I saved my Air from RTHing away

[Epp1983]

Hey guys, yesterday was an experiment day.
I'm usually a believer of knowing the limitations of equiment. If you know what they are capable of you know what to expect from them.
I was a Mavic Pro owner for over an year I think I have tried every aspect of it.
Yesterday I decided to see the Mavic Air response to loss of connection.

The basics:
- No wind
- Constant LOS
- Home point set
- 60m RTH height
- No sensor errors (checked right before turning remote control off, and no issues while flying before the test)

I hovered the drone roughly 400 feet away from me at a height of 120 feet and switched off the remote.
After about 3 seconds the Air started climbing to the predefined RTH height. So far so good.
But then it turned to the oposite direction and started to move further away from the home point.
I immediatly turned on the remote and a little after I regained downlink.
Before canceling the RTH I checked the map to see if the home point had changed, and it had not.
The aircraft was indeed moving away from homepoint, and still no sensor errors.
RTH was then cancelled and I flew the drone back to me (the homepoint) for a successful retrieval.

After this incident, I popped a fresh battery and repeated the experiment under the same conditions and while hovering in the previous location (at the time I tought it could be some kind of magnetic interference at that location).
The drone executed the RTH correctly.

Now, this first RTH was pretty odd as everything was correctly set.
I've seen a couple of flyaway happening but usually they are due to pilor errors.

I'll leave the flightlog for you to check if you want to.
DJI Flight Log Viewer - PhantomHelp.com

Remote control off is at 12m 5s.
Downlink restored at 12m 54.3s.

Let me know your thoughts about this incident.
Thanks in advance.

Best regards,

Emanuel

 

[RayOZ]

Warning about taking off in Class C airspace? Not particularly sure what that means. I wonder if there’s conflicting commands; go home or get outside restricted airspace?

 

[Epp1983]

Warning about taking off in Class C airspace? Not particularly sure what that means. I wonder if there’s conflicting commands; go home or get outside restricted airspace?

Thanks for your reply.
I don't think it's related to the Class C airspace warning..
After this incident, I popped a fresh battery and repeated the experiment under the same conditions and while hovering in the previous location (at the time I tought it could be some kind of magnetic interference at that location).
The drone executed the RTH correctly.
I didn't included this extra information in the main post as I did not find it relevant, but maybe it actually is. I'll update the post.

 

[Jameslavis]

I live near a class D and it does nothing to my drone.

 

[rfc]

Warning about taking off in Class C airspace? Not particularly sure what that means. I wonder if there’s conflicting commands; go home or get outside restricted airspace?

Don't know how RTH vs. Restricted Airspace works, but the reason for the Class C warning is a no brainer: 10 miles from Lisbon International.

 

[sar104]

That's a strange event, and nothing leaps out as the cause. The homepoint was correctly recorded and, not that it should have mattered, the app itself had the same location data for the entire flight. Uplink disconnect happened at 726.5 s with the aircraft at 131 m from the HP and facing -124°, which was approximately towards the HP. There were no data recorded for 48 s until the RC reconnected, at which point the aircraft had oriented itself on a heading of 93° (facing away from the HP) and traveled away from the HP at an average speed of around 5 m/s, which is slower than the regular RTH speed.

 

[Tinmania]

Compass heading matched true heading?



Mike

 

[sar104]

Compass heading matched true heading?



Mike

Yes. Heading and track align.

 

[swooshdave]

Compass heading matched true heading?

Mike

I check that before every flight. Wait for the GPS lock then confirm compass heading.

 

[swooshdave]

That's a strange event, and nothing leaps out as the cause. The homepoint was correctly recorded and, not that it should have mattered, the app itself had the same location data for the entire flight. Uplink disconnect happened at 726.5 s with the aircraft at 131 m from the HP and facing -124°, which was approximately towards the HP. There were no data recorded for 48 s until the RC reconnected, at which point the aircraft had oriented itself on a heading of 93° (facing away from the HP) and traveled away from the HP at an average speed of around 5 m/s, which is slower than the regular RTH speed.

I noted in another thread that after the last app update the warning when you are flying near an airport now says it can affect the RTH but of course it doesn’t say how. I may need to dig into the DJI Go release notes.

 

[Kevjones777]

In work on the principle if u dont get the message on your screen "Home Point Set" ... it is not set.

 

[sar104]

In work on the principle if u dont get the message on your screen "Home Point Set" ... it is not set.

According to the log file it was set before takeoff, and it was correct, but it is interesting that there was no listed notification of that event. I wonder if it was attempting to return to a previous home point.

 

[msinger]

I wonder if it was attempting to return to a previous home point.

Ever see something like that in a flight log? I haven't.

This flight would probably make more sense if we had the DAT flight log.

 

[Epp1983]

According to the log file it was set before takeoff, and it was correct, but it is interesting that there was no listed notification of that event. I wonder if it was attempting to return to a previous home point.

That would be odd. The app itself had the homepoint set. I took off using auto-takeoff with percision landing, and "H" was shoing on the map.
Also, on the log it correctly measures the distance to the homepoint.

 

[Epp1983]

This flight would probably make more sense if we had the DAT flight log.

As soon as I get home I'll try to place it here.

 

[Tinmania]

That would be odd. The app itself had the homepoint set. I took off using auto-takeoff with percision landing, and "H" was shoing on the map.
Also, on the log it correctly measures the distance to the homepoint.

Do the log HP coordinates match actual HP location? GPS health was good?

Has to be something here to explain it.


Mike

 

[sar104]

Ever see something like that in a flight log? I haven't.

This flight would probably make more sense if we had the DAT flight log.

No - never seen that. I was just trying to come up with some kind of hypothesis for why the aircraft would quite deliberately proceed in the wrong direction. What was it thinking, etc.

One of the DAT files would be very useful - the log entries might reveal something about the decision process, and possibly explain why there was never a "home point recorded" message even though it logged an appropriate home point and set the homepoint recorded flag.

 

[sar104]

That would be odd. The app itself had the homepoint set. I took off using auto-takeoff with percision landing, and "H" was shoing on the map.
Also, on the log it correctly measures the distance to the homepoint.

Right - and the log data clearly show the homepoint at the takeoff location. It reset slightly after a couple of seconds, but the change was less than 1 meter.

 

[Epp1983]

Here are the DAT files.
Flight Record & Sensing: flightrecord_DJI_ASSISTANT_EXPORT_FILE_2018-04-19_21-41-27.DAT
Navigation: nav_DJI_ASSISTANT_EXPORT_FILE_2018-04-19_21-42-31.DAT

I hope it sheds some light on the subject.

 

[sar104]

Here are the DAT files.
Flight Record & Sensing: flightrecord_DJI_ASSISTANT_EXPORT_FILE_2018-04-19_21-41-27.DAT
Navigation: nav_DJI_ASSISTANT_EXPORT_FILE_2018-04-19_21-42-31.DAT

I hope it sheds some light on the subject.

Neither of those files appears to be a flight log. Did you select a flight control entry? The DAT file from the mobile device might be easier to find.