RHS Feedback - AFRF
View Issue Details
0004185AFRF[All Projects] Generalpublic2018-06-02 02:282018-06-02 22:18
Cosmogator Deckard 
reyhard 
noneminoralways
newopen 
 
 
No
1.82
No
0004185: Tu-95 Flight Model need to be fixed.
Now Tu-95 has jet fighter climb/acceleration, and very bad turn rate.

Some references for flight model tuning.

1) Aircraft MUST turn (and lower nose) in banked flight. Playing with draconicTorqueXCoef should help. Simple -- it must acquire 2g at 60deg bank, or turn time 50s for 360deg turn. Without changing altitude.

2) I presume that TU-95 is able to reach his max 2,5g, so 360deg turn time at the 500 km/h (around 70deg bank) should be less than 40s. 1.5 min now.

3) Acceleration that I measured was more than 4 m/s2. With 130t mass (tu-95 max 185t) and takeoff thrust 11tf per engine max acceleration should be less than 3 m/s2. Even lesser if we add air resistance and consider that engine thrust must decrease with speed. Acceleration in level flight at 500 km/h should be around 2 m/s2, I think.

4) Climb (500 km/h) now around 50 m/s (1min to 3000) at speed 500. Closest analog Tu-114 (minimal weight) requires 4,2 min for reaching 3000m on nominal power. Climb at full power should not be more than 15 m/s, I suppose.

5) Tu-114 (with minimal weight) requires 1500m for takeoff or landing. Tu-95 now requires around 700m. Altis airport - 1350m. 1000-1100 (for gameplay purposes) should be ok, I think.
No tags attached.
Issue History
2018-06-02 02:28Cosmogator DeckardNew Issue
2018-06-02 10:03reyhardSeveritymajor => minor
2018-06-02 13:15reyhardNote Added: 0007800
2018-06-02 13:15reyhardAssigned To => reyhard
2018-06-02 13:15reyhardStatusnew => feedback
2018-06-02 22:01Cosmogator DeckardNote Added: 0007801
2018-06-02 22:01Cosmogator DeckardStatusfeedback => new
2018-06-02 22:06Cosmogator DeckardNote Added: 0007802
2018-06-02 22:18Cosmogator DeckardNote Added: 0007803

Notes
(0007800)
reyhard   
2018-06-02 13:15   
if I remember correctly draconicTorqueXCoef had some serious negative impact when using some extreme numbers & banking without those extremes was not working.

take off distance was reduced to 700m because otherwise AI was not able to take off. Probably it could be tweaked a little bit but it's veeeery low priority assets so if you have some values to share then I might take a look at them
(0007801)
Cosmogator Deckard   
2018-06-02 22:01   
Now I'm in trouble. I never thought that anyone wants to let default arma AI control such craft.
I mean - when I experimented with it, AI was so eager to perform CFITs, that I thought it should be able only to fly waypoints higher then 1000m and fire missiles.
I never tried to make it takeoff and land, nad today experiment was disappointing. Takeoff is not so bad, AI don't want to land. With both flight models. With my model it was able to land somehow and crushed into the building - this is the best result...

About draconicTorqueXCoef - AI definitely don'l like it and have tendency to lose some height. It does not matter on high altitudes, but when trying to land it may perform CFIT when trying to fly at height 50 m with 70-80 deg bank (WHY he doing this anyway???). Anyway he is unable to land with both models.
If you can remember some more problems - post it here, mb there is some workaround.

I think only thing that is possible here - make two airplanes with defferent FMs - one for AI (with ability to make nice takeoffs and landings), and one for player (AI only for waypoint following and missile launching). But with AI 70deg bank - these flights won't looks real anyway.
(0007802)
Cosmogator Deckard   
2018-06-02 22:06   
armor=50;
        irTarget=1;
        irTargetSize=1.2;
        visualTarget=1;
        visualTargetSize=1.6;
        radarTarget=1;
        radarTargetSize=1.6;
        gearUpTime=26;
        gearDownTime=24;
        wheelSteeringSensitivity=4;
// it should be able to turn around one of main gears.. which is prohibited, ofc.
        cabinOpening=0;
        airBrake=0;
        lightOnGear=1;
        driverCanEject=0;
        class EjectionSystem
        {
        };
        landingAoa="rad 10";
        throttleToThrustLogFactor=3;
// we need this with a new thrust setting
        angleOfIndicence="rad 3";
        landingSpeed=240;
// I removed acceleration, because can't understand how it works.
        takeoffSpeed=290;
        altFullForce=9000;
        altNoForce=20000;
        maxSpeed=920;
        slowSpeedForwardCoef=0.60000002;
        aileronSensitivity=0.34999999;
        elevatorSensitivity=0.44999999;
        rudderInfluence=0.8;
// some experiment with rudder control. May not be interested, but I fly with these now.
        aileronControlsSensitivityCoef=1.2;
        elevatorControlsSensitivityCoef=1;
// there was typo in elevatorControlsSensitivityCoef.
        rudderControlsSensitivityCoef=0.80000001;
        elevatorCoef[]={0,0.6,1.15,2.6,2.7,2.8,2.35,2,1.35,0.6,0.3,0.2,0.1,0,0,0};
// elevatorCoef tuned to imitate AoA manuever limitation in speed less then 500, and max g manuever limitation in speed more 500.
        aileronCoef[]={0,0.40000001,0.75,1,1.1,1.1,1.1,1,0.85000002,0.64999998,0.40000001,0.1,0,0,0,0};
// tried to increase aileronCoef to help AI, when staying in realistic values. Not necessary.
        rudderCoef[]={0,0.2,0.7,1.0,1.2,1.2,0.9,0.7,0.4,0.2,0,0,0,0,0,0};
// some experiment with rudder control. May not be interested, but I fly with these now.
        flaps=1;
        flapsFrictionCoef=0.4;
        gearsUpFrictionCoef=0.3;
// added gearsUpFrictionCoef with editing flapsFrictionCoef to allow normal takeoff/landing. Not necessary.
        draconicForceXCoef=0.5;
// tried to imitate heavier airplane with editing rudderCoef and draconicForceXCoef. Not necessary.
        draconicForceYCoef=0.25;
// some experiment with rudder control. May not be interested, but I fly with these now.
        draconicForceZCoef=0.1;
        draconicTorqueXCoef = 28;
// here it is, problematic coefficient.
        draconicTorqueYCoef[]={1,1.8,1.4,1,0.1,0,0,0,0,1.0,1.4};
// this allow nose down momentum when aircraft near stall speeds, when it should not fly trimmed, and imitate some control loss close to M.
        envelope[]={0,0.2,0.60,1.55,1.82,2.8,4.1,5.55,6.05,6.05,2,1,0};
// a bit experiments. increased take-off speed, removed high-speed lift loss (there should be different effects, no way to emulate them). Still need to be tuned.
        thrustCoef[]={0.45,0.55,0.65,0.7,0.7,0.78,0.82,0.83,0.95,1.05,1,0.5,0.25,0,0,0};
// reduced acceleration and verical speed. Now it accelerates like a 170t plane, and climbs like a 100t... Idk what to do with it. Still better.
(0007803)
Cosmogator Deckard   
2018-06-02 22:18   
Maybe some of these would make some intererst.
It looks like there is something wrong with aircraft mass and air resistance, but i don't kow how arma emulate it.
It looks like some model parameters stored elsewhere.
airlanex and air friction coeffs seems does not work.