Platinum preforming barrel rolls

[gvsukids]

Took off from front of boat and quad immediately flew backwards at high rate of speed. I was trying to give altitude as it was flying backwards and was able to get it up above other boats, and then it performed multiple barrel rolls before I was able to regain control where I brought it back to the boat.
Restarted the mavic and it flew fine the rest of the day.

 

[Meta4]

Took off from front of boat and quad immediately flew backwards at high rate of speed. I was trying to give altitude as it was flying backwards and was able to get it up above other boats, and then it performed multiple barrel rolls before I was able to regain control where I brought it back to the boat.

I was surprised to see the data confirms your drone really did roll upside down (7 times).
To be accurate it didn't actually perform any barrel rolls, but it probably looked like it did.
IT was rolling very quickly until upside down and then snapping back without going right around.
I've never seen flight data showing that (or barrel rolling).

Your problem was that you had a yaw error.
Can you describe where you powered up the drone and launched from?
What steel items or live electric cables were close by?

 

[sar104]

We really do need to see the mobile device DAT file (or aircraft DAT file) for that flight. Filename ending FLY054.DAT

 

[gvsukids]

We really do need to see the mobile device DAT file (or aircraft DAT file) for that flight. Filename ending FLY054.DAT

View attachment 105576

Here it is attached.

 

[sar104]

Here it is attached.

Very interesting - I've never seen anything like that. It appears to have been temporary IMU failure of some kind. Also, the attitude data from the DAT are different from the txt log shown above which is unusual. According to the DAT file, which is pretty definitive on these things, it wasn't rolling but actually did seven full back flips - while climbing:



The direct cause is clear - the motors were responsible of course, but why the FC was doing that is not clear:



The event stream indicates multiple control failures, none of which shed any further light on the problem. It may have been a computational failure rather than hardware.

1.221 : 1257 [L-FLYMODE][Ctrl<5>] REQ_APP AUTO_TAKEOFF ctrl_auto_takeoff​

2.919 : 1342 [L-FMU/FSM]not near ground​

3.417 : 1366 [L-NS][AHRS]height max obv num [3/16]​

3.577 : 1374 [L-NS][AHRS] yaw aligned 1​

4.064 : 1399 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

4.064 : 1399 [L-FDI][CTRL]: fault on , takeoff_fail​

4.084 : 1400 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

...​

5.023 : 1447 [L-FDI][CTRL]: fault off, takeoff_fail​

5.680 : 1480 [L-FDI]NS(0) GPS(0): fault off, abrupt​

6.242 : 1508 [L-FLYMODE]CTRL reset all by auto take off lanch fail​

6.263 : 1509 [L-FLYMODE][Ctrl<1>] REQ_RC_NORMAL ATTI ctrl_atti​

7.683 : 1580 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

...​

8.382 : 1615 [L-NS] [AHRS] gyro mag inconsist on​

8.680 : 1630 [L-FDI]NS(0) GPS(0): fault on , abrupt​

8.879 : 1640 [L-FDI]NS(0) GPS(0): fault off, abrupt​

10.122 : 1702 [L-NS][AHRS] gyro mag inconsist off​

10.304 : 1711 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

10.304 : 1711 [L-FDI][CTRL]: fault on , takeoff_fail​

10.323 : 1712 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

... etc.​

 

[gvsukids]

Very interesting - I've never seen anything like that. It appears to have been temporary IMU failure of some kind. Also, the attitude data from the DAT are different from the txt log shown above which is unusual. According to the DAT file, which is pretty definitive on these things, it wasn't rolling but actually did seven full back flips - while climbing:

View attachment 105609

The direct cause is clear - the motors were responsible of course, but why the FC was doing that is not clear:

View attachment 105610

The event stream indicates multiple control failures, none of which shed any further light on the problem. It may have been a computational failure rather than hardware.

1.221 : 1257 [L-FLYMODE][Ctrl<5>] REQ_APP AUTO_TAKEOFF ctrl_auto_takeoff​

2.919 : 1342 [L-FMU/FSM]not near ground​

3.417 : 1366 [L-NS][AHRS]height max obv num [3/16]​

3.577 : 1374 [L-NS][AHRS] yaw aligned 1​

4.064 : 1399 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

4.064 : 1399 [L-FDI][CTRL]: fault on , takeoff_fail​

4.084 : 1400 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

...​

5.023 : 1447 [L-FDI][CTRL]: fault off, takeoff_fail​

5.680 : 1480 [L-FDI]NS(0) GPS(0): fault off, abrupt​

6.242 : 1508 [L-FLYMODE]CTRL reset all by auto take off lanch fail​

6.263 : 1509 [L-FLYMODE][Ctrl<1>] REQ_RC_NORMAL ATTI ctrl_atti​

7.683 : 1580 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

...​

8.382 : 1615 [L-NS] [AHRS] gyro mag inconsist on​

8.680 : 1630 [L-FDI]NS(0) GPS(0): fault on , abrupt​

8.879 : 1640 [L-FDI]NS(0) GPS(0): fault off, abrupt​

10.122 : 1702 [L-NS][AHRS] gyro mag inconsist off​

10.304 : 1711 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

10.304 : 1711 [L-FDI][CTRL]: fault on , takeoff_fail​

10.323 : 1712 [L-FDI][CTRL] tilt ctrl failed when takeoff,75.0 deg​

... etc.​

This flight was performed with an iPad using the cable provided by Apple. Fist time using this combination, usually I have a different cable for the iPad.

So am I to continue flying as usual or are there issues that need to be addressed? It was my first time seeing this behavior as well. Like I said, the quad performed well the rest of the day while hooked up to my Android phone.

 

[Meta4]

This flight was performed with an iPad using the cable provided by Apple. Fist time using this combination, usually I have a different cable for the iPad.
So am I to continue flying as usual or are there issues that need to be addressed?

The issue was 100% unrelated to the tablet and cable.
What can you say about where you powered on and launched from?

 

[gvsukids]

The issue was 100% unrelated to the tablet and cable.
What can you say about where you powered on and launched from?

I powered on from the front of a motorboat that was powered off. And launched from the front of the same boat, as I had done many times and also later that day.

 

[Meta4]

I powered on from the front of a motorboat that was powered off. And launched from the front of the same boat, as I had done many times and also later that day.

The front of a motorboat doesn't tell me anything except that there may have been water nearby.
Was the drone sitting on a surface?
What was that made of?
What was underneath it or close to it, particularly steel items or live electric cables?

 

[sar104]

The front of a motorboat doesn't tell me anything except that there may have been water nearby.
Was the drone sitting on a surface?
What was that made of?
What was underneath it or close to it, particularly steel items or live electric cables?

Just to clarify - this wasn't caused by magnetic interference - the IMU and compass agreed well through the duration of the flips:

 

[sar104]

This flight was performed with an iPad using the cable provided by Apple. Fist time using this combination, usually I have a different cable for the iPad.

So am I to continue flying as usual or are there issues that need to be addressed? It was my first time seeing this behavior as well. Like I said, the quad performed well the rest of the day while hooked up to my Android phone.

As @Meta4 noted - this had nothing to do with the mobile device being used. I would check the IMU values and do an IMU calibration.

 

[gvsukids]

The front of a motorboat doesn't tell me anything except that there may have been water nearby.
Was the drone sitting on a surface?
What was that made of?
What was underneath it or close to it, particularly steel items or live electric cables?

The drone took off from my wife's hand.

 

[Stan1915]

Time to buy a lottery ticket!

 

[Camino Ken]

The drone took off from my wife's hand.

Does she have a “magnetic“ personality? ?

 

[Motorcycle Rider]

Does she have a “magnetic“ personality? ?

Well he married her...

 

[Chip]

... According to the DAT file, which is pretty definitive on these things, it wasn't rolling but actually did seven full back flips...

And then returned home without breaking a sweat. Incredible!

 

[Flygmaskin]

This behavior, whatever caused it, must be reverse engineered and exploited! I have been wanting the ability to perform flips with a Mavic for a very long time now!

 

[InterstateP]

Does she have a “magnetic“ personality? ?

*rimshot*