correcting instability

[jjlehto 15]

I found out that some aircraft have an incredible tendency to instability on the yaw axis.

At pretty much all speeds, touching the rudder will cause the aircraft to start swinging left and right, making hitting (with guns) anything smaller than an aircraft carrier almost impossible.

I've been trying to fool around with several parameters in the data.ini file, in particular the "damping" ones, and then some of those in the "rudder" section.

Sadly, these changes have brought no real improvement. I even tried (for testing purposes) to substitute the entire "rudder" section with another from another similar plane: in this case the rudder ceased to work altogether (which weirdly enough, made aiming the plane easier).

My question is, which parameters to I need to tweak in order to remove this anomaly?

For example, the MiG-15 (both TW and user made) all have this tendency, which is also unrealistic. The Sabres, the french and the british early jets  are not plagued at all by this problem, their rudder is silky. Historical and technical publications tend to agree that the Fagot would become unstable only nearing mach 1, which is very much a higher speed than those I'm testing it.

Thanks for your attention

[+VonS 1,424]

While my tips are probably more helpful for prop. aircraft, I recommend nonetheless tinkering with the following values outlined below (I've had success with this for my FE2 tweaks, also for prop. aircraft in SF2).

Relevant section to tinker with is of course [Rudder] in the appropriate data.ini.

-----

[Rudder]
SystemType=CONTROL_SURFACE
InputName=YAW_CONTROL
ReverseInput=TRUE
MaxDeflection=25.0 -----> this is often the crudest but most effective value to tweak first (see info. for entry immediately below)
MinDeflection=-25.0 -----> this is often the crudest but most effective value to tweak first (reduce the value in increments of about 2.0 and retest to satisfaction, higher values give more rudder response since greater arc of movement, lower values give smaller arc, etc.; for example, 20.0/-20.0 will give less rudder movement than 25.0/-25.0)
CDdc=0.0073 -----> drag effect on rudder movement (higher value, more drag, lower value, less drag, usually doesn't require tinkering with)
Cydc=0.0362 -----> side force when moving rudder left/right (decrease in increments of about 0.005 and retest to satisfaction, for less effect during rudder movement)
Cldc=0.0010 -----> roll moment when moving rudder left/right (helps to pitch wings up/down during rudder movement, higher value for pitch up, lower for pitch down or neutral, usually avoid tweaking this unless you are flying a staircase as some of us do in FE2 - triplanes, etc., and tweak in increments of about 0.002)
Cndc=-0.0307 -----> yaw moment/strength when moving rudder left/right (decrease towards value of zero in increments of about 0.005 and retest to satisfaction, for less yaw during rudder movement, and obviously increase in same increments for more yaw effect; for example, a value of -0.08 should give more yaw than -0.03, don't be confused by the fact that in math class a negative no. closer to zero is a "bigger" no., it's the other way around for the yaw moment entry in the data inis)

CydcAlphaTableNumData=37
CydcAlphaTableDeltaX=10.00
CydcAlphaTableStartX=-180.00
CydcAlphaTableData=1.000,0.970,0.883,0.750,0.587,0.413,0.250,0.117,0.030,0.000,0.030,0.117,0.250,0.413,0.587,0.750,0.883,0.970,1.000,0.970,0.883,0.750,0.587,0.413,0.250,0.117,0.030,0.000,0.030,0.117,0.250,0.413,0.587,0.750,0.883,0.970,1.000
CldcAlphaTableNumData=37
CldcAlphaTableDeltaX=10.00
CldcAlphaTableStartX=-180.00
CldcAlphaTableData=-1.000,-0.985,-0.940,-0.866,-0.766,-0.643,-0.500,-0.342,-0.174,0.000,0.174,0.342,0.500,0.643,0.766,0.866,0.940,0.985,1.000,1.040,0.871,0.643,0.339,0.170,0.201,0.140,0.185,0.200,0.224,0.294,0.400,0.531,0.669,0.800,0.906,0.976,1.000
CndcAlphaTableNumData=37
CndcAlphaTableDeltaX=10.00
CndcAlphaTableStartX=-180.00
CndcAlphaTableData=-1.000,-0.985,-0.940,-0.866,-0.766,-0.643,-0.500,-0.342,-0.174,0.000,0.174,0.342,0.500,0.643,0.766,0.866,0.940,0.985,1.000,0.971,0.986,0.990,0.718,0.163,0.000,0.000,0.000,0.000,0.112,0.294,0.400,0.531,0.669,0.800,0.906,0.976,1.000
MaxControlSpeed=50.25 -----> in meters/sec (reducing this helps to give more rudder sloppiness at higher speeds, test in increments of about 10.0, also keep in mind that actual speed at which this effect kicks in in FE2/SF2 will usually be about 20.0 or 30.0 m/s higher than value specified for "MaxControlSpeed")
ControlRate=4.0 -----> controls sluggishness/snappiness of control input (test in increments of about 0.5, up or down, for more or less snappiness of moveable control surface, in this case of rudder)
ModelNodeName=Rudder
ReverseModelOrientation=TRUE
AnimationID=1

-----

NOTE 1: "Cydc" and "Cndc" values are in a relationship of sorts (proportional to one another) - usually test changes with only one of those values first, to satisfaction - if no success then change other value and retest - if again no success then tweak both, keeping in mind that more "cydc" will point the nose more quickly in the direction you want to go, and more "cndc" will give you more yaw but will keep you going more so in the same direction - try to visualize the relationship and the changes to get to something realistic - for example, too little "cydc" (flight path direction upon rudder movement) and too much "cndc" (yaw along the same path of flight) sometimes gives a strange "rudder reversal" effect at speeds above about 500 kph in SF2 where "cydc" will start functioning as an inversion of itself (giving opposite direction to rudder input).

NOTE 2: entry above is for the tweaked, late Albatros D.II for FE2 (not to be used for jet aircraft - merely included for info.)

Happy flying,

Von S

Edited August 2, 2020 by VonS
Added info.

[jjlehto 15]

thanks for your reply, this is valuable information.

I'll post back here if I ran into any other doubt while testing

 

[jjlehto 15]

..Success.

I started tweaking the parameters you suggested, only to notice that, even if rudder dynamics were changing, the original problem was not corrected.

It was like in fact the problem would occur at another level, possibly "higher" than the rudder itself.

This made me look upon the "vertail" section, which is parent of the rudder one.

So I simply substituted the aerodynamic parameters contained there with those from another plane with a known functioning rudder (f-86d), in block, just to see if those were the culprits. Ran a single mission, and bang!, the problem was gone. The aircraft still don't feels right, obviously, but I know what to tweak now.

I'll gladly accept any suggestion on the single parameters pertaining to the vertical fin section, as there are even more params than in rudder, and I don't have many clues about them (I would proceed changing them one by one and perform a test between every change). It appears that the vertical empennage has a counterforce that makes the aircraft bounce to the opposite direction to where the rudder has been depressed, roughly with the same force.

The correct behavior would be an aircraft that keeps its nose horizontally in the direction of rudder action, while taking a slight roll in the same direction (to be compensated by ailerons) proportional to rudder force.

Thanks again.

[+VonS 1,424]

Glad the info. was helpful. Yes, often, tweaking of control surface sections (such as the [Rudder]) may require looking into the parent component too - in the case of [Rudder] it's usually called [VertTail] or [VertStabilizer].

Here's the relevant section from the same data ini used as an example in my previous post.

-----

[VertTail]
ParentComponentName=Tail
ModelNodeName=VertStabilizer
DestroyedNodeName=
ShowFromCockpit=TRUE
DetachWhenDestroyed=TRUE
DamageRating=DESTROYED
MassFraction=0.030 -----> the lower the massfraction value, the more fragile the part and more susceptible to ammo. damage, etc. (test in increments of 0.01 approximately, up or down)
HasAeroCoefficients=TRUE -----> leave this as "true" if you want the panel/section to be a part of the overall flight model, otherwise it is purely an aesthetic part
CD0=0.0017 -----> drag value at zero (no) lift
Cyb=-0.0609 -----> side force due to sideslip
Cyp=0.0017 -----> side force due to roll
Cyr=0.0261 -----> side force due to yaw
Clb=-0.0005 -----> roll due to sideslip
Clp=-0.0001 -----> roll damping value
Clr=0.0004 -----> roll due to yaw
Cnb=0.0122 -----> yaw due to sideslip
Cnp=-0.0017 -----> yaw due to roll
Cnr=-0.0177 -----> yaw damping value

-----

For the c-values above, I would tweak in small increments, of about 0.001 (where a bigger no. equals more effect - and as indicated in my previous post, don't be confused by the negative value in some entries, thinking that nos. are getting smaller as you go further below zero - the negative value refers to "opposite" or inverted forces in the TW data inis).

Other values indicated in the section and that I've not included here, such as mach values and alpha nos., I usually don't bother touching - except in some cases alpha nos. but only for wing sections (and left/right stabs) - to tamper with rate of climb, if necessary. Often, cutting-and-pasting of mach and alpha entries from the [VertTail] section of an aircraft data ini that works well into another data ini that is problematic solves problems quickly, but might require further tweaking of some more parameters. Good luck with your tweaks.

Von S