Mavic Mini Short Flyaway

[vhdblood]

Here is my log: DJI Flight Log Viewer - PhantomHelp.com

FW version is v01.00.0300 - App is 1.0.1

(Flyaway is at 0m 40s and appears to start when the 10th satellite connects.)

I took my Mini out to grab a shot of the sunrise this morning, gusts were only 15mph at max. I wanted to get a quick shot and I was kinda late getting out, so I just set the drone down and immediately hit both sticks down and in to start the drone and flew right out into the sky, as opposed to using the button in the app.

Once I got out and tried to go up, I was still waiting on GPS satellites and I was height limited, so I just let it sit and find the GPS in the air. It stayed stable and as soon as GPS showed up, I went straight up to get a good shot.

(Note: I was using the propeller guards.)

As soon as I got about 100ft up, the drone bee-lined sideways towards a park I normally fly in, without me touching anything. I tried to correct course and come back, but it just kept going until it was close to the park. I turned the camera towards me and it was then flying backwards, so it seemed like it was going in a specific direction on purpose. The drone did not appear to fight with any wind and was tilted as though I had the stick point the way it was flying. I was able to try and get it close to the ground in an open area in the park once it chilled out and stopped running away, but the drone left wifi range as I got to the ground (and visual range for that matter).

So since it lost the controller, it ended up returning to home, and it came back towards the correct location, the place I launched it from.

Does anyone know what happened here? I thought that maybe it used an old RTH location in the park I had set as the boundary for the propeller guards or something, but then it came home correctly when it lost the controller and it shouldn't have been out of that range anyway. I didn't do a screen recording (I will from now on) so I'm not sure if there was some useful error or anything.

I can say I will never launch again and go up in the air without a home point set and all GPS connected.

 

[Fly Dawg]

As soon as I got about 100ft up, the drone bee-lined sideways towards a park I normally fly in, without me touching anything. I tried to correct course and come back, but it just kept going until it was close to the park.

Initial look says yet another IMU positioning issue. Will dig deeper when I have the chance. The MM's seem to have this recurring issue thus far. You may want to post your FW versions for those who fly the mini and could give you more info on FW isuues.

 

[Meta4]

Does anyone know what happened here?

Your description has some elements of a Yaw Error situation but the data doesn't have enough information for me to be certain of the cause.
Can you describe the spotyou launched from?
This looks like a case for @sar104 or @BudWalker to look into.

I thought that maybe it used an old RTH location.

That is not possible.
The saved home point is lost when the drone is powered down.

it shouldn't have been out of that range anyway.

The data shows no Max Distance Limit set

I didn't do a screen recording (I will from now on) so I'm not sure if there was some useful error or anything.

The recorded flight data shows any error messages

I can say I will never launch again and go up in the air without a home point set and all GPS connected.

That had nothing to do with the incident and it is a perfectly safe practice.

 

[vhdblood]

This is great info, thank you for the straight facts.

Edited my first post with Firmware info.

Your description has some elements of a Yaw Error situation but the data doesn't have enough information for me to be certain of the cause.
Can you describe the spot you launched from?

I launched from my front porch. I had powered the drone and controller on inside on a level table, and then after a couple minutes walked out my front door, set the drone on the concrete porch, and did the controller launch with the stick command.

The data shows no Max Distance Limit set

I had the propeller guards on, so I should have had a distance limit from Payload mode.

 

[BudWalker]

........
I launched from my front porch. I had powered the drone and controller on inside on a level table, and then after a couple minutes walked out my front door, set the drone on the concrete porch, and did the controller launch with the stick command.

.......

Most likely Yaw would've been compromised due to a distorted geomagnetic field inside the house. That compromised Yaw was still in effect at launch. The Yaw data shows the MM facing NW at launch. do you recall the actual heading?

if you want to know more we'll need the .DAT from the tablet

How to retrieve a V3.DAT from the tablet

 

[vhdblood]

Most likely Yaw would've been compromised due to a distorted geomagnetic field inside the house. That compromised Yaw was still in effect at launch. The Yaw data shows the MM facing NW at launch. do you recall the actual heading?

if you want to know more we'll need the .DAT from the tablet

How to retrieve a V3.DAT from the tablet

I launched the drone with it facing east, so it certainly seems to have that wrong. I will always power on the drone outside in the flight location in the future.

I'll pull the .dat later when I get home just for reference.

 

[hiflyer201]

That concrete could contain rebar?

 

[vhdblood]

Here is that .dat for the flight.

 

[Fly Dawg]

@vhdblood Question. Did you power up the aircraft before placing it on the launch surface? It seems from your description and the before motor start data that you were moving it around. Immediately after you sat the aircraft down for takeoff there was a compass error. You should not move the aircraft around after power up.

 

[vhdblood]

@vhdblood Question. Did you power up the aircraft before placing it on the launch surface? It seems from your description and the before motor start data that you were moving it around. Immediately after you sat the aircraft down for takeoff there was a compass error. You should not move the aircraft around after power up.

View attachment 88077

Yes I did move it around quite a bit. I didn't realize how much of an issue that could be but I see now for sure.

 

[BudWalker]

Moving the AC after power up isn't a problem. The problem is powering it up in a geomagnetically distorted location which causes an incorrect Yaw. Moving the AC to a location that isn't geomagnetically distorted won't fix the incorrect Yaw. That's what happened here.

 

[Fly Dawg]

Moving the AC after power up isn't a problem.

Right. After initialization certainly it should be no problem. My point here is the fact that as the data began the aircraft was being moved around, which to me suggests that either the OP powered up the aircraft in hand, or immediately picked it up before it was fully initialized. NavHealth was at zero until 5 seconds before takeoff.

 

[BudWalker]

Right. After initialization certainly it should be no problem. My point here is the fact that as the data began the aircraft was being moved around, which to me suggests that either the OP powered up the aircraft in hand, or immediately picked it up before it was fully initialized. NavHealth was at zero until 5 seconds before takeoff.

I routinely power up my drones in hand and move them around before placing them on the ground. What part of the initialization would be compromised?

The OP has stated that he powered up inside the house and then moved outside to the launch site.

 

[Fly Dawg]

What part of the initialization would be compromised would be compromised?

Can't answer the question quite honestly. Only a theory, and from habit I suppose. I always power up on the most level surface as possible with the aircraft on the ground. Grass, asphault or otherwise and verify the headings ( Amoung other pre-flight things ) before starting the motors. Just to minimize error possibilities. That might be over kill, but Im so used to triple reduntancy its habit.

 

[vhdblood]

Right. After initialization certainly it should be no problem. My point here is the fact that as the data began the aircraft was being moved around, which to me suggests that either the OP powered up the aircraft in hand, or immediately picked it up before it was fully initialized. NavHealth was at zero until 5 seconds before takeoff.

I did not do this, I powered the aircraft on a level surface and the controller, and I waited a minute or two while I was getting ready to go outside. It sounds like the aircraft should have had time to initialize but there was some other issue.

 

[Fly Dawg]

It sounds like the aircraft should have had time to initialize but there was a magnetic interference issue.

Agreed, however there is only ~18 seconds of pre-motor start data.

 

[vhdblood]

Agreed, however there is only ~18 seconds of pre-motor start data.

How long does it take to initialize once it starts logging?