wobbles behavior roll axis Fixed Wing

[Fabio-FPV]

I love INAV, and everything works perfectly except for those slow oscillations on the roll axis that occur when the wind speed is very low and the plane should be flying well-stabilized, but instead it oscillates left and right a few degrees slowly and continuously in CRUISE, CRSH, and RTH modes, making flying unpleasant.
Of course, I've tried everything with the PIDFF settings, and paradoxically, even setting the P, I, and D term to zero and changing the FF and roll rate values higher or lower, the behavior remains the same.
It seems as if the plane remains within a dead band on the roll axis and only takes action when it exceeds a certain threshold (degrees of bank).
At this point, I can't figure out if this is a physiological behavior of the algorithm managing the stabilized mode with altitude control, or if it's something that can be improved but I don't know how to find a solution.
I would like to point out that I have done many autotuning tests, I have flown with a calibrated compass several times and without a compass, but the behavior remains the same.
I hope somebody can help me to save my FPV Passion in INAV !
Thanks in advance for find a solution or to understand better the cause of this anomaly.

ARWING PRO INAV 8 Diff All 02_01_2026.txt

Attached is the link to the video where the anomaly is clearly visible.

https://youtu.be/5RGsvLhyhAc

[Jetrell]

@Fabio-FPV Looking at your magnetometer calibration. It isn't good. The X and Z zero offsets are huge. Likely caused by something magnetic in the plane interfering. Its hard to say for sure without seeing the hardware layout.

Look here for what you should expect to see after a good calibration.

Other links to help with tuning of fixedwing navigation PID controllers can be found here.
But don't touch them until you have the magnetometer calibration correct; if the issue is still there.

The other thing you can do for an easy test is to disable the mag. And go for a flight. If the issue is gone, that will confirm it.

[Fabio-FPV]

Yes is aligned respect the yaw axis.

Anyway i am confident that changing one of those values i hope will change the behavior because I noticed that in angle mode (where is not present the issue) the normal PID have affect but no in cruise/RTH/WAYPOINT mode and if in flight i will notice some effect for sure that will be the ruote right for improve or find the solution.

Thank again @Jetrell . I'll let you now!

[Fabio-FPV]

Hi @Jetrell ,
I finally found a light-wind day to fly and test the in-flight changes to the parameters you suggested. Unfortunately, these didn't solve the problem because, as you can see from the video, they didn't change the aircraft's often noticeable oscillating behavior. In the end, I've come to the conclusion that this, which I find an annoying problem, is actually the normal behavior of INAV in calm winds. I'd certainly like to improve it, but since I've already tried everything I can, I hope that making the video and log file available can at least help improve INAV in future versions. My suggestion is to be able to adjust the aircraft's roll response more or less aggressively in automatic modes without creating that annoying dead band where the aircraft oscillates inside. Thank you in advance for any feedback and comments.

https://youtu.be/LZbGzgn51Os
https://drive.google.com/file/d/1z_42kGtxyc3-Fqc2yvyHRF5SJFu-vbBF/view?usp=sharing

[Fabio-FPV]

Hi @Jetrell,
I apologize for the previous links not working. These links are now correct.
Best regards.
https://youtu.be/LZbGzgn51Os
https://drive.google.com/file/d/1z_42kGtxyc3-Fqc2yvyHRF5SJFu-vbBF/view?usp=sharing

[Jetrell]

I had a quick look through the video and log you left. And the issue is also present in Angle mode. This can be seen by the slow oscillating of the roll feedforward.
When I first looked at your Diff settings. I thought the Rates and Feedforwards look out of sorts from what I would normally expect to see on most elevon planes. But I just assumed you must have had very low control surface throws.

I'd recommend running autotune again.
The roll and pitch axis's have to reach the maximum rotation rate the plane is capable of at full stick deflection. To get a correct tune... Not just a quick snap left and right with the stick.

[Fabio-FPV]

Hi @Jetrell ,
I've performed several autotunes, performing complete rotations on both the pitch and roll axes at full stick until the OSD value remains more or less stable. I also tried changing the mechanical throw physically modifying the linkage of control surface. AT the end, I also tried manually adjusting the feedforward and maximum rate in flight via the OSD. Performing this last operation, I noticed that the oscillatory behavior didn't change, neither increasing or decreasing the FF or the maximum rate. For this reason, I finally decided to ask for help. Anyway, I'll make a video where I redo the autotune so you can see how I do it. Anyway, I wanted to thank you so much for your time and help, hoping to perhaps find a solution. Thanks so much, and I'll let you know as soon as possible.

[trailx]

I think this is a P oscillation in the nav XY controller. The nav controller works much slower than we're used to in the PID stabilization loop, so its oscillation speed is slower. Due to the speed, I thought it could be an I-term oscillation, but the I-term in the PID is low, the oscillation is seen in the setpoint, and the I-term in the nav controller isn't oscillating.

What is oscillating in the blackbox log is the P-term for the FWposP. Which I believe is tied to nav_fw_pos_xy_p. Same as the top-left term in your OSD. Your setup is just more responsive to roll, and requires a smaller number. Considering you're at 75 and seeing the oscillation, I'd try reducing it by half, to like 37 for starters. But thats just a guess. See if you can eliminate the oscillation by reducing it.

You can see the BB log below. The bottom half is the nav controller. If the nav controller is oscillating, that generally means its driving the stability PID oscillations, plus the oscillations you're seeing are just too slow to likely be originating in the stability PID or in your FF/rates (as would be affected by autotune).

I also saw you were adjusting HP in your OSD. What variable is that exactly? I'm not super familiar. Is that a MC variable?

[trailx]

In your video you were adjusting the HP value. Do you know what that is tied to? I'm not familiar with it. I suspect it could be the "horizontal position" variable? If so, I think that is a multirotor value, and thus why it doesn't do anything on a fixed wing.

[Fabio-FPV]

I'm not sure what the HP value, which is called heading hold in the configurator, does. However, the only effect changing nav_fw_pos_xy_p in the video has is making it much slower to find the route during RTH engagement if I lower it. I'd like to change what the configurator writes as "Nav Heading," but it's not possible to change it in flight. I just hope I can find that damned parameter that changes the oscillation behavior during automatic modes because so far I haven't been able to find it (if it exists).

[trailx]

Based on context clues, it looks like the Heading Hold is linked to nav_fw_heading_p

The row beneath it, NAV Heading, appears to be linked to: nav_fw_pos_hdg_p, nav_fw_pos_hdg_i, and nav_fw_pos_hdg_d

Oddly enough, there is also nav_fw_pos_xy_p, nav_fw_pos_xy_i, and nav_fw_pos_xy_d.

I wish I knew, but I'm not sure what the exact difference is between the pos_hdg and pos_xy PID controllers, and I don't exactly know how these three systems work together. But it looks like the heading controller works more to maintain a direction, while the xy may be more important for other maneuvers.

It sounds like you flew it, did the slow oscillation go away when you reduced nav_fw_pos_xy_p? I assumed that was the issue because blackbox showed an oscillation in what it called fwPosP

Based on what you said, you may want to try reducing nav_fw_pos_hdg_p by half, and see what that effect has. That may be the culprit. This does look like a slow P-oscillation, regardless... I think its a matter of finding the right one.

[trailx]

Oh, and you may want to consider doing this via CLI rather than trying to figure out which is which in the configurator tabs, which somewhat confounds what variable is what.

[Fabio-FPV]

Hi @trailx,
(sorry, in another video) when I modified the nav_fw_pos_xy_p parameter in flight, I had reached the lowest value of 50, but unfortunately the day was a bit windy and it wasn't clear whether the anomaly had improved. However, it was clear that while I engaged RTH, the settling onto the home track after the turn was much slower. The only change after all these attempts this year that significantly improved the behavior, as you can see now, was setting all the YAW PID values ​​(PIDFF) to zero and the YAW rate to minimum, otherwise the oscillation is even more noticeable (which is paradoxical because the YAW shouldn't be involved in an elevon). However, in the configurator and in the INAV documentation, it's not entirely clear what certain specific parameters on the page do (Additional PID Gain), and especially whether they're involved in the FW or MC. I hope to find another calm day as soon as possible to retest the nav_fw_pos_xy_p parameter in flight, decreasing it as you suggested, and then I'll post the video with the result. Thanks again for your time and help. Happy flying!