Mavic 2 Flyaway Crash - Not Enough Force/ESC Error

[Tbredhed]

@sar104 Would hitting pause force it to stop in place or would it continue trying to reconcile the compass and the IMU error?
Cheers

 

[ff22]

Absolutely. Although to be precise it's not that it stops GPS input - there was nothing wrong with the GPS data - it simply stops the FC from trying to control the aircraft when it has wildly incorrect heading data.

Okay, I figured out "FC" means Flight Controller (there's an acronym file I downloaded ages ago) BUT is that inside the DRONE or the Controller/device?

 

[Drbobk]

Is it only heading data? I thought the interaction of wrong heading data with changing position data was the issue. If it was just heading data, wouldn’t the craft just rotate until it thought it was heading correctly? I thought it was trying to move back to some starting position (assuming no stick input) but kept pushing itself the wrong way which then made its position comparison worse, and so on.

 

[sar104]

Okay, I figured out "FC" means Flight Controller (there's an acronym file I downloaded ages ago) BUT is that inside the DRONE or the Controller/device?

That's the computer on the aircraft that manages flight control.

Is it only heading data? I thought the interaction of wrong heading data with changing position data was the issue. If it was just heading data, wouldn’t the craft just rotate until it thought it was heading correctly? I thought it was trying to move back to some starting position (assuming no stick input) but kept pushing itself the wrong way which then made its position comparison worse, and so on.

Yes - but it's only the heading data that are wrong. When the FC detects positional drift it attempts to apply thrust to oppose the drift. If it has an incorrect yaw value then it will push in the wrong direction. If the yaw value is around 180° off then it will push in the opposite direction to what is required, leading to a positive feedback look and uncontrolled acceleration.

 

[kilomikebravo]

Can you explain that course of locating North initially with landing pad to M2P?

Wyojet: There are no stupid questions and yours is a good one. But the simple answer is, I just use the sun for a reference to east or west (if it's not high noon). If it's overcast or high noon and I'm in an area where I don't have a good geographical orientation, I just use the Compass app in my phone to find North.

 

[PropSpin]

Absolutely. Although to be precise it's not that it stops GPS input - there was nothing wrong with the GPS data - it simply stops the FC from trying to control the aircraft when it has wildly incorrect heading data.

Wait, how do you switch to ATTI mode? Has there been a firmware update that allows this? I haven't updated since last November - I don't remember the firmware version though.

 

[PropSpin]

@sar104 Would hitting pause force it to stop in place or would it continue trying to reconcile the compass and the IMU error?
Cheers

My guess is no since it would still attempt to maintain position which would be impossible given that the compass and IMU were out of sync.

 

[sar104]

Wait, how do you switch to ATTI mode? Has there been a firmware update that allows this? I haven't updated since last November - I don't remember the firmware version though.

No - ATTI is still not an option in the default configuration. You have to change one of the fswitch parameters to map it to one of the mode switch positions.

 

[sar104]

My guess is no since it would still attempt to maintain position which would be impossible given that the compass and IMU were out of sync.

That's correct. It's already trying to pause (hold position) and failing because of the yaw error. Telling it again won't help.

 

[paulatkin73]

That's correct. It's already trying to pause (hold position) and failing because of the yaw error. Telling it again won't help.

I wonder - was something like that also common for 1st gen mavic pros or is it something new on mavic 2 models? How many ppl posted about such flyaways already?

 

[sar104]

I wonder - was something like that also common for 1st gen mavic pros or is it something new on mavic 2 models? How many ppl posted about such flyaways already?

I wouldn't describe it as common but yes - there were similar events with the MP. The firmware does seem to have been iteratively modified to make the FC try harder to resolve the errors before aborting to ATTI, which I'm not sure has really been an improvement.

 

[Shon]

Absolutely. Although to be precise it's not that it stops GPS input - there was nothing wrong with the GPS data - it simply stops the FC from trying to control the aircraft when it has wildly incorrect heading data.

I’m wondering if attempting to recalibrate the compass before flight would have helped. Seems like the manual suggests this is NOT a good idea in areas with high interference, but they also suggest you do it if you haven’t flown in an while or if you are a certain distance away from your last flight location. How will you know if an unfamiliar area is experiencing interference?

My experience with other drones has indicated that if there is significant interference, one of the red flags, will be significant problems successfully calibrating your compass.

 

[sar104]

I’m wondering if attempting to recalibrate the compass before flight would have helped. Seems like the manual suggests this is NOT a good idea in areas with high interference, but they also suggest you do it if you haven’t flown in an while or if you are a certain distance away from your last flight location. How will you know if an unfamiliar area is experiencing interference?

My experience with other drones has indicated that if there is significant interference, one of the red flags, will be significant problems successfully calibrating your compass.

There is no way for calibration to help if the takeoff location is magnetically distorted. Whatever you do the initialized yaw value will be wrong, and when it becomes correct after takeoff the same problems will occur.

 

[Shon]

There is no way for calibration to help if the takeoff location is magnetically distorted. Whatever you do the initialized yaw value will be wrong, and when it becomes correct after takeoff the same problems will occur.

If there is interference though, the drone would fail calibration. So.... you wouldn’t take off... would you?

My old drone didn’t even give me the option, calibration fail, no flight. I suppose that’s not a thing with this one?

 

[paulatkin73]

There is no way for calibration to help if the takeoff location is magnetically distorted. Whatever you do the initialized yaw value will be wrong, and when it becomes correct after takeoff the same problems will occur.

as of 'takeoff location is magnetically distorted' - i wonder if folks do take off from the car`s hood or some metal plates? that will for sure affect compass.

it is just i have quite a bit of models, 11 now, all of them have gps and i am yet to see in any of them a severe gps to compass deviation after takeoff. it simply never happens. so it is so odd to see it in mavic, but, data seems indeed to indicate exactly that. so odd...

 

[sar104]

If there is interference though, the drone would fail calibration. So.... you wouldn’t take off... would you?

My old drone didn’t even give me the option, calibration fail, no flight. I suppose that’s not a thing with this one?

It may or may not fail calibration. Often the magnetic distortion, for example due to rebar in concrete, seems to be limited to within a few inches. A calibration done a few feet above the ground would not be affected.

 

[sar104]

as of 'takeoff location is magnetically distorted' - i wonder if folks do take off from the car`s hood or some metal plates? that will for sure affect compass.

it is just i have quite a bit of models, 11 now, all of them have gps and i am yet to see in any of them a severe gps to compass deviation after takeoff. it simply never happens. so it is so odd to see it in mavic, but, data seems indeed to indicate exactly that. so odd...

I've seen it with all models of Phantom and Mavic. And yes - taking off from the roof of a car or the bed of a truck is a common cause.

 

[Meta4]

I’m wondering if attempting to recalibrate the compass before flight would have helped.

A lot of people have an idea that calibrating the compass "fixes" problems but that's not what it does at all.
Calibrating the compass simply measures the magnetic fields that are part of the drone itself so that they can be ignored in flight.

Seems like the manual suggests this is NOT a good idea in areas with high interference, but they also suggest you do it if you haven’t flown in an while or if you are a certain distance away from your last flight location. How will you know if an unfamiliar area is experiencing interference?

Parts of the manual are confusing and poorly written.
There is no physical reason that calibration would be necessary at new sites or due to time since last flight.
The manual should advise not to launch from steel or reinforced concrete surfaces though.

If there is interference though, the drone would fail calibration. So.... you wouldn’t take off... would you?

as of 'takeoff location is magnetically distorted' - i wonder if folks do take off from the car`s hood or some metal plates? that will for sure affect compass.

When the outside magnetic field is too strong and easily identified as en external source, you get a warning of magnetic field interference and can't take off.
Like this example on a steel car roof

Unfortunately DJI's poor wording of the message confuses people by suggesting move the aircraft or calibrate.
Moving (after powering down) is the only correct action, calibrating will never fix that problem.

The problem comes when the magnetic field doesn't ring alarm bells but does cause a yaw error.
This view of what's inside a concrete sidewalk gives some clues:

You could put the Mavic down where its not close enough to enough steel to cause a problem.
.... Or you might put it down on top of an overlap area or a crossing point etc where the magnetic field is larger and right below the compass.
Reinforced concrete is an (un)lucky dip and should be avoided.
It's the most common cause of these issues.

it is just i have quite a bit of models, 11 now, all of them have gps and i am yet to see in any of them a severe gps to compass deviation after takeoff. it simply never happens. so it is so odd to see it in mavic, but, data seems indeed to indicate exactly that. so odd...

It's quite common but you won't see it unless you are launching from (magnetically) dangerous sites.
Incidents of this kind are reported here every week.

 

[Wilbur]

Because I've listen to Meta4 I have calibrated my compass once, and before I launch make sure the compass is lined up with the orientation of the AC. No fly aways, or erratic flights thus far. Definitely should be part of pre-flight check list.

 

[paulatkin73]

A lot of people have an idea that calibrating the compass "fixes" problems but that's not what it does at all.
Calibrating the compass simply measures the magnetic fields that are part of the drone itself so that they can be ignored in flight.

Parts of the manual are confusing and poorly written.
There is no physical reason that calibration would be necessary at new sites or due to time since last flight.
The manual should advise not to launch from steel or reinforced concrete surfaces though.


When the outside magnetic field is too strong and easily identified as en external source, you get a warning of magnetic field interference and can't take off.
Like this example on a steel car roof

Unfortunately DJI's poor wording of the message confuses people by suggesting move the aircraft or calibrate.
Moving (after powering down) is the only correct action, calibrating will never fix that problem.

The problem comes when the magnetic field doesn't ring alarm bells but does cause a yaw error.
This view of what's inside a concrete sidewalk gives some clues:
....

It's quite common but you won't see it unless you are launching from (magnetically) dangerous sites.
Incidents of this kind are reported here every week.

So, it is possible then to conclude that a safest way to avoid issues with compass is to hand launch?