[WIP] R.E.L Skylon C2. Alpha Released. FAR config broken. (08 Dec 2014)

[nli2work]

Does that mean the wing mesh origin, or the part origin, because I can't move the CoM all the way to the root; that's really hacky. I think he's saying I have to make the offset move somewhere. To the middle of the root?

basically set the wing position in unity such the stack attach node is at 0,0,0; that's roughly the mid point of the base chord IIRC. you can use the offset vector to shift the attach node. or use CoMOffset to shift the CoM. two paths to the same end.

Are you sure you mean surface area? Because that model is somewhat complex; it has a very large surface area, even without the base. The main mesh is like a pot, then there are a few concentric rings, and the pointy bit.

Not the surface area of everything, just the surface area of the cone that encloses the intake, it's collision mesh basically. minus the base that attaches to the precooler.

[CaptainKipard]

The mid chord sweep angle is nothing like what his calculation would have you believe; not for a trapezoidal shape anyway. That wiki tutorial is just bad.

edit

ok I just did the calculations as shown, and then measured all the wing and control surfaces manually, and I get different numbers for all midchordsweep angles.

Edited December 9, 2014 by Cpt. Kipard

[Climberfx]

Hey Cpt. Kipard

My SAS is on 80 already. and 60 for rotation. I did a bunch os changes, and know is stable to manual fly, because MJ keep oscillating in pitch, don't mater the attitude adjustments i put. Even with Pitch assistive from FAR. (That Helps a lot on manual, in joystick.).

So, manual on joystick, AOA on front canards in -200, and other in 45.

Canards only on pitch, stabilizer only on yaw, and wings control surface on rotate and pitch, with flaps on (15 degrees max for now, put not yet ok).

Sabre with 10 on Gimbal, and SOMA too.

You can fly around, but still not finish.

I put 14k on from liquid tank, and 10k on the tale.

200 on nose MP and 75 on tale MP.

40º on pitch authority on canards, and 40º on the wings control surface too.

40º on the stabilizer too.

So, and because that -200 on AOA, it keeped that canards in all cases pointed to wind speed, so it help a lot on stabilization, been a little hard to flip back the vessel on pitch, but if you loose control on yaw, the things get wrong. Roll a lot when yaw is of center. Another motive why it can't handle on MJ.

Managed to return too (FAR and DR installed). orbit around 100k, slow until periapsis on 20km, point front with 20º up from Vertical velocity, all the way. Will start to doing up again on middle of deceleration, but keep 20º (Now on Mj). So, when it get around 800, 600, 500, will be lower then 20k, that moment the Mj will not be stable anymore, so, but on on Stock SAS and Pitch from FAR and go manual. I could approach it ned landing, but it lost a lot of velocity, near 150, and that velocity it roll to right and lost control. No landing this time. Oh, i use flaps on second position to approach to land.

Orbit, manual, joystick, problem solved. Returning from orbit too.

Now, fine-tuning everything, and trying to solve the landing problem.

If interesting some one, i can put my configs here, but i prefer to advance a little more, and manage the landing problem, then o can share a better setup.

Cheers.

Cpt., if i can suggest, you could keep the model finalizations in this mean time, If you please, like the stabilizer are a full control surface not only on back of it, and the intakes that need to close, the black color. But only if this all is on your next to do list. In the mean time, i and other here, will keep trying and adjusting...

p.s.: Canards pointing to the wind sweep on the way back. I forgot that i did this reentering with almost half tank and the big orange tank in. So when empty, things can change, like this angle of attack on reentering.

Skylon by Climberfx, on Flickr

Edited December 10, 2014 by Climberfx

[CaptainKipard]

I corrected the rudder a while ago actually, I was just waiting until I can fix the FAR patch to release it. I think I'm close. I'll upload it after testing it once at least. Soon.

I'm not sure about the intakes. I tried to animate them but the module I used doesn't seem to work for me.

[Climberfx]

Great news Cpt.

Did you see how B9 do it close? I'm not sure but i think the new ones from Squad close too.

Edited December 9, 2014 by Climberfx

[nli2work]

no need to reexport anything;

just add

node_attach = 0,0,0,-1,0,0,1 for right wing and canards

node_attach = 0,0,0,1,0,0,1 for left wing and canards

will center the CoL

I also added drag model for the Intakes.

{
MODULE
{
name = FARBasicDragModel
S = 22 //Surface Area
CdCurve //Drag coefficient at various angles
{
key = -1 0 //backwards
key = 0 0.2 //sideways
key = 1.0 0 //forwards
}
ClPotentialCurve
{
key = -1 0 //Lift coefficient, decreases with Mach number
key = -0.5 -0.03
key = 0 0
key = 0.5 0.03
key = 1 0
}
ClViscuousCurve
{
key = -1 0 //Lift coefficient, remains nearly constant with Mach number
key = -0.5 -0.2
key = 0 0
key = 0.5 0.2
key = 1 0
}
CmCurve //Moment coefficient
{
key = -1 0
key = -0.5 -0.01 //keeping angle and moment signs the same results in pitch instability; it will try to flip over
key = 0 0 //making them opposite signs results in pitch stability; it will try to angle fully forward
key = 0.5 0.01
key = 1 0
}
localUpVector = 0,1,0
localForwardVector = 1,0,0
majorMinorAxisRatio = 1.53
cosCutoffAngle = 0
taperCrossSectionAreaRatio = 0.288;
CenterOfDrag = 0,0,0 //a vector defining the part's CoD
}
}

@PART[KipEngSkylonIntake]:NEEDS[FerramAerospaceResearch|NEAR]

Now she flies rather well. a bit of a beast to wrangle under 10km. from there up it's pretty smooth sailing to 300km apogee with ~50% of fuel remaining. switched to rocket mode around 25km and 1150m/s. ~5sec burn to circularize with main engines full. plenty of fuel. Monopropellant not as much. Lost control during reentry, never recovered; pilot error I'm sure.

Edited December 10, 2014 by nli2work

[Virtualgenius]

I noticed the RCS jets thrust was set to 5, i checked through a couple of other mods and the RCS was set to 1 so i modified mine and its a lot better on RCS fuel now

[CaptainKipard]

nli2work that module you posted is kinda similar to the automatic one that ferram told me to generate. Did you simplify it from the generated one or did you make it from scratch?

[nli2work]

nli2work that module you posted is kinda similar to the automatic one that ferram told me to generate. Did you simplify it from the generated one or did you make it from scratch?

CoEfficient curves are default;

S

TaperCrossSectionRatio

maxMinAxisRatio

are derived from measurements of the intake.

If it's close to what FAR generated, then that means I did the calculations (well, arithmetics really) correct.

EDIT: I lowered b_2 on the canards to about half of main wing pieces, just by eyeballing the sizes in SPH. less tendency to nose up sharply; but harder to raise the nose on takeoff. trade offs I suppose.

Edited December 11, 2014 by nli2work

[CaptainKipard]

This is what FAR generated automatically for the intake.

@PART[KipEngSkylonIntake]:NEEDS[FerramAerospaceResearch|NEAR]

{

MODULE

{

name = FARBasicDragModel

S = 22 //Surface Area

CdCurve //Drag coefficient at various angles

{

key = -1 0 //backwards

key = 0 0.2 //sideways

key = 1.0 0 //forwards

}

ClPotentialCurve

{

key = -1 0 //Lift coefficient, decreases with Mach number

key = -0.5 -0.03

key = 0 0

key = 0.5 0.03

key = 1 0

}

ClViscuousCurve

{

key = -1 0 //Lift coefficient, remains nearly constant with Mach number

key = -0.5 -0.2

key = 0 0

key = 0.5 0.2

key = 1 0

}

CmCurve //Moment coefficient

{

key = -1 0

key = -0.5 -0.01 //keeping angle and moment signs the same results in pitch instability; it will try to flip over

key = 0 0 //making them opposite signs results in pitch stability; it will try to angle fully forward

key = 0.5 0.01

key = 1 0

}

localUpVector = 0,1,0

localForwardVector = 1,0,0

majorMinorAxisRatio = 1.53

cosCutoffAngle = 0

taperCrossSectionAreaRatio = 0.288;

CenterOfDrag = 0,0,0 //a vector defining the part's CoD

}

}

[CaptainKipard]

Is there any benefit to using one of those over the other?

[Climberfx]

Hey Cpt. Kipard.

Any news on the final FAR setup and model?

Can't wait to put it on my gameplay. In the actual setup and little changes i made, i couldn't land the thing.

[CaptainKipard]

I can release the next revision for everyone to test out if they want. I've made most of the changes I needed already, but I haven't tested it myself yet.

There is no "final" set up yet. I'm under no illusion that refining everything will be anything but a long process.

Edited December 13, 2014 by Cpt. Kipard

[Climberfx]

I Want. so that way i can match up if the changes i made, and try to make a way to land it and be 100% returnable.

[CaptainKipard]

Here. I'm getting pretty sick of this. This pos wont even fly straight now.

[blowfish]

Here. I'm getting pretty sick of this. This pos wont even fly straight now.

Under FAR? Might it have to do with dynamic stability? Based on my experience with FAR, I doubt that the Skylon would be dynamically stable by default. Though this raises the question of whether the RL version is supposed to be dynamically stable, with good aerodynamics being replaced with advanced PIDs so often these days.

Still though, if this is the case, it might be worth contacting ferram4 directly, since he's the only one who might possibly know how to make this fly correctly.

[Climberfx]

He was already contacted...

[Virtualgenius]

So will the release still fly without far or near installed as I dont use either and have no intention of using them

[nli2work]

how did the CoL get so far ahead of CoM?

[CaptainKipard]

Under FAR? Might it have to do with dynamic stability?

Yes. I don't know.

So will the release still fly without far or near installed as I dont use either and have no intention of using them

Yes

how did the CoL get so far ahead of CoM?

I don't know and at this point I don't really care. This is too frustrating.

Here are the measurments if you want to work it out yourself.

Edited December 14, 2014 by Cpt. Kipard

[Climberfx]

Cpt. Kipard, you must care. this ship is amazing and have a tremendous possibilities for all the users.

You must have a version for FAR, because it is used a lot, and the ones who really care about the Flight will ask for it.

Now you shared the measurements i will try some more to help you.

But we need to consider that we don't really know if the COL and COM is really there in the real world, and how is the weight distributed.

I would consider changing all the masses to have a COM ahead of COL a little bit, and the center of mass exactly in middle of cargo bay, unchangeable by expending resources.

May the Force be with You!

[CaptainKipard]

Here's the deal.

AFAIK there are only two mods in the whole community that have wings with stack nodes, and the CoM in the place where it should be, i.e. the middle of the wing. Those mods are. Mini Shuttle, and Skylon. All other wings have the CoM at the root of the wing, and are surface attachable.

The makers of Mini Shuttle forewent FAR compatibility. I don't know why, but it's just odd that there's this pattern emerging. B9 has no problems with FAR, and their wings are like stock wings. There are also FAR configs for other mods with wings that I can't remember right now.

I wont be making this FAR compatible any more until I can expect some consistent behaviour. FAR was fun to play with but it's not essential to me. I'm continuing development with only stock in mind. Sorry.

If you don't like it, bug ferram. And keep in mind, Skylon was designed by real engineers. It should fly just fine, so either I'm doing something wrong (ferram himself said the configs are probably fine) or FAR is not fit for this particular purpose.

This is all speculation mind you. Everyone is welcome to make their own FAR config for this, but don't tear your hair out doing it.

Edited December 17, 2014 by Cpt. Kipard

[Voculus]

Sorry to hear of your frustrations, Kipard. At any rate, it's still a gorgeous work of art!

[blowfish]

I've improved the FAR cfg to the best of my abilities according to ferram4's suggestions and other tidbits in this thread. It's still incredibly flip-happy, owing to the CoL being in front of the CoL, but it's almost manageable with SAS. It still probably needs a lot of work. I might PM ferram about it at some point since my understanding of aerodynamics is more or less limited to what can be found on this forum.

Other notes from my testing:

• The vehicle mass is a lot lower than the real Skylon. The thrust might be higher than it needs to be in certain regimes too. The REL trajectory data suggests that the starting TWR is less than 0.5, whereas I'm getting more like 2.5 right now.
  
• The front landing gear doesn't turn that much, making it very hard to perform even basic taxi maneuvers. Not that you ever need to do that in KSP...
  

@PART[KipEngSkylonWing*]:NEEDS[FerramAerospaceResearch|NEAR]
{
@module = Part
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
!MODULE[ModuleLiftingSurface] {}

MODULE
{
name = FARControllableSurface
MAC = 10.385
MidChordSweep = 17.165
b_2 = 4.02
TaperRatio = 0.56
e = 0.87
nonSideAttach = 0
maxdeflect = 15
ctrlSurfFrac = 0.12
}
}

@PART[KipEngSkylonWingLeft]:NEEDS[FerramAerospaceResearch|NEAR]
{
node_attach = 0, 0, 0, 0.994, 0, -0.107
@MODULE[FARControllableSurface]
{
rootMidChordOffsetFromOrig = 1.8, 0.845, -0.18
}
}

@PART[KipEngSkylonWingRight]:NEEDS[FerramAerospaceResearch|NEAR]
{
node_attach = 0, 0, 0, -0.994, 0, -0.107
@MODULE[FARControllableSurface]
{
rootMidChordOffsetFromOrig = -1.8, 0.845, -0.18
}
}

@PART[KipEngSkylonCanard*]:NEEDS[FerramAerospaceResearch|NEAR]
{
@module = Part
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
!MODULE[ModuleLiftingSurface] {}

MODULE
{
name = FARControllableSurface
MAC = 1.985
MidChordSweep = 60.642
b_2 = 1.8
TaperRatio = 0.25
e = 0.87
nonSideAttach = 0
maxdeflect = 15
ctrlSurfFrac = 1
}
}

@PART[KipEngSkylonCanardLeft]:NEEDS[FerramAerospaceResearch|NEAR]
{
node_attach = 0, 0, 0, 1, 0, 0
@MODULE[FARControllableSurface]
{
rootMidChordOffsetFromOrig = 0.68, 0.8, 0
}
}

@PART[KipEngSkylonCanardRight]:NEEDS[FerramAerospaceResearch|NEAR]
{
node_attach = 0, 0, 0, -1, 0, 0
@MODULE[FARControllableSurface]
{
rootMidChordOffsetFromOrig = -0.68, 0.8, 0
}
}

@PART[KipEngSkylonStabiliser]:NEEDS[FerramAerospaceResearch|NEAR]
{
@module = Part
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
!MODULE[ModuleLiftingSurface] {}

node_attach = 0, 0, 0, 0, 0, -1

MODULE
{
name = FARControllableSurface
MAC = 3.27
MidChordSweep = 89.227
b_2 = 3.8
TaperRatio = 0.213
e = 0.87
nonSideAttach = 0
maxdeflect = 15
ctrlSurfFrac = 1
rootMidChordOffsetFromOrig = 0, 1.7, -1.48
}
}

@PART[KipEngSkylonSabreEngine]:NEEDS[FerramAerospaceResearch]:AFTER[FerramAerospaceResearch]:NEEDS[!AJE]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
@MODULE[ModuleEnginesFX],0
{
@maxThrust *= 1.3
}
}

@PART[KipEngSkylonIntake]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
MODULE
{
name = FARBasicDragModel
S = 22
CdCurve
{
key = -1 0
key = 0 0.2
key = 1.0 0
}
ClPotentialCurve
{
key = -1 0
key = -0.5 -0.03
key = 0 0
key = 0.5 0.03
key = 1 0
}
ClViscuousCurve
{
key = -1 0
key = -0.5 -0.2
key = 0 0
key = 0.5 0.2
key = 1 0
}
CmCurve
{
key = -1 0
key = -0.5 -0.01
key = 0 0
key = 0.5 0.01
key = 1 0
}
localUpVector = 0, 1, 0
localForwardVector = 1, 0, 0
majorMinorAxisRatio = 1.53
cosCutoffAngle = 0
taperCrossSectionAreaRatio = 0.288;
CenterOfDrag = 0, 0, 0
}
}

@PART[KipEngSkylonPayloadBay]:NEEDS[FerramAerospaceResearch]
{
@name = KipEngSkylonCargoBay
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
!MODULE[ModuleLiftingSurface] {}
@RESOURCE[LiquidOxygen]
{
@amount = 11000
@maxAmount = 11000
}
}

@PART[KipEngSkylonNoseFuel]:NEEDS[FerramAerospaceResearch]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
@RESOURCE[LiquidHydrogen]
{
@amount = 12000
@maxAmount = 12000
}
}

@PART[KipEngSkylonTailFuel]:NEEDS[FerramAerospaceResearch]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
@RESOURCE[LiquidHydrogen]
{
@amount = 12000
@maxAmount = 12000
}
}

@PART[KipEngSkylonCommand1]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonCommand2]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonMainGearLeft]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonMainGearRight]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonNoseGear]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonNoseMP]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonTailMP]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonOrbitalEngine]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonPrecooler]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

@PART[KipEngSkylonSideDoor]:NEEDS[FerramAerospaceResearch|NEAR]
{
@maximum_drag = 0
@minimum_drag = 0
@angularDrag = 0
}

[CaptainKipard]

Thanks!

People have been reporting a lot of resources left over at even a HKO. It occured to me recently that maybe the wings are stalling because the extra mass is requiring a high AoA. I've played with the stall visualisation in FAR and at the time of testing the wings were very red.

What's a good gear turning angle?