Aircraft Yaw CW - Neutral Controller

[Amarand]

The yaw behavior looks fine until around 400 seconds, after which there does appear to be an offset in the rudder stick input, which leads to a steady CW rotation:

View attachment 77386

Since it starts mid-way through the flight it seems unlikely to be a stick calibration issue or a magnetic connector. It's very steady, which rather rules out an external magnetic field. I'd suspect a faulty (sticky?) stick module.

So a decent way to test this without any additional hardware would to fly/hover direct from iPhone over WiFi? Or, as mentioned previously, hover with gestures? See if I can hover on a full battery without any yaw?

Then, after testing that, reintroduce the RC to see if CW yaw returns?

 

[Amarand]

The yaw behavior looks fine until around 400 seconds, after which there does appear to be an offset in the rudder stick input, which leads to a steady CW rotation:

View attachment 77386

Since it starts mid-way through the flight it seems unlikely to be a stick calibration issue or a magnetic connector. It's very steady, which rather rules out an external magnetic field. I'd suspect a faulty (sticky?) stick module.

Also, may I ask which app you’re using to visualize the .DAT file? I’m fairy technical, could watch my testing after the fact, iterate.

 

[sar104]

Also, may I ask which app you’re using to visualize the .DAT file? I’m fairy technical, could watch my testing after the fact, iterate.

Actually the plot that I showed was from the txt log, not the DAT file. I looked at the DAT file to check for yaw errors, but the MA mobile device doesn't include the RC stick inputs for some reason. The program I'm using is Wavemetrics Igor Pro, which you probably won't have, but you can look at both the txt and DAT logs with @BudWalker's CsvView. You will need to compare OSD_yaw with RC_rudder.

 

[sar104]

So a decent way to test this without any additional hardware would to fly/hover direct from iPhone over WiFi? Or, as mentioned previously, hover with gestures? See if I can hover on a full battery without any yaw?

Then, after testing that, reintroduce the RC to see if CW yaw returns?

That sounds like a good plan.

 

[BudWalker]

Actually the plot that I showed was from the txt log, not the DAT file. I looked at the DAT file to check for yaw errors, but the MA mobile device doesn't include the RC stick inputs for some reason. The program I'm using is Wavemetrics Igor Pro, which you probably won't have, but you can look at both the txt and DAT logs with @BudWalker's CsvView. You will need to compare OSD_yaw with RC_rudder.

That's strange. Works for me with version 3.6.6. ????

 

[sar104]

That's strange. Works for me with version 3.6.6. ????
View attachment 77389

I was using 3.6.3 on a different Mac. You are right - the RC data convert just fine with 3.6.6. Thanks.

 

[Amarand]

That sounds like a good plan.

So I took the aircraft up for a full battery with my iPhone alone, and it hovered perfectly. Flew it around a bit, then brought it back, perfect hover. I couldn't do anything speedy (like I normally do), because I couldn't figure out how to put the phone's controller into Sport mode. Lots of obstacles in my backyard, so it goes fairly slow. (Which is why I use Sport mode in the backyard to begin with.)

I then replaced the battery, let the craft cool down a bit, and re-paired it with the RC. Flew on a full battery as normal. When I took it up to 300' AGL, it started to have a minor yaw in the CCW direction (hands off sticks). But when I brought it back down, I couldn't reproduce the behavior.

I'll do some additional testing over the next few days, and also see if I can't secure one of the analysis software packages recommended.

If nothing else, this got me out of my comfort zone, and using my phone as a controller. Less than ideal (I much prefer physical controls) but at least it allowed me to test.

 

[sar104]

So I took the aircraft up for a full battery with my iPhone alone, and it hovered perfectly. Flew it around a bit, then brought it back, perfect hover. I couldn't do anything speedy (like I normally do), because I couldn't figure out how to put the phone's controller into Sport mode. Lots of obstacles in my backyard, so it goes fairly slow. (Which is why I use Sport mode in the backyard to begin with.)

I then replaced the battery, let the craft cool down a bit, and re-paired it with the RC. Flew on a full battery as normal. When I took it up to 300' AGL, it started to have a minor yaw in the CCW direction (hands off sticks). But when I brought it back down, I couldn't reproduce the behavior.

I'll do some additional testing over the next few days, and also see if I can't secure one of the analysis software packages recommended.

If nothing else, this got me out of my comfort zone, and using my phone as a controller. Less than ideal (I much prefer physical controls) but at least it allowed me to test.

If you connect to the DJI Assistant software you can check the stick calibration and see the inputs in real time. Is it possible that you have anything magnetic in the vicinity of the controller while you are flying?

 

[Amarand]

That's strange. Works for me with version 3.6.6. ????
View attachment 77389

Oh that CsvView program is fun. So much data!

How are you getting the Y-axis grid lines to show up in the plot area?

 

[Amarand]

If you connect to the DJI Assistant software you can check the stick calibration and see the inputs in real time. Is it possible that you have anything magnetic in the vicinity of the controller while you are flying?

I'll connect the RC to DJI Assistant Software now, and re-calibrate the sticks.

Is my Apple Watch 2 considered "magnetic?" I wear it on my left wrist, which is where the left stick is....

 

[sar104]

I'll connect the RC to DJI Assistant Software now, and re-calibrate the sticks.

Is my Apple Watch 2 considered "magnetic?" I wear it on my left wrist, which is where the left stick is....

Not the watch itself, but if you have one of the straps with a magnetic clasp then that would certainly be enough to affect the stick sensors.

 

[Amarand]

I'll connect the RC to DJI Assistant Software now, and re-calibrate the sticks.

Odd.... When I bring up "RC for Mavic Air" under Connected Devices, I only have two options: firmware and tools. Tools only shows "Export Log." Am I supposed to have an RC calibration there? Or is that the DJI GO app? Sorry for the confusion....

 

[Amarand]

Not the watch itself, but if you have one of the straps with a magnetic clasp then that would certainly be enough to affect the stick sensors.

No, not the Milanese Loop. Too fancy for my tastes. I just have the stock silicone strap.

 

[sar104]

No, not the Milanese Loop. Too fancy for my tastes. I just have the stock silicone strap.

That's not going to do it then. I really like the Milanese, but it tends to shake loose so not suitable for sports (or flying drones), so I mostly wear the stock one too.

 

[Amarand]

Is it possible that you have anything magnetic in the vicinity of the controller while you are flying?

I went back and watched the Nest camera recording of the flight, and I am basically holding the RC in my left hand, thumb an inch or two from the stick at all times. Decades of "trigger discipline" make me aware of when I'm intentionally touching the stick or not. And during that two minute test, I was thumbs-off.

 

[Amarand]

That's not going to do it then. I really like the Milanese, but it tends to shake loose so not suitable for sports (or flying drones), so I mostly wear the stock one too.

My problem with the Milanese is that it seems like it would cut into my skin. I'm sure that's just perception and not reality. But yeah, I also would not enjoy removing my Watch while flying, calibrating, etc.

 

[BudWalker]

Oh that CsvView program is fun. So much data!

How are you getting the Y-axis grid lines to show up in the plot area?

Do a right mouse click on the IMU_ATTI(0):Yaw label, then select Y Grid

 

[sar104]

My problem with the Milanese is that it seems like it would cut into my skin. I'm sure that's just perception and not reality. But yeah, I also would not enjoy removing my Watch while flying, calibrating, etc.

No - it's very comfortable.

 

[silverdragon]

Problem? Aircraft has a propensity to yaw clockwise with hands off the sticks

Was unit in a crash? No

What have you tried so far?: Full calibration in an electronic and metal-free backyard on grass (compass, IMU, RC, gimbal)

What device are you using? Mavic Air RC with iPhone Xs Max

What firmware are you running ( aircraft, remote controller)? Most current

What Go app version are you using? Latest version in App Store (4.3.20)

Any modification? Stock unit with Polar Pro landing gear

Landing Gear | Mavic Air

The PolarPro DJI Mavic Leg Extensions raise the camera 1 inch off the ground keeping it safe from dust / dirt / sand. The DJI Mavic Air Leg Extensions are made of

www.polarprofilters.com

Did you change anything or install any apps? I have only used the DJI GO 4 APP with this aircraft

Do you have a video or pictures of the problem?

(I hope this is what I was supposed to do for the log file!)

DJI Flight Log Viewer - PhantomHelp.com

Mine has been doing that for several months now. About 12 degrees to the right and 6 or so when it magically yaws to the left.

 

[Amarand]

No - it's very comfortable.

Ahh, cool.

Also, I downloaded an app for my iPhone called Tesla Bot. In the area where I use the controller, there are only minor variations. Same with my hand/arm/Apple Watch. I brought the iPhone up to my front door (where there are dozens of Neodymium magnets) and the phone went wild. ?