My first almost crash!

[anykill]

Today I was surprised , I was flying with my phantom 3 for a moment to another he began to make sudden movements and shot without a trace , then I pressed RTH and turned off the radio , I was white and thought I had lost my bird , I prayed to not have hit someone , a few minutes later he came back and we had a happy ending ! What could have happened? Attached LOG flight .
I lost signal, and a red cross appears on my software like some error... but i dont remember what it was, I was using DJI GO!
Now I'm afraind to happan again!


FLY363.DAT

Thanks for any help!

 

[StratocasterDave]

"It made sudden movements and shot without a trace"?

Can you elaborate a bit?

 

[anykill]

Run away and none of my stick commamds seems to work

 

[Ghost Rider]

It probably went into Atti mode. I've had that happen before but it came back online. So far, each time my signal started sketching out and I felt like I needed to abort, my aircraft came back safe.

 

[StratocasterDave]

Anykill, why did you shut the RC off after you pressed RTH?

 

[anykill]

I did not put in Atti, it acted like if it was folowing a prey... Or had been taked by an eagle.... But de font have Eagles here.
I shut the radio because i lost signal and the error appears on thr screen then i tough if its my radio that is the problem, then shut it will Rth and it will be safe.
You can see the event on the .DAT

 

[BudWalker]

I looked at this some. The event was at about 420.0 secs. The P3 had descended from 63 meters to 28 meters. The gpsHealth went from 5 (normal) to 3 and there was a drop in the number of satellites from 16 to 13.

The real story is what happened to the compass. There was a spike in the magMod value coincident with nonGPSCause = SPEED_ERROR_LARGE and a switch to ATTI.

The effect on the FC can be seen in this plot

magYaw is computed by DatCon and is based on the magnetometer values and the FC's values for roll and pitch. When the compass error happens magYaw indicates that the P3 is rotating CW but the gyroX values indicate that the P3 is not turning. The FC uses the magnetometer values for a short time but then has to declare a compass error and stops using magnetometer data for it's computation of Yaw. It would seem that the magnetometer readings are back to normal at about 423 but by this time there magYaw (the correct value for the heading and Yaw (the value used by the FC) differ by 107 degrees. The FC may have been rotating the P3 CCW to compensate for the perceived CW rotation. The Yaw error causes the "fly away".

As for the reason that there was a magMod spike I can only speculate that the P3 came close enough to the building at -51.1924748, -29.16801073 which are the coordinates where the incident took place. As mentioned above the P3 had descended to 28 meters; possibly low enough to be influenced by any geomagnetic distortions associated with that building.

@anykill do you have a video of that part of the flight that can be used to determine the heading during the incident? If so, then that could help decide if magYaw was the correct heading.

 

[anykill]

I will post the video, but not the right time of the incident because I was on photo mode.

 

[anykill]

Your explanation was very proffessional and helpfull, I was filming the building and I think I came to close to the iron bars.
Thank you very much