Mavic Air took off to the right and scared me stupid.

[skyturnstoblack]

So I've only owned it for like a week and I've had 2 issues.

The first I think was from me not realizing it didn't have gps lock because the only drones I had played with before were just controller ones with no brains. It took off said it hit it's maximum height at like 10 feet. Started bobbing down right above the ocean I brought it back in and dropped it asap. That one after looking it up was probably because I needed to let the gps lock with 8 sats first.

I watched about 50 youtube videos trying to avoid stupid mistakes but it happened. the next problem was pretty similar but very different.

So I get a home lock I take off and head out the phone gives me the warning about being by my local airport which I called and informed. So as i'm trying to click through that I look up and the mavic air is shooting to the right. Over the water again and towards a large hotel. I pulled the logs and have almost 0 way to understand them. It went about 50 feet. I"m assuming it is gps again i'm taking off by a 12 story building. but am hoping someone can glance at the logs and tell me if one of the sensors needs calibrating. I could calibrate everything but am worried i'll do more damage than good if I just start stabbing in the dark.

It happens right around the 20 second mark. There was a good breeze but it was out to sea not across the beach.

DJI Flight Log Viewer - PhantomHelp.com

 

[Meta4]

It happens right around the 20 second mark.

Are you using Mode 2 for your joysticks?
If you are, the data is scrambled.

If you are using a different joystick mode, it makes it hard to analyse the flight data.

 

[msinger]

Did you try flying on a day when there wasn't strong winds to compare?

 

[Lastrexking]

Some heavy wind there. Also, took off before GPS locked -if you weren't aware of the need for GPS then I strongly recommend reading the manual before you fly again. No doubt you spent a lot of money on your drone - reading the manual will save you losing that money.

Winds on the coast are often a lot stronger even a short distance from the beach.

 

[JSKCKNIT]

I'll be in Kahului next week, I'll shoot you a DM.

I've flown in similar winds and my Air never drifted off that much. Even though you were in "GPS" mode, the more satellites you have locked, the better.

 

[Meta4]

Also, took off before GPS locked -if you weren't aware of the need for GPS then I strongly recommend reading the manual before you fly again. No doubt you spent a lot of money on your drone - reading the manual will save you losing that money.

The recorded flight data clearly shows that he had GPS before launching.
Reading the data saves you making incorrect assumptions.

 

[skyturnstoblack]

Are you using Mode 2 for your joysticks?
If you are, the data is scrambled.

If you are using a different joystick mode, it makes it hard to analyse the flight data.

I'm using controller 3. It's more like a fps. I kept trying to use the stock control but it was like writing with the wrong hand so I set it up closer to a video game controller.

 

[skyturnstoblack]

Did you try flying on a day when there wasn't strong winds to compare?

View attachment 59230


View attachment 59231

This data actually seems like it might be closer to explaining what happened. With the building at my back it felt like the wind was blowing out to sea. But I thought it could handle winds like these. May be I'm just used to the winds in maui but it really didn't seem any higher than the videos I've seen for sure. And it handled whe wind without drifting right after that. Is there like a time before it starts compensating? And it may have just been the shock but it didn't seem to react to controls for like a second until I spun it. Then it seemed to stop drifting and start reacting to imputs.

 

[msinger]

Is there like a time before it starts compensating?

No. With a good GPS connection, it should always be trying to hold its position.

 

[Meta4]

This data actually seems like it might be closer to explaining what happened. With the building at my back it felt like the wind was blowing out to sea. But I thought it could handle winds like these. May be I'm just used to the winds in maui but it really didn't seem any higher than the videos I've seen for sure. And it handled whe wind without drifting right after that. Is there like a time before it starts compensating? And it may have just been the shock but it didn't seem to react to controls for like a second until I spun it. Then it seemed to stop drifting and start reacting to imputs.

The wind doesn't explain the movement.
There would have been similar wind affecting the drone throughout the flight but it reacted normally.

At 0:27.8 you moved one of your sticks (aileron ?) to it's extreme one of your sticks and the speed jumped from zero.
At 0:28.7 you moved the other stick (rudder ?) to one of its extremes and your Mavic rotated about 130 degrees while still accelerating forward.
You went hands-off at 0:30.2 and the speed decreased until 0:38 when you pulled/pushed another stick and turned the drone about 90 degrees.
Sorry but I find it very hard to read the data when I can't tell which stick you were moving.
It's would be very easy to understand the data if you were using the standard joystick mode.

But it looks to me like the movement you are concerned about was entirely due to your joystick inputs.

 

[skyturnstoblack]

The movement I'm referring to is at about 20 seconds. And the stick stuff is just swapped it's not custom. It started to respond at about 28 seconds again and act normally. There were no stick inputs from the 20 second to like 28 ish. But the drone moved like 50 feet. Just imagine the right stick is left and the left stick is right and you should better understand what happened. From about 20 seconds to like 28 ish I was dismissing the warning message on the phone. With no controller activity. I looked up and it had shifted way too the right. I pushed left didn't seem to get a response so I spun it to see if it was responding at all. Doing that seemed to smack it outta whatever was happening. And it started holding position and acting normally.

 

[Meta4]

The movement I'm referring to is at about 20 seconds. And the stick stuff is just swapped it's not custom. It started to respond at about 28 seconds again and act normally. There were no stick inputs from the 20 second to like 28 ish. But the drone moved like 50 feet. Just imagine the right stick is left and the left stick is right and you should better understand what happened. From about 20 seconds to like 28 ish I was dismissing the warning message on the phone. With no controller activity. I looked up and it had shifted way too the right. I pushed left didn't seem to get a response so I spun it to see if it was responding at all. Doing that seemed to smack it outta whatever was happening. And it started holding position and acting normally.

Perhaps the times you are referring to don't match the times in the full data that I've been going over (for much too long - it really is hard to make sense of when the joysticks aren't normal).
You went hands off from 0:17 until 27.8. In that time the only movement was the drone continuing forward for 20 ft and slowing down.
But all the movement I see corresponded to joystick inputs you made.

 

[skyturnstoblack]

The distance from home at 20 seconds is 108 from home at 28 seconds it's about 138. There are no controller inputs. I'm struggling to see how that is because of controller inputs. If you just imagine the sticks swapped it's the same controller set up. Just swapped the sticks.

 

[skyturnstoblack]

This part of the trip here is the drifting.

 

[Meta4]

You went hands off from 0:17 until 27.8. In that time the only movement was the drone continuing forward for 20 ft and slowing down.

<<Correction>> that distance to slow down should be about 70 ft (I was looking at the distance in metres column)

The distance from home at 20 seconds is 108

Good .. that confirms we are looking at the same time/position.

There are no controller inputs. I'm struggling to see how that is because of controller inputs.

I've already listed all the control inputs above and you are correct, there are no joystick inputs corresponding to the move from west to north.
The Mavic stayed pointing to the west even though it was flying north.
It was reporting its position correctly.
From 17 secs to 27.8 secs the speed was slowing down from 14.8 mph to zero.
Although the direction of travel is to the north, the drone stays facing toward the west.
That makes it look like the drone was in atti mode and pushed by the wind.
But the flight log shows 9 sats and a GPS health of 4 and that the drone was in PGPS mode the whole time.
I'm not sure what to make of that or explain those few seconds when the rest of the flight, before and after appears normal.

If you just imagine the sticks swapped it's the same controller set up. Just swapped the sticks.

So you said .. but it still doesn't make it a simple exercise to read position, speed and yaw angle against control inputs in the data here.

[/QUOTE]

 

[skyturnstoblack]

Yeah all of that is what was confusing. My initial thought was that the building behind me screwed up the gps. But for that to be the case I imagine the gps would have reported not moving as it was actually moving or some other weird GPS data. But honestly it shows almost exactly what my first hand point of view was. I'm wondering if as I clicked through the screen prompts if it for some reason forced it out of a GPS controlled state. What you say seems to shoot that one down. So I'm left nervous it's going to just drift off on me. I probably wouldn't even have noticed it if it wasn't for the prompt stopping me from using the remote for a few seconds. I was hoping someone would just say hey stupid you forgot _____ and I could be ashamed and learn for future flights. And the wind seems like the common theme of ideas which is totally ok with me it's just when something acts differently than expected it makes me feel uneasy about depending on it. Also if I wreck this my wife will be very unkind to my ego.

 

[skyturnstoblack]

Also thanks everyone for trying on this. I have no idea what motivates people to stare at spreadsheets to help other people for free, but I'm glad it happens.

 

[BudWalker]

So I've only owned it for like a week and I've had 2 issues.

The first I think was from me not realizing it didn't have gps lock because the only drones I had played with before were just controller ones with no brains. It took off said it hit it's maximum height at like 10 feet. Started bobbing down right above the ocean I brought it back in and dropped it asap. That one after looking it up was probably because I needed to let the gps lock with 8 sats first.

I watched about 50 youtube videos trying to avoid stupid mistakes but it happened. the next problem was pretty similar but very different.

So I get a home lock I take off and head out the phone gives me the warning about being by my local airport which I called and informed. So as i'm trying to click through that I look up and the mavic air is shooting to the right. Over the water again and towards a large hotel. I pulled the logs and have almost 0 way to understand them. It went about 50 feet. I"m assuming it is gps again i'm taking off by a 12 story building. but am hoping someone can glance at the logs and tell me if one of the sensors needs calibrating. I could calibrate everything but am worried i'll do more damage than good if I just start stabbing in the dark.

It happens right around the 20 second mark. There was a good breeze but it was out to sea not across the beach.

DJI Flight Log Viewer - PhantomHelp.com

This could be a case of launching from a geomagentically distorted site. At 28 secs the left aileron input caused the north component of the velocity to increase indicating the MA was pointing east. But, the Yaw value is a westerly direction. I.e. the Yaw value was compromised which could have come from a geomagnetically distorted launch site.

The xSpeed signal is the GPS's value for the north velocity. Rudder is actually aileron because mode 3 was being used.

I suspect the drifting started because the FC was trying to correct the location using a compromised Yaw value.

The only way to know for sure is to look at the tablet .DAT. Look here to see how to retrieve it.

 

[skyturnstoblack]

This could be a case of launching from a geomagentically distorted site. At 28 secs the left aileron input caused the north component of the velocity to increase indicating the MA was pointing east. But, the Yaw value is a westerly direction. I.e. the Yaw value was compromised which could have come from a geomagnetically distorted launch site.
View attachment 59378
The xSpeed signal is the GPS's value for the north velocity. Rudder is actually aileron because mode 3 was being used.

I suspect the drifting started because the FC was trying to correct the location using a compromised Yaw value.

The only way to know for sure is to look at the tablet .DAT. Look here to see how to retrieve it.

 

[BudWalker]

The part of the flight that we're interested in isn't contained in those 2 ....FLY052.DAT files. Are there any other files with names ending in FLY052.DAT?