried to get input on the map view, but it was not responding. Initiated RTH command, but don't believe it was ever received by the drone. The controller was continuously beeping and displaying "connecting..." Thinking the failsafe would kick in and it would come home, I waited about 5 min. Nothing. Walked directly under the last known point, still attempting to connect. Nothing. Took controller and drove in direction of flight path. Nothing.

Someone waited at the launch point in case it returned, but it did not.

Not being that experienced, it feels like a total in flight power failure, but I was hoping for some feedback.

DJI is reviewing records now. I attached the flight log.

Any advice appreciated.

Lincoln

Алексей. Неудачное Вы место для взлета выбрали. Лежит он где нибудь на крыше разбитый. Может вы батарею плохо зафиксировали, она и вылетела.

 

[DanMan32]

OK, so if the AC uses the compass mainly at startup, that would explain reported issues taking off with compass error that otherwise clears once off the ground.
My P3 won't even start with a compass error, and I know it is because of where it is rather than a bad or miscalibrated compass. Now I understand why it won't even start.

 

[sar104]

OK, so if the AC uses the compass mainly at startup, that would explain reported issues taking off with compass error that otherwise clears once off the ground.
My P3 won't even start with a compass error, and I know it is because of where it is rather than a bad or miscalibrated compass. Now I understand why it won't even start.

It's generally not going to clear up if the initial heading is incorrectly initialized. And while the compass is used for that at startup it is also essential to the sensor fusion scheme during flight to correct for rate gyro drift - without the compass the FC is not able to navigate or hold position.

 

[Cheech Wizard]

Sar104,

Your trajectory was dead on. Neighbor recovered it in his back yard exactly where you pointed.

Thank you so much.

Not sure if I should continue my warranty claim or use DJI Care Refresh as it seems like it did just fall out of the sky. Totally busted up.

Thanks to everyone else who helped with advice as well.

I Love a Happy Ending, that's cool you got it back. @sar104 is a drone Guru

 

[Eraser339]

Wait. What. No. It's not over. Can you upload the flight info from the DRONE. We have not determined if it is a signal loss or power fail

 

[Meta4]

We have not determined if it is a signal loss or power fail

Signal loss doesn't bring a drone down to earth until it completes RTH.

 

[Eraser339]

Met4 the data analysed from the phone stopped when there was a signal loss. If there is data from the DRONE after the signal loss to the RC/Phone then it wasn't a power failure. Did it reset the home point and just land. Probably not as the AC was wrecked. Some have reported a shutdown due to high battery Temps. Is there anymore info from the DRONE.

 

[Meta4]

Did it reset the home point and just land.

What could reset the home point after the signal is lost?
That's beyond unlikely.

 

[sar104]

Met4 the data analysed from the phone stopped when there was a signal loss. If there is data from the DRONE after the signal loss to the RC/Phone then it wasn't a power failure. Did it reset the home point and just land. Probably not as the AC was wrecked. Some have reported a shutdown due to high battery Temps. Is there anymore info from the DRONE.

There have been several cases like this, where the log just stops without any prior indication of problems. It does appear that the M2 may have a problem, possibly in the firmware, that infrequently and somewhat randomly crashes the FC, but no solid data yet.

The fast record in the aircraft DAT file would probably shed some more light on this, but those files are encrypted on the M2 and not currently readable by DatCon.

 

[BudWalker]

Met4 the data analysed from the phone stopped when there was a signal loss. If there is data from the DRONE after the signal loss to the RC/Phone then it wasn't a power failure. Did it reset the home point and just land. Probably not as the AC was wrecked. Some have reported a shutdown due to high battery Temps. Is there anymore info from the DRONE.

The Mavic 2 .DAT is encrypted and can not be decoded. This is also true of the Mavic Air. The only .DAT that can be decoded resides on the tablet.

 

[Eraser339]

Thanks Sar. I now remember you saying that in a previous thread. Do we know what type of encryption. RSA? if we could find the certificate we may decode it. Can the data from the fc be downloaded to the phone. Hence decrypted. Eg. Do a flight turn RC off(or just the phone). After the AC lands play the flight and see if you any data of the return to home.

 

[sar104]

Thanks Sar. I now remember you saying that in a previous thread. Do we know what type of encryption. RSA? if we could find the certificate we may decode it. Can the data from the fc be downloaded to the phone. Hence decrypted. Eg. Do a flight turn RC off(or just the phone). After the AC lands play the flight and see if you any data of the return to home.

@BudWalker is the expert on DAT files and encoding schemes. I think that he is looking into the possibilities to read them.

 

[ChopperLinc]

Wait. What. No. It's not over. Can you upload the flight info from the DRONE. We have not determined if it is a signal loss or power fail

My inexperienced opinion is that the aircraft suffered a power loss, although the FC crash is a possibility. Here’s why I think it lost power:

1). Sar104 calculated the trajectory based on it vector when the signal was lost and I found the drone literally in the exact spot indicated by his calcs.
2). If the FC crashed and the drone reset home point and landed, I think I would have seen it hovering. I was flying at night and have found it to be much more visible with the lights than in the daytime, and it was only 300 feet away. It was immediately invisible.
3). Signal loss seems extremely unlikely. In the short time I had it I did some non-scientific range tests. In one instance I was transmitting from between two high tension power transmission lines and the drone was a mile and half away. When I lost sight of it for a a few seconds, the video transmission became choppy. I increased altitude and it returned. You all have flown much longer than me, but signal loss seems to be gradual, not immediate and catastrophic.

I guess the FC crash could be a culprit, and the drone could have attempted to land, crash through some trees and happened to fall on the exact spot it would have if it lost power and fell from the last know height and speed...

FYI. When I recovered it, I replaced the battery which had fallen in the pool. Everything was all broken and banged up, include large parts of propeller missing...and it still flew perfectly (which I did for about 30 seconds in front of me in my living room.)