investigating The rudder should not be using sideslip feedback for yaw dampening (causing a very slow yaw response)

[DummyCatz]

Our current Auto Flap Up FCS implementation is using sideslip feedback to control the rudder, which I consider as a bug, because the feedback is erroneously applied to the rudder rather than to the aileron and differential stabilators. 

The sideslip and sideslip rate feedback should be fed to the aileron and diff-stab above 20 deg AOA, as in FCC OFP v10.7 IRL. For now, I'm seeing rudders move with sideslip changes even if AOA < 20°.

 

Test procedure:

1. Set an extremely turbulent weather.

2. Flaps to auto.

3. Press F4 to get a closer look at the rudder

4. Check if the rudder is moving with turbulence/sideslip changes. Use active pause so that there's no lateral acceleration and yaw rate interference to the rudder. Test with AOA below or above 20 degrees.

 

References:

1. Park, David J., "Development of F/A-18 Spin Departure Demonstration Procedure with Departure Resistant Flight Control Computer Version 10.7. " Master's Thesis, University of Tennessee, 2004. https://trace.tennessee.edu/utk_gradthes/2312

2. Mitchell, Eric John, "F/A-18A-D Flight Control Computer OFP Versions 10.6.1 and 10.7 Developmental Flight Testing: Out-of-Controlled Flight Test Program Yields Reduced Falling Leaf Departure Susceptibility and Enhanced Aircraft Maneuverability. " Master's Thesis, University of Tennessee, 2004. https://trace.tennessee.edu/utk_gradthes/2372

3. Simulation Model of a Twin-Tail, High Performance Airplane, NASA TM 107601, including a set of FCC OFP v10.1 block diagrams (https://ntrs.nasa.gov/api/citations/19920024293/downloads/19920024293.pdf)

4. NATOPS manual which is not quoted here but contains relevant info.

 

According to the Directional Auto Flap Up CAS block diagram from reference 3, there's no sideslip or sideslip rate feedback in v10.1 as there was no sideslip measurements available to the aircraft. The feedbacks were only included in v10.6.1 (a test version of 10.7) and v10.7, together with the sideslip estimator.

From reference 2 describing the sideslip estimator:

Quote

The following features were incorporated in v10.6.1 (Heller, 2003):
1. Sideslip Rate Feedback: An accurate estimator of sideslip rate (β dot) can be calculated from true airspeed, g, Ny (lateral g), pitch and roll angles and rates, and AOA.
2. Sideslip Feedback: With the production F/A-18A-D lacking a sideslip probe, sideslip (β) is calculated by simply integrating the β dot estimate discussed above.

From reference 1 describing the sideslip and sideslip rate feedback:

Quote

The most significant upgrades to the Flight Control Computer Operation Flight Program were the incorporation of sideslip and sideslip rate feedback to the ailerons and differential stabilators. Previous to v10.7, the usual cause of departure in the Hornet was due to the increasing roll or yaw as a result of increasing sideslip. This increase in roll or yaw eventually overcame the control surface authority and resulted in departure from controlled flight. Therefore, the key to departure prevention in the Hornet was to minimize the sideslip with control surfaces before it became a problem. Since the F/A-18 lacks any external measuring equipment to measure the actual sideslip, estimates must be computed.

Also from reference 1:

Quote

A total of 62 tailslides were performed during the v10.7 evaluation and many more have been performed to date during departure demonstration flights. There were many cases where the departure conditions may have resulted in severe motion with v10.5.1, but no sustained out-of-control motion, such as sustained spin or sustain falling leaf, was observed with v10.7. In general, any rolling and yawing motion would quickly damp whenever the angle of attack cycled high during the oscillations. This is primarily a result of sideslip and sideslip rate feedback driving the ailerons and differential stabilators to damp the roll and yaw motions. With v10.5.1, departure and recovery motions were unpredictable and severe while the altitude loss generally ranged from 8,000 feet to 12,000 feet with some extreme cases exceeding 20,000 feet.[6] With v10.7, departure and recovery motion have become very mild and predictable while the altitude loss has consistently ranged from 8,000 feet to 10,000 feet. Vertical departure and recovery motion with v10.7 could be categorized in two typical examples: upright recovery and inverted recovery.

Hronet rudder moves with sideslip.trk

Edited November 18, 2023 by DummyCatz

[DummyCatz]

• Update flight model for ground effect, takeoff pitch effects, auto-pilot based on FPM, touch and go handling, and other remaining flight model issues [In Progress]

Hi, I've already noticed that the Flight Model and FCS has been WIP for more than 5 years. I may ask, what are the other remaining issues and has this particular bug been reported? I'm not seeing the reasoning behind the 'as stated in the road map' cuz I'm doing a bug report (regarding the rudder FCS) of an implemented feature (sideslip feedback), not a wishlist item that is not implemented (like missing logic in FCS). It's basically including all subsequent FM related bug reports into this category automatically, and it's just like saying that all the remaining issues that could possibly exist are WIP. (of course)

The fix should be simple and it’s not related to the aerodynamics. Just pure control logic:

Remove the sideslip feedback to the rudder, and add it to the aileron and diff-stab instead. See, it’s not a missing logic, but a bug.

Edited November 19, 2023 by DummyCatz

[atercygnus]

Thanks ED for the new hornet update! I appreciate all the hard work, done by the team.

Nevertheless, this issue wasn`t addressed despite the fact it was planned to.

Any comments?

[atercygnus]

@BIGNEWY any news for that one?

[BIGNEWY]

Hi, 

I have asked the team to review this topic. 

thank you