Mavic 2 Flyaway Crash - Not Enough Force/ESC Error

[NerdGoingOutside]

So, last September I was in northern Idaho with my brand new Mavic 2 Pro. At the time it has <5 houts of flight time.

I was down in the bottom of a canyon with no wind to speak of, and when I took off I had trouble getting a GPS signal. I hovered for a moment, and after a minute or so I got a GPS signal. I started ascending, and the home point was set successfully. I hadn't moved the drone yet other than to ascend. At about 100ft of elevation, all on its own, the drone suddenly picked up speed and began to fly forward without my input. I tried to regain control with no success, and a few seconds later the drone slammed into the side of the canyon at about 40mph in a place where I couldn't recover it. Knowing DJI Care Refresh wouldn't help without the drone and I was probably screwed out of $1500, I got frustrated, ended my vacation early, and tried to forget about it.

...then, yesterday, I realized my iPad had preserved the flight log. A .csv of the flight log is attached, zipped. Reviewing the flight log, you can see that before I touched the right stick the drone took off on its own, and shortly after it began moving started throwing 'Not Enough Force/ESC Error' errors.

Have you guys ever seen this before with the Mavic 2s? Flight log attached.

 

[BD0G]

@sar104


You should totally use the information gleaned from the .dat file and organize some searchers that are in the area. Referring to the Regional Section of this forum for people in the area. Offer a reward and some "treasure" hunters that are good souls may offer to search using the GPS Coordinates. Heck, people Geo Cache in crazy areas just for fun and this would be a much more meaningful hunt for your Mavic 2

I put out the Bat Signal for SAR104 above as he is a pro at deciphering Flight Logs.

Keep us updated.

 

[Thomas B]

After SAR104’s analysis you may have the info you need to pursue the issue with DJI. Even more so if you can recover the remains of the drone. Clearly you had contact with the drone, and for whatever reason it would not respond to some of the controller signals.

 

[ff22]

I don't know what's in the file you posted but generally using the following link and than posting the results link can help with analysis

DJI Flight Log Viewer | Phantom Help

 

[sar104]

This was a case of magnetic interference at takeoff.



Note that the recorded gimbal yaw at 46 seconds, when video was started, is -50° (310°), i.e. northwest. But the camera view at that time was this:



Comparison with GE shows that the camera view was actually to the southeast (140°) - the green line rather than the red line (-50°):



So the yaw, at this point, was almost exactly 180° out. As a result, when the aircraft started to use GPS and began to correct for slight drift, the FC applied pitch and roll in exactly the opposite direction to what was required. As the aircraft began to move in the wrong direction the FC increased the incorrect attitude and just made it worse. This is the classic linear uncontrolled flight that follows a yaw error of around 180°.

 

[ff22]

Even if the OP understands your analysis - could you translate for the unskilled? Was this a pilot error or might this be a warranty claim?

 

[sar104]

Even if the OP understands your analysis - could you translate for the unskilled? Was this a pilot error or might this be a warranty claim?

I guess that it depends on how you categorize taking off from a magnetically distorted location. It's easy enough to check for before takeoff (verify that the aircraft direction indicator on the map is pointing in the right direction), but nowhere does DJI advise doing that. In this case they also may point to taking off with no GPS position fix, but that didn't significantly contribute to the problem.

 

[sar104]

It would be useful to see the mobile device DAT file as well, if possible:

Mobile device DAT file: How to retrieve a V3.DAT from the tablet

 

[ff22]

I guess that it depends on how you categorize taking off from a magnetically distorted location. It's easy enough to check for before takeoff (verify that the aircraft direction indicator on the map is pointing in the right direction), but nowhere does DJI advise doing that. In this case they also may point to taking off with no GPS position fix, but that didn't significantly contribute to the problem.

You, sir, are a gentleman and scholar and incredibly helpful.

 

[FormerMember07122019_2]

I guess that it depends on how you categorize taking off from a magnetically distorted location. It's easy enough to check for before takeoff (verify that the aircraft direction indicator on the map is pointing in the right direction), but nowhere does DJI advise doing that. In this case they also may point to taking off with no GPS position fix, but that didn't significantly contribute to the problem.

I thought that you should get a message “MAG INTERFERENCE” on the controller/app? If not, I guess I should be a lot more diligent about checking the map versus aircraft orientation.

 

[sar104]

I thought that you should get a message “MAG INTERFERENCE” on the controller/app? If not, I guess I should be a lot more diligent about checking the map versus aircraft orientation.

No - generally you will not, unless the interference is so strong that the magnetic field components are outside what we assume are pre-defined bounds. At takeoff the compass and IMU yaw values agree - as they must since the IMU yaw is initially just given the compass yaw value (+ local declination). Even though the yaw value is wrong the FC cannot know that, and so there is no error.

Errors and unstable flight occurs once the compass starts reading correctly but the IMU yaw is still wrong - IMU yaw is primarily updated by the rate gyros with the compass just correcting for drift and bias, and it cannot reconcile a sudden change in compass reading without the rate gyros having recorded a similar rotation.

 

[NerdGoingOutside]

This is very interesting sar104, thank you. I, too, had assumed I'd be warned of any problematic magnetic interference by the app - I've had that happen before - so I didn't worry about that without a warning. Thinking back, the location is on pavement near an old mine site, so there could be some metal in the environment fairly easily. Rebar in the pavement, maybe.

I was confused when reviewing the flight log yesterday, because my recollection was that I had the nose pointed along the green line you've indicated and that the drone had suddenly flown backwards. When the indicator on the flight log showed the opposite I was perplexed but assumed I must have misremembered. This explains it.

I assume the 'Not Enough Force' error stems from trying to rev the engines high enough to compensate for the position drift?

I'll try and get the DAT file as well.

 

[Meta4]

This is very interesting sar104, thank you. I, too, had assumed I'd be warned of any problematic magnetic interference by the app - I've had that happen before - so I didn't worry about that without a warning. Thinking back, the location is on pavement near an old mine site, so there could be some metal in the environment fairly easily. Rebar in the pavement, maybe.

Similar incidents show up here every week.
The most common cause is launching from reinforced concrete surfaces.

 

[NerdGoingOutside]

Similar incidents show up here every week.
The most common cause is launching from reinforced concrete surfaces.

Yeah I've had that happen before, but always gotten an interference warning in the app. Had no idea there could be interference bad enough to cause this kind of uncontrolled flight but not bad enough to trigger an alert. If it happens that often I'm surprised there isn't a solution.

 

[Meta4]

Yeah I've had that happen before, but always gotten an interference warning in the app. Had no idea there could be interference bad enough to cause this kind of uncontrolled flight but not bad enough to trigger an alert. If it happens that often I'm surprised there isn't a solution.

See post #7 .. or simply avoiding launching from reinforced concrete surfaces will go a long way to preventing such mishaps.

 

[kilomikebravo]

As SAR said, so will double checking the aircraft heading on the display with the actual magnetic direction. DJI should absolutely add this procedure to the manual as it would certainly save a few drones. After my "almost" event, I do this every time I launch. The Hoodman landing pad helps a lot because it has cardinal markings on it so I just orient it to North and point the aircraft in that direction then, check the screen. Works like a champ and pretty much eliminates the possibility of a mag interference on takeoff flyaway.

 

[Joe_Silvia]

This is what happened to mine. I took off everything was normal and I noticed it drifting around. I turned around and with no forward input it took off at 64 MPH and 40 MPH down. I sent it to DJI and they replaced it no problem.
I asked them what happened and they never responded back.

 

[Wyojet]

As SAR said, so will double checking the aircraft heading on the display with the actual magnetic direction. DJI should absolutely add this procedure to the manual as it would certainly save a few drones. After my "almost" event, I do this every time I launch. The Hoodman landing pad helps a lot because it has cardinal markings on it so I just orient it to North and point the aircraft in that direction then, check the screen. Works like a champ and pretty much eliminates the possibility of a mag interference on takeoff flyaway.

KilBravo, I realize this may sound stupid as I am in the fence deciding whether to get the M2P vs phantom 4pro plus obsidian. Yet I sense great importance to your explanation as to seeking true North Orientation. As a flyaway is a major concern for myself. Although I'll never without the use of Marco Polo.

That said, for dummies like myself. Can you explain that course of locating North initially with landing pad to M2P? I think that issue is quite important from all the many specific crashes n flyaways I've seen on utube. Thanks Ty

 

[Drbobk]

In a situation like this one, would switching to ATTI mode help by stopping GPS input?

 

[sar104]

In a situation like this one, would switching to ATTI mode help by stopping GPS input?

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.