Lost of control M2P possible issue with "Fly Immediately" on SC

[Deleted member 114292]

I have a fair amount of stick time with this AC never an issue. I turned on "Fly Immediately" for the first time on my SC.
All of a sudden after a few minutes it was turning and the camera yawed hard left.
I hit RTH and it came back in control so I flew back shut off "Fly Immediately" and reset everything. The rest of the flight was uneventful.
Needless to say I will not try "Fly Immediately" again.

 

[slup]

Jeez ...yeah, that doesn't look normal. We need the DAT log for this, look here under point 3 (Mobile device DAT files (DJI GO 4 & DJI Fly) over how --> Flight log retrieval and analysis guide then attach the log here in a new post. (If now the DAT location is the same in SC?)

 

[JAW]

Not being familiar with the SC, what is “Fly Immediately” supposed to do?

 

[Deleted member 114292]

If you power up the ac first, the sc goes right into go4 ready to fly,

 

[Deleted member 114292]

Try this.
Found in
This PC\rm500\Internal shared storage\DJI\dji.go.v4\FlightRecord\MCDatFlightRecords

 

[slup]

@sar104 ... what is this, is it a compass error together with that that function M2P have to align Yaw with magYaw which is unique for that model?

 

[Deleted member 114292]

@slup,
I guess @sar104 is not interested in this one.
Am I correct that the saw tooth in the yaw means spinning? While the sticks are centered?

 

[sar104]

@sar104 ... what is this, is it a compass error together with that that function M2P have to align Yaw with magYaw which is unique for that model?

View attachment 93376

There's no compass error here - the compass and inertial data agree closely. The aircraft starts the CCW yaw at the request of the full left rudder command at 366 seconds and rotates at the appropriate rate, but when the rudder is centered at 376 seconds it continues the rotation. There are discernible blips in the rate when full CW rudder is applied briefly at 377 and 380 seconds, but it largely continues to rotate CCW until 407 seconds.



The log sees the actual rudder inputs, but the aircraft doesn't respond. I've not seen that before, and I don't see anything in the logs to account for it.

@BudWalker?

@slup,
I guess @sar104 is not interested in this one.

Or you may have incorrectly assumed that I'm on call 24/7 to respond to this forum.

 

[slup]

There's no compass error here - the compass and inertial data agree closely. The aircraft starts the CCW yaw at the request of the full left rudder command at 366 seconds and rotates at the appropriate rate, but when the rudder is centered at 376 seconds it continues the rotation. There are discernible blips in the rate when full CW rudder is applied briefly at 377 and 380 seconds, but it largely continues to rotate CCW until 407 seconds.

View attachment 93467

The log sees the actual rudder inputs, but the aircraft doesn't respond. I've not seen that before, and I don't see anything in the logs to account for it.

@BudWalker?


Or you may have incorrectly assumed that I'm on call 24/7 to respond to this forum.

No ... didn't think it was a compass error deep down either, but all looked really strange, hadn't seen this behavior earlier (which isn't so strange as I'm new into this). Then came to think of what you explained about M2's better/or different solution to deal with yaw/mayaw disagreements when one aligns with the other.

To the OP ... is it something special going on around that property you flew over when this happened?

 

[Deleted member 114292]

There's no compass error here - the compass and inertial data agree closely. The aircraft starts the CCW yaw at the request of the full left rudder command at 366 seconds and rotates at the appropriate rate, but when the rudder is centered at 376 seconds it continues the rotation. There are discernible blips in the rate when full CW rudder is applied briefly at 377 and 380 seconds, but it largely continues to rotate CCW until 407 seconds.

View attachment 93467

The log sees the actual rudder inputs, but the aircraft doesn't respond. I've not seen that before, and I don't see anything in the logs to account for it.

@BudWalker?


Or you may have incorrectly assumed that I'm on call 24/7 to respond to this forum.

Oh thank you for your time sar I can imagine there are many demands. Sorry if I disrespected you.

My only concern is if there could be a safety Issue with this particular instance.
I have flown three or four flights with it since and not seen Unexplained anomalies.
I flew a real estate inspection mission which required careful maneuvering and saw no issues.
Lacking your, and the forums, deeper knowledge of the history of events such as this I thought maybe it was an infrequent but known class of events.
Thank you for determining that it is unusual if not unique.

 

[Deleted member 114292]

No ... didn't think it was a compass error deep .
To the OP ... is it something special going on around that property you flew over when this happened?

I actually had the same thought. I flew around the area to see if it might happen again but not with enough diligence to ensure I had adequately duplicated the flight path.
I will look for a tool that will allow me to take the exported KLM file and re-create the mission.
The area the neighborhood is in could have a massive iron hull from a long forgotten scuttled iron schooner. Very unlikely but possible.

 

[new2mavic]

My observation is essentially the same as that of sar104 except that the craft did respond to rudder input as shown in the graph below. The anomaly here is that the craft started to yaw to the left on it’s own at 378 second. Then it stopped auto yawing after switching to GO HOME mode.

After 50 seconds ( ie, at 428 second ), the GO HOME mode was exited and the auto yawing behavior came back but the yawing was in the opposite direction and the rate was much lower. This behavior continued all the way to the end of the flight so the OP’s impression“…. The rest of the flight was uneventful….” is incorrect.

As said by sar104, there was no significant discrepancies between the magnetic yaw and the gyro yaw so the problem was unlikely to be caused by sensor errors or magnetic interference. It appears that the flight controller just went nut thinking that it was required to track something ? The behavior does look like some kind of tracking actions to me.

 

[BudWalker]

There's no compass error here - the compass and inertial data agree closely. The aircraft starts the CCW yaw at the request of the full left rudder command at 366 seconds and rotates at the appropriate rate, but when the rudder is centered at 376 seconds it continues the rotation. There are discernible blips in the rate when full CW rudder is applied briefly at 377 and 380 seconds, but it largely continues to rotate CCW until 407 seconds.

View attachment 93467

The log sees the actual rudder inputs, but the aircraft doesn't respond. I've not seen that before, and I don't see anything in the logs to account for it.

@BudWalker?


Or you may have incorrectly assumed that I'm on call 24/7 to respond to this forum.

Interesting. The only observation I can offer is that the uncommanded rotations stopped when flyCState switched from GPS+ATTI to GoHome.

EDIT: @new2mavic I should've read to the end of the thread....

 

[sar104]

My observation is essentially the same as that of sar104 except that the craft did respond to rudder input as shown in the graph below. The anomaly here is that the craft started to yaw to the left on it’s own at 378 second. Then it stopped auto yawing after switching to GO HOME mode.

After 50 seconds ( ie, at 428 second ), the GO HOME mode was exited and the auto yawing behavior came back but the yawing was in the opposite direction and the rate was much lower. This behavior continued all the way to the end of the flight so the OP’s impression“…. The rest of the flight was uneventful….” is incorrect.

As said by sar104, there was no significant discrepancies between the magnetic yaw and the gyro yaw so the problem was unlikely to be caused by sensor errors or magnetic interference. It appears that the flight controller just went nut thinking that it was required to track something ? The behavior does look like some kind of tracking actions to me.

View attachment 93494

Good spot - I was so distracted by the obvious CCW yaw issue that I didn't notice the later anomalies. There is no evidence that it was in any kind of tracking mode, and it is clear that all the sensors saw the yaw occurring, which implies to me that the FC commanded it. I have no idea why though - this looks like a processor problem.

 

[sar104]

Oh thank you for your time sar I can imagine there are many demands. Sorry if I disrespected you.

I wasn't concerned about respect or otherwise - just unrealistic expectations.

 

[Deleted member 114292]

Quite possible I missed that slight yaw to the right. I was flying home via radar as during the event I had lost LOS.
After that Hard left yaw It probably seemed normal at that point.

 

[Deleted member 114292]

[

I have no idea why though - this looks like a processor problem.

And if by that you mean an odd bug in the firmware that was my swag. Yesterday I was flying it and did a lot of hovering amongst obstacles to get the shots I needed.
No hint of auto yaw.
The only factor I have a correlation to is trying
out the fly instantly feature. But it’s a correlation with only one data point which I don’t plan to test again.
I can’t afford to give up my source of income.
Should I return it to DJI and purchase a back up in the meantime?

 

[slup]

[

And if by that you mean an odd bug in the firmware that was my swag. Yesterday I was flying it and did a lot of hovering amongst obstacles to get the shots I needed.
No hint of auto yaw.
The only factor I have a correlation to is trying
out the fly instantly feature. But it’s a correlation with only one data point which I don’t plan to test again.
I can’t afford to give up my source of income.
Should I return it to DJI and purchase a back up in the meantime?

If you could live with loosing your drone for a while ... you could contact DJI for further investigation as they have access & can read the AC DAT log. Otherwise this might be a ticking bomb just waiting to go off again.