Fly aways due to compass errors

[Meta4]

Ok maybe DJI reads this, perhaps add a manual mode M in addition to Position, Sport and Cinema?

DJI don't read this
I'm not wanting to sound impolite but it's helpful if you actually understand the issue before trying to come up with solutions.

 

[Jeff A]

I won't fly with known compass errors. Who does this?

 

[Meta4]

I won't fly with known compass errors. Who does this?

Most victims of yaw errors (that's what this thread is about), have no idea until it's too late because there is no warning message to alert them..

 

[Jeff A]

I'm new tro this equipment but on my controller says Compass Error. Maybe not true with others. I have heard you should the align compass every 30 days or if you go more than 31 miles from your home location

 

[Meta4]

I'm new to this equipment but on my controller says Compass Error. Maybe not true with others.
I have heard you should the align compass every 30 days or if you go more than 31 miles from your home location

Your controller would not show any warning if you were setting yourself up for a Yaw Error.
A compass error warning is different and is usually just the compass warning that you have placed the drone close to something steel.
There is no need to calibrate your compass for 30 days or 31 miles, but DJi has mysteriously made the Mavic 2 ask for it, which just adds to the confusion already around compass issues.

 

[MAvic_South_Oz]

I'm new tro this equipment but on my controller says Compass Error.

As said by meta4, different issue with yaw error, and no warning is given.
Per my post 10 and reply at 11, when you are ready to fly (before start up waiting for home point to be updated), just look at the little map red arrow icon, and make sure that is very (very) close to direction drone is facing on the ground.

 

[Polyvista]

Surely compass information is just where North is and which way the drone is pointing. No information on actual position, so why would the coding tell the Drone to move at all?

 

[MAvic_South_Oz]

Surely compass information is just where North is and which way the drone is pointing. No information on actual position, so why would the coding tell the Drone to move at all?

From reading flight analysis posted by others on such incidents, it seems that the conflicting info between the 2 compass readings go the way of the wrong one.
The drone senses movement from wind etc, tries to correct, but it is the wrong way, usually common this is 180 degrees, so the drone keeps moving further away, and faster.
Until it is just belting away at full speed, and either flies away or crashes into something.

Sometimes the error can correct itself, but usually it goes out of signal range or crashes before this occurs.

There was such a flight I recall where the error was more like 90 degrees, was over water (the sea) and from a boat I am sure . . . the pilot was able to bring it back by some strange and complex stick inputs, very close to losing it, but in the end managed.

 

[Meta4]

From reading flight analysis posted by others on such incidents, it seems that the conflicting info between the 2 compass readings go the way of the wrong one.

It's a problem with conflicting data from the compass and the gyros in the IMU rather than between two compasses.
The gyros initialise at startup and are set to what the compass data indicates is north.
If the drone is in an area of magnetic interference, the compasses might not be giving an accurate reading.
When the drone leaves the magnetic interference area, the compass reverts to reading properly but is now conflicting with the IMU gyros that were initialised incorrectly.

 

[JAW]

It's a problem with conflicting data from the compass and the gyros in the IMU rather than between two compasses.
The gyros initialise at startup and are set to what the compass data indicates is north.
If the drone is in an area of magnetic interference, the compasses might not be giving an accurate reading.
When the drone leaves the magnetic interference area, the compass reverts to reading properly but is now conflicting with the IMU gyros that were initialised incorrectly.

Very clear explanation. Gracias.

 

[LivinLarge]

It is entirely possible to fly and return to home without any compass at all. Skydio does it. Perhaps DJI can figure it out too.

 

[sar104]

It is entirely possible to fly and return to home without any compass at all. Skydio does it. Perhaps DJI can figure it out too.

That's ATTI mode. They figured it out years ago.

 

[Prismatic]

That's ATTI mode. They figured it out years ago.

By which I'm sure you mean "DJI figured it out years ago."
But it begs the question: do Skydio drones not use an onboard compass? If they don't, how do they determine their orientation in the Earth frame-of-reference? That orientation info seems utterly essential for RTH, among other critical functions.
That's possibly an off-topic question in MavicPilots; if so, I apologize.

 

[LivinLarge]

That's ATTI mode. They figured it out years ago.

But the DJI drones cannot return to home in ATTI mode.

 

[sar104]

But the DJI drones cannot return to home in ATTI mode.

No drone can return home without using a compass to determine direction.

 

[radman]

From page 11 of the Skydio 2 manual:
“Tapping the Return to Home button gives you the ability to have Skydio 2 automatically return to either launch point or your current location.”
The Skydio 2 does not use a magnetic compass. It uses powerful magnets to hold the battery and filters in place, which would interfere with magnetic compass operation. It rather relies on optical methods to navigate around obstacles. Is it possible that it uses differences in GPS position to determine an approximate direction home? It appears that the Skydio 2 will stop and hover in place if it loses GPS signals. I’m in batch 3 so it will be at least a couple of months before I can get one to try out.

 

[cybervet]

No drone can return home without using a compass to determine direction.

Some years ago I had a Syma toy drone that had home lock and a kind of RTH (while there was a signal), I am pretty sure that it had no compass and GPS.

 

[sar104]

From page 11 of the Skydio 2 manual:
“Tapping the Return to Home button gives you the ability to have Skydio 2 automatically return to either launch point or your current location.”
The Skydio 2 does not use a magnetic compass. It uses powerful magnets to hold the battery and filters in place, which would interfere with magnetic compass operation. It rather relies on optical methods to navigate around obstacles. Is it possible that it uses differences in GPS position to determine an approximate direction home? It appears that the Skydio 2 will stop and hover in place if it loses GPS signals. I’m in batch 3 so it will be at least a couple of months before I can get one to try out.

Yes - they claim that they only use optical orientation to determine heading - presumably similar to the "vision compass" in the later Mavics. It will be interesting to see how that works in the absence of suitable fixed features, or whether it turns out to be robust enough for general use.

 

[DanMan32]

I've said before, GPS delta can also provide heading. My Tab E has no compass, but if I'm moving, it can give me direction on map apps. I grant you it can't help standing still, but if uncommanded movement occurs, there's your true heading into, assuming GPS data is still good. That should reset errors and stop uncontrolled movement.

 

[Arthur.N]

It's possible to build a path to HP with good GPS signal only. Probably it will require more computations and will look like driving a car (drone) by drank driver, since drone has to probe a direction(s) before move close to HP, but it's doable. When compass work properly it will be straightforward flight with less computations (and may be much faster), but such fallback when changes of GPS coordinates contradict with compass direction to HP may save some drones from flyaway. Isn't it?