fkindome Posted at 2017-5-16 14:32
I'm beginning to think this erratic behavior of the Mavic when it complains about  "Magnetic Field Interference. Exit P-GPS Mode" is a BUG!
The message leads to a flyaway of the AC and I think DJI (if someone is reading) should seriously look into it and provide a patch. One possible fix is for the AC to activate and auto RTH similar to cases where it loses signal.

If your aircraft looses gps it can't return home because it doesn't know where it is...

I took a look at fkindome's .txt. IMHO it's a lot like fans18d204c5's incident. I.e., not a launch-from-geomagnetically-distorted-site incident. Further, the "compass error" occurred at 50 meters above the launch site making it unlikely the cause was any kind of "magnetic interference". If fkindome would retrieve and Dropbox the .DAT we could know more.

One question though. Are you guys running .0700 firmware? The reason I ask is that the behavior is reminiscent of the P3 behavior. I.e, a more sensitive compass error detection algorithm which leads to false positives.

I got the same warning saturday.  Out in the open, 25 meters high, .650 sw.  19 GPS sats falling to no signal, "Current RTH will pass a no-fly zone"  , then  "magnetic field interferance, Exit P-gps mode" and the Mavic drifting to the left while me trying to gain control by giving full right stick right, to no avail. Drifted for 10 seconds or so, then GPS was re-established and i got control and flew back.  I later (on video) discovered a metal hunting tower not far away from where I flew, did'nt got around to investigate it, but I would guess such contraptions are made of non-magnetic lightweight alu rather than steel, so I do'nt know if this is the explanation.  I'm on .700 now, so perhaps this is not so important to investigate.

Kiowa Posted at 2017-5-16 16:39
I got the same warning saturday.  Out in the open, 25 meters high, .650 sw.  19 GPS sats falling to no signal, "Current RTH will pass a no-fly zone"  , then  "magnetic field interferance, Exit P-gps mode" and the Mavic drifting to the left while me trying to gain control by giving full right stick right, to no avail. Drifted for 10 seconds or so, then GPS was re-established and i got control and flew back.  I later (on video) discovered a metal hunting tower not far away from where I flew, did'nt got around to investigate it, but I would guess such contraptions are made of non-magnetic lightweight alu rather than steel, so I do'nt know if this is the explanation.  I'm on .700 now, so perhaps this is not so important to investigate.

Sorry, should read .600 sw release.

Here is my DAT file via DropBox: https://www.dropbox.com/s/9zwnw7 ... 7_00-10-03.DAT?dl=0  
I'm running the latest firmware .700 which I installed on May 9th.
Thanks to you all.

fkindome Posted at 2017-5-16 21:32
Here is my DAT file via DropBox: https://www.dropbox.com/s/9zwnw7 ... 7_00-10-03.DAT?dl=0  
I'm running the latest firmware .700 which I installed on May 9th.
Thanks to you all.

I have forwarded your flight data to our engineers for review.
If we need any further information or have any update, we will let you know. Thank you for your feedback.

DJI Mindy Posted at 2017-5-16 22:50
I have forwarded your flight data to our engineers for review.
If we need any further information or have any update, we will let you know. Thank you for your feedback.

Mindy, I provided my info as well, but just in case:

kds1041@yahoo.com   5/13/2017

fans18d204c5 Posted at 2017-5-17 06:36
Mindy, I provided my info as well, but just in case:

   5/13/2017

Let me know if you need the .dat file

fkindome Posted at 2017-5-16 21:32
Here is my DAT file via DropBox: https://www.dropbox.com/s/9zwnw7 ... 7_00-10-03.DAT?dl=0  
I'm running the latest firmware .700 which I installed on May 9th.
Thanks to you all.

I looked at this .DAT. In summary, it seems the "incident" was caused by a combination of high winds (possibly turbulent) in conjunction with abrupt control inputs (mostly aileron). The Flight Controller couldn't keep up, decided it was confused and switched to ATTI - essentially saying to the pilot fly this thing. This is exactly what it should do. It could be argued that the FC should not have to give up and switch to ATTI. Better to recognize a problem and switch.

At time 245 the Mavic was at 52 meters altitude with all 4 motorSpeeds relatively equal. Then with no elevator or aileron inputs the back motorSpeeds decreased indicating the wind had increased from the SW. Then at time 317 elevator and aileron inputs were applied. magYaw (computed from the magnetometers) oscillated but the Yaw (derived from the IMU, mostly gyros) remained stable. Finally, at time 331 the switch to ATTI is made.

I know the OP stated the Mavic wasn't responding to inputs but the data indicates that was responding. It was drifting with the wind in ATTI. Maybe it just seemed unresponsive. Essentially, the pilot and the wind were both controlling the Mavic. There were several short control inputs after the switch to ATTI. I'll speculate that, wtih the Mavic drifting in the high wind, none were long enough to observe the effects.


If DJI provides their findings for this "incident" please post them

BudWalker Posted at 2017-5-17 06:48
I looked at this .DAT. In summary, it seems the "incident" was caused by a combination of high winds (possibly turbulent) in conjunction with abrupt control inputs (mostly aileron). The Flight Controller couldn't keep up, decided it was confused and switched to ATTI - essentially saying to the pilot fly this thing. This is exactly what it should do. It could be argued that the FC should not have to give up and switch to ATTI. Better to recognize a problem and switch.

At time 245 the Mavic was at 52 meters altitude with all 4 motorSpeeds relatively equal. Then with no elevator or aileron inputs the back motorSpeeds decreased indicating the wind had increased from the SW. Then at time 317 elevator and aileron inputs were applied. magYaw (computed from the magnetometers) oscillated but the Yaw (derived from the IMU, mostly gyros) remained stable. Finally, at time 331 the switch to ATTI is made.

Good job, are you finding any connection to new firmware.

Also if you don't mind are yaw error and speed error connected to compass error or turbulence?

hallmark007 Posted at 2017-5-17 06:52
Good job, are you finding any connection to new firmware.

Also if you don't mind are yaw error and speed error connected to compass error or turbulence?

No connection to firmware that I'm aware of.  Just to emphasize, IMHO, the FC did exactly the right thing. It can't be perfect and will get confused from time to time. The best thing to do is switch to ATTI. There are other situations, in particular the launch-from-geomagnetically-distorted-site situation, where the FC should switch to ATTI but doesn't.

SPEED_ERROR_LARGE, YAW_ERROR_LARGE, COMPASS_ERROR_LARGE are values that the nonGPSCause field can have. All these values mean is that the FC sees a value that it hasn't predicted. E.g., in the launch-from-geomagnetically-distorted-site situation the predicted value of Yaw doesn't match the value seen by the magnetometers.

BudWalker Posted at 2017-5-17 07:08
No connection to firmware that I'm aware of.  Just to emphasize, IMHO, the FC did exactly the right thing. It can't be perfect and will get confused from time to time. The best thing to do is switch to ATTI. There are other situations, in particular the launch-from-geomagnetically-distorted-site situation, where the FC should switch to ATTI but doesn't.

SPEED_ERROR_LARGE, YAW_ERROR_LARGE, COMPASS_ERROR_LARGE are values that the nonGPSCause field can have. All these values mean is that the FC sees a value that it hasn't predicted. E.g., in the launch-from-geomagnetically-distorted-site situation the predicted value of Yaw doesn't match the value seen by the magnetometers.

Thank you for that, if only we had a switch to go to Atti !

BudWalker Posted at 2017-5-17 06:48
I looked at this .DAT. In summary, it seems the "incident" was caused by a combination of high winds (possibly turbulent) in conjunction with abrupt control inputs (mostly aileron). The Flight Controller couldn't keep up, decided it was confused and switched to ATTI - essentially saying to the pilot fly this thing. This is exactly what it should do. It could be argued that the FC should not have to give up and switch to ATTI. Better to recognize a problem and switch.

At time 245 the Mavic was at 52 meters altitude with all 4 motorSpeeds relatively equal. Then with no elevator or aileron inputs the back motorSpeeds decreased indicating the wind had increased from the SW. Then at time 317 elevator and aileron inputs were applied. magYaw (computed from the magnetometers) oscillated but the Yaw (derived from the IMU, mostly gyros) remained stable. Finally, at time 331 the switch to ATTI is made.

BudWalker...thank you very much for you humble analysis...it looks interesting interpreting the numbers...great job! I will surely post any analysis I receive from DJI.

fans18d204c5 Posted at 2017-5-17 06:37
Let me know if you need the .dat file

Yes, please export the flight data via DJI assistant then upload it to dropbox or google drive.
Thank you. Besides, what's your current firmware version?

The bigger question we are not addressing is why it loses gps in the first place, just happened to me. Went into atti on my second flight after getting it back from a 6 week repair. Was windy but all was fine till it lost gps at 40 meters over my house.  (No magnet interference anywhere near by, and upon startup everything was nominal with everything in the green) Wind took it away and novice pilot that I am, I managed to fly it back and land but my hands were shaking.  Have lost all faith in this thing and am afraid to fly it now.  Went from 18 gps locked to nothing in a heartbeat. Don't care what you say, this was a malfunction of the drone!!! P.S. First flight after I updated the firmware.  
DJI needs to address this fast as I will be looking for a replacement if this happens again!

DJI Mindy Posted at 2017-5-17 20:16
Yes, please export the flight data via DJI assistant then upload it to dropbox or google drive.
Thank you. Besides, what's your current firmware version?

I'll try and get the .dat file this weekend. I know I'm running the latest firmware. The issue occurred on my first flight after I updated it.

Hey guys!

Reading this thread I remember a similar situation that I experienced. I have uploaded a video including the flight playback and my analysis.

DJI Mindy Posted at 2017-5-17 20:16
Yes, please export the flight data via DJI assistant then upload it to dropbox or google drive.
Thank you. Besides, what's your current firmware version?

Mindy, current firmware version is: 01.03.0700   Link to Google drive folder with .dat files is below

https://drive.google.com/drive/f ... QTZTajQ?usp=sharing

Sorry for your Exit P-GPS Mode experience. I had a look into the flight data and found there was a disagreement for gyroscope sensor,compass sensor and gps sensor at
the time when tips showed. The craft was confused about the true yaw information,
then it choosed to switch to ATTI mode to
stop the craft  fly-away randomly.Which would be worse than drifting ATTI mode. However, the big wind nearly 4-5m/s made the craft hard to be controlled. The bad compass calibration data or compass interference  or bad gps sensor data are the most reason for  this confusing condition. If this repeats after all kinds of
calibration, you may send it back for a carefully check.

Dronoob Posted at 2017-5-19 23:05
Hey guys!

Reading this thread I remember a similar situation that I experienced. I have uploaded a video including the flight playback and my analysis.

Dronoob...thanks for your video and explanation. The thing with my incident is that my compass was reading correctly during take off. I also watched my flight using the Go4 app and my incident was different from yours. Yours was truely a case where the aircraft was confused at take off. With my incident it was not confused. The confusion started in mid-air during flight when with no stick input the aircraft switched to ATTI mode and drifted away.

wisnia.sebastia Posted at 2017-5-21 18:12
Sorry for your Exit P-GPS Mode experience. I had a look into the flight data and found there was a disagreement for gyroscope sensor,compass sensor and gps sensor at
the time when tips showed. The craft was confused about the true yaw information,
then it choosed to switch to ATTI mode to

Wisnia...thanks for your input...I believe your explanation is more inline with what I suspected happened. After watching the flight in the Go4 app I noticed moments when without any stick inputs the AC displayed "weak gps signal", GPS indicators (bars) were turned off, then the switch to ATTI mode...I'm guessing after the AC reset itself to fix the issue with the GPS signal and that was when it flipped back to GPS mode from ATTI mode.

Also had this when taking off, I can control it, but it still wants to pan when I don't touch control, I is not the wind. Because it is really tilting so I compensate. Till I feel it gets steady. But it is odd.

I find it interesting the exact same thing has happened to multiple owners. I didn't have any issues until I did the most recent firmware update. Could be totally unrelated...who knows?

I had this issue happen to me again this weekend flying in the same general area different take off spot. it only lasted a few seconds. Long enough to get a brief warning on the screen.  

Hi,

I can recall when I upgraded to firmware 0.700 I experienced a small drifting in about 70 meters altitude. It was my first drifting in after 80 flights at the same spot with earlier firmware revisions. Refreshed the firmware and haven't seen any drifting yet.
That day, it was windy, but not so windy that I couldn't fly it. I have flown in situations much more windy and never had a problem like that.

Hmm.  I had something similar happen to me awhile ago, where I went from 4 bars (signal strength) with a similar number of satellites to haywire in a second.  And I've only recently started getting "Magnetic Field Interference. Exit P-GPS Mode" messages, but before I've taken off, so I've moved my takeoff location.

I'm wondering if this is a new warning (and whether it was mag interference that happened with me, despite the fact that I was in a fairly open rural area where magnetic interference was unlikely, but not impossible).  "@DJI Mindy", could you confirm?

I also started to have this problem since a week or so. No matter where I go, after a while in air it occurs intermittently. The device is useless!
My log from recent flight http://www.phantomhelp.com/LogViewer/4EL608USZQFWZXEJKN2A/#

I know what it is.. I think most of us been here know what it is after the shielding work I did. I can fly by cell towers now, sport the whole way no more issues at all. I used to calibrate like crazy, now just pull it out and fly. It's the emf coming off the wiring to the rear motors.

I've seen it fly off in one direction before the mod too. It was with no input to the sticks as I was coming down. What it usually did was spin in toilet bowl circles when the compass was screwed and it wouldn't drop into atti. Now, there's none of that. It's not something you can solve correctly with firmware. That wiring needs to be double shielded.

I fixed the problem by calibrating the compass. Simple as that.

fkindome Posted at 2017-5-15 21:11
Date: May 14, 2017

We apologize for the late reply, please email to support.us@dji.com with your DJI account, they will create ticket for you to make data analysis in warranty.

DroneFlying Posted at 2017-5-15 06:08
I took off from the side walk.

That's what I suspected, and that's probably what caused this to occur. Concrete usually contains (steel) rebar that will create compass problems that in turn lead to behavior like this during flight.

Hi,
I wondering, do you think the metal inside the landing pad (the metal spring) could affect the compass calibration procedure ?

Igordur Posted at 2017-6-5 02:39
Hi,
I wondering, do you think the metal inside the landing pad (the metal spring) could affect the compass calibration procedure ?

That's an interesting question. I suppose it's possible if the landing pad uses steel for the outer ring that keeps it taught  when expanded. I use one myself and I've never had a problem, but then I also put the Mavic more or less in the center of the landing pad for takeoff. But I suppose if it were placed near the outer edge it's possible that there could be enough steel to affect the compass; I'll have to try that the next chance I get and see if I notice any compass deviations.

@droneflying, I will interested by yours conclusions.
Now, my Mavic went to repair for a fuzzy image problem...

Hi,

One question, do you was in non fly zone?
Because i also see this message: Current RTH route will pass a No-Fly zone. Pay attention to the aircraft's position to avoid RTH failure.

Hello all. Been droning for a couple of years now. I own P3A, no mods, bog standart.

Ok, recently i rediscovered litchi autonomous waypoint mode. And i've been having so much fun out of it. But.
Both of the last wp flights i took ended in yaw error in logs that immediately led to Atti mode, but as AC was out of RC reach, it autotriggered RTH.
Both times.
I'm no pro at log analysis, but i have my dat files (i am willing to upload those whereever). I looked closely at first occurrence and it totally resembles what happens to people in this thread.
On video only a sliiiight wobble can be seen (at least visually) before RTH triggers. Not a bird.

I also have videos of both incidents uploaded to yt, can be correlated with .dats i have.

Would love some pro insight into those. Tnx in advance.

Last f/w 1.11.20.

I have also been getting this error, link-> Compass error and got no respond from DJI regarding this issue. This is most certainly some sort of manufacture issue that is not being answered by DJI, as they provide no answers or solution. What I can see by other post the first solution is always "user error “I have lost respect for DJI prod and they need to act on this if they want to avoid liability and negative product respond. There are so many posts about same issue that goes ignored.

  

  

original link was with different email address, I had to change it when I registered Mavic to new email account.

  

Dronulator Posted at 2017-7-26 10:09
Hello all. Been droning for a couple of years now. I own P3A, no mods, bog standart.

Ok, recently i rediscovered litchi autonomous waypoint mode. And i've been having so much fun out of it. But.

If you upload the DAT file from a flight where this occurred I'll take a look and see if I can figure out what might have happened.

Attila_Ice Posted at 2017-7-26 11:45
I have also been getting this error, link-> Compass error and got no respond from DJI regarding this issue. This is most certainly some sort of manufacture issue that is not being answered by DJI, as they provide no answers or solution. What I can see by other post the first solution is always "user error “I have lost respect for DJI prod and they need to act on this if they want to avoid liability and negative product respond. There are so many posts about same issue that goes ignored.     original link was with different email address, I had to change it when I registered Mavic to new email account.

I have also been getting this error . . . and got no respond from DJI regarding this issue.

Didn't you contact them a couple of weeks ago? It's disappointing to hear that they haven't gotten back to you yet. I'd try sending a Personal Message to one or more of the DJI employees who participate on this forum. I've seen posts from Susan and Thor a bit lately, so you might try contacting one or both of them. And if that doesn't work, either add an entry to your original thread or create a new one and mention that you haven't gotten a response from them yet.

DroneFlying Posted at 2017-7-26 12:57
I have also been getting this error . . . and got no respond from DJI regarding this issue.

Didn't you contact them a couple of weeks ago? It's disappointing to hear that they haven't gotten back to you yet. I'd try sending a Personal Message to one or more of the DJI employees who participate on this forum. I've seen posts from Susan and Thor a bit lately, so you might try contacting one or both of them. And if that doesn't work, either add an entry to your original thread or create a new one and mention that you haven't gotten a response from them yet.

they did contact me but could not give any answer, they wanted me to send it to Netherlands but I am reluctant to do that as they can’t provide any idea or explanation about this error with the data I gave them, I am going to give it another try in safe are for testing.

do you guys suggest doing compass calibration everytime?