Drone crash - what did I do?

[sar104]

Depends on your location.

I've only had this happen to me once, and my pre-programmed response to things being weird is to switch to sports mode and go Full UP. This got me well away from the car that was apparently causing interference, and gave my Mavic Pro time to gets its **** together, so I could then land in a controlled manner a few moments later.

Its much easier to control something in a large wide open space, and you will generally have more obstacles closer to the ground so up has always been a better option for me.

This will only work in unusual circumstances - namely when the interference wasn't at the takeoff point but occurred after the flight began. Climbing may correct the yaw error in that case. But if the problem was incorrect yaw initialization to start with, then the only way out is either ATTI mode (if you have that selectable) or get the aircraft on the ground as fast as possible.

 

[JAW]

I actually happened across a group of people looking for "orbs" in Fredericksburg, VA near Hamilton's crossing (lots of civil war battles in the area) almost 20 years ago now. These orbs were supposedly balls of energy and they could be photographed, but not seen w/ the naked eye. If you had the right equipment, you could record audio of the voices of the long since departed according to these people. They said that they were not ghost hunting, mind you, in spite of the fact that they did most of their searching in places where there were lots of death had occurred in the past. Go figure. They related it all to quantum mechanics and also said that these orbs could drain the power from batteries. While we were talking, one of them snapped a photo and showed me the result in the camera's display. There certainly was something next to me that looked like a golf ball sized bubble. Not sure what to make of it all; I tried looking info up on the subject, but never really found anything.

Got any more of that stuff?

 

[RicGPDX]

It looks like another case of magnetic interference at the takeoff location.

Note that the yaw value is initialized at -116°. After takeoff, the yaw changes to -90° and forward elevator is applied, causing it to pitch forwards as expected. It should therefore move west, but it actually moves east.

(Diagrams clipped out to save space

That's reflected clearly in the GPS flight path, but note that the IMU initially thinks that it is moving west, based on its accelerometers and incorrect yaw value. It's probable that it was actually facing east. The mobile device DAT file contains the necessary magnetometer data to confirm that.

sar104,
Could you point me at a resource that will explain these flight reports? I just had a lost-contact experience (no video, no control) with my Mavic Air that I'd like to investigate. (It did a surprisingly accurate return-to-home, however.)

Thanks,
Ric

 

[sar104]

sar104,
Could you point me at a resource that will explain these flight reports? I just had a lost-contact experience (no video, no control) with my Mavic Air that I'd like to investigate. (It did a surprisingly accurate return-to-home, however.)

Thanks,
Ric

I don't think that such a resource really exists yet, beyond the explanations scattered amongst the threads on these forums. I think that there are only a few of us who do these kinds of analysis, and with a variety of approaches, most of which require some familiarity with mathematical methods and somewhat specialist software. If you have specific questions I'd be happy to try to answer them though.

 

[Dingo]

Just wondering if you did a compass calibration at the site before take off ?? The other day i was about to take off from a rock wall held together with steel wire mesh. I got the green all clear for take off( no magnetic interferance or other warnings)but decided to check if the red triangle on my map( drone) was facing the island approx 300M away. It was facing the wrong way even though my drone was poniting straight at it.. I only picked this info up a few weeks ago on this forum. I powered off the drone and tx, app. relocated to a non metal take off point approx 2 metres away and then checked the red triangle was now pointing the correct way/same as drone was pointing. This my mavic pro 1. Sounds like I was lucky or could have experienced the same issue you had.

 

[Meta4]

Just wondering if you did a compass calibration at the site before take off ??

Calibrating the compass or not calibrating the compass makes no difference in these cases.
Calibrating the compass doesn't "fix" anything and is not the solution.

 

[buzzbomb]

Speed started to take off around 18 seconds and the crash was at 38 seconds with speeds up to 55 mph.

maybe speed caused the crash.

there is topic about sports mode causing some interference when lot of current goes from to engines

I experienced it when flying in sport mode... received a red warning about compass calibration so I decided not to use sport mode any more

 

[sar104]

maybe speed caused the crash.

there is topic about sports mode causing some interference when lot of current goes from to engines

I experienced it when flying in sport mode... received a red warning about compass calibration so I decided not to use sport mode any more

The difference in current is not that great, but it could possibly be enough to trigger a warning. This was magnetic interference due to the launch point though, so unrelated.

 

[CaptPete]

So when the compass error is present, does the drone simply go away from instead of towards a gps coordinate it is trying to maintain? Should it not be possible for software in the app to see this error taking place (conflict between heading and closing distance), and self-correct by changing "heading" incrementally until it is moving towards instead of away from a home-point? That way a RTH could overcome even this flyaway scenario, or have I oversimplified the error? It seems that ignoring compass with a software failsafe like this would allow navigation to home point with compass completely ignored.

I am not saying that this failsafe could correct compass (because of wind speed influence), just would enable navigate home.

I would be interested in hearing from the abundance of knowledgeable pilots and software engineers, and if possible have this feature put forward to be incorporated in a future GO4 update.

 

[sar104]

So when the compass error is present, does the drone simply go away from instead of towards a gps coordinate it is trying to maintain? Should it not be possible for software in the app to see this error taking place (conflict between heading and closing distance), and self-correct by changing "heading" incrementally until it is moving towards instead of away from a home-point? That way a RTH could overcome even this flyaway scenario, or have I oversimplified the error? It seems that ignoring compass with a software failsafe like this would allow navigation to home point with compass completely ignored.

I am not saying that this failsafe could correct compass (because of wind speed influence), just would enable navigate home.

I would be interested in hearing from the abundance of knowledgeable pilots and software engineers, and if possible have this feature put forward to be incorporated in a future GO4 update.

The FC tries to do just that - rotate the aircraft while holding the IMU yaw value constant. Unfortunately, for large errors, the flight rarely lasts long enough for that strategy to work before it comes to an untimely end.

 

[Jim Padgett]

A couple screws in the wood? Well, what about these 'Magnetic' detachable USB Cables? Like the ones recommended to help save the fragile Power Port on the controller. I don't even like throwing those into the same bag with anything they could kill.

 

[kadras]

A couple screws in the wood? Well, what about these 'Magnetic' detachable USB Cables? Like the ones recommended to help save the fragile Power Port on the controller. I don't even like throwing those into the same bag with anything they could kill.

I was thinking the same thing and I decided not to keep the cable in with all the other equipment.

 

[BudWalker]

A couple screws in the wood? Well, what about these 'Magnetic' detachable USB Cables? Like the ones recommended to help save the fragile Power Port on the controller. I don't even like throwing those into the same bag with anything they could kill.

The question is how close is it to the magnetometers. That's a pretty strong magnet (I have one of these) but I would think that 10" is enough separation. But, the absolute easiest way to know for sure is to use the AC's magnetometers to see if they are affected. Just make sure that the red triangle heading indicator on the map display is consistent with the AC's actual orientation.

 

[PropSpin]

Got any more of that stuff?

Nope!

 

[Steve L Reid]

My units have always warned me during calibration of interference.

 

[MAvic_South_Oz]

My units have always warned me during calibration of interference.

As per numerous threads of late, it doesn't always.
That is the issue, and why a simple addition to check orientation of aircraft and map direction when doing pre flight checks is pretty important.

 

[Steve L Reid]

Are you saying some will calibrate anyway in a strong magnetic field?

 

[AMann]

Asphalt is generally safe as it doesn't have steel reinforcing like concrete.
But it can sometimes cover steel plates or pipes etc that can cause this sort of issue.

He was at a cemetery, maybe a secret vault or? Should go back with a shovel and find out!

Sounds like checking orientation on the map prior to takeoff is a good idea.

Is it possible that the compass needed to be degaussed?