crzyrndm / b9-pwings-fork Goto Github PK

@Crzyrndm please ignore for now - this is the same issue as discussed on the forums, posted here for the benefit of others who check github first. (If that's you - the forum link has relevant information for 2 pages ATM).

The parts appear textured and striped in the part browser; and have pitch-black surfaces when placed, with a normal texture on the side. Example screenshot by Konnor.

The issue seems to originate from Unity, possibly compounded by video drivers (Unity only claims support for vendor-provided drivers AFAIK).

JIC, my system:

• Arch Linux 64-bit;
• open-source Radeon drivers (xf86-video-ati 7.7.1);
• KSP 1.1.3 (started with LD_LIBRARY_PATH="/usr/lib:$LD_LIBRARY_PATH" LC_ALL=C ./KSP.x86_64 -force-glcore45 -adapter 0);
• B9-PWings-Fork 0.40.7 (installed via CKAN as a dependency for RP-0).

I've been helping test RSS/RO on 1.1.2  Using the latest fork of this as linked above, I'm seeing two problems: about half the time in the VAB, the option to press J and configure the wings doesn't appear.  Leaving the VAB scene and coming back will usually, but not always, fix this.  Also, and perhaps more importantly, it's doing something very wonky with mass.  As a test case, I made a sample vessel, nothing more than a probe core and 4 unmodified pwings.  Vessel mass as shown by mechjeb was 2.656t in the VAB, but then on the pad it shows 4.570t.  This was easily reproducible, only happened to vessels with pwings on them, and the difference in vab and pad mass scaled based on how large the wings were.

No matter how small I set the wings to, the engineer report always reports the wings as 4m. http://i.imgur.com/ItqqM7P.jpg is an example, the tank is 2.1m in diameter, with the wings set to 3m, total span should be 8.1m, but is reported as 10.1m, not allowing me to launch the craft. This is under a Realistic Progression Zero set of mods with version 2.1 of the B9 Wings mod.

Thickness not change the wing mass. I think it should.

Also it seems that the thickness doesn't affect the coefficient of drag value, which I think it should. too

Hi there,

The only thing pending for this mod to work on KSP 1.4.x is the shader.

Not sure if you plan to upgrade the shader to Unity 2017 but if you don't have time I'm willing to do it.

But I will need the source code of wingShader (KSP/Specular Layered)

I haven't found it looking at your repo ;(

I posted in your thread, but I figured that I should post here too. When this mid is installed alongside RO and it's suggested/recommended mods through CKAN, the wingshader.ksp file stops the KSP EDIT from functioning. Removing or moving the shader file into its own folder solves the issue, but breaks B9 Pwings. Output logs below.

Pwings not installed - https://www.dropbox.com/s/z96756wa94gqo9j/output_log%20%28No-PWings%29.zip?dl=0

Pwings installed - https://www.dropbox.com/s/adn0rjvoigemug7/output_log%20%28With-PWings%29.zip?dl=0

Rotating the vessel about the vertical axis in the VAB is commonly used to orient the ship for launch along the equator.
This and many other non-default orientations cause the detection of which wing surface is top or bottom to be incorrect and often asymmetrical.

To reproduce in VAB:

1. Place a root part eg FL-T800 fuel tank
2. Attach wings with radial symmetry, perpendicular to the floor of the VAB.
3. Select the fuel tank and rotate the top down so the wings are no longer perpendicular to the floor of the VAB.
4. Continue to rotate about this axis and observe:
   a) The bottom (HRSI surface) is the North-most surface rather than the top-most.
   b) When wings are parallel to the floor, one has top surface facing up, one facing down.
5. Start again and this time attach wings with mirror symetry. Attach wings such that they are not parallel with eachother, ie at a dihedral/anhedral angle.
6. Rotate 90 degrees about the vertical axis and observe errant top/bottom detection.

Setup:

• KSP 1.2.1.1604
• B9-PWings-Fork 0.40.10
• No other mods

When the Deploy tweakable is used in the editor mirrored procedural control surfaces deploy in opposite directions. Normal control surfaces behave correctly. See screenshot:

Logs and minimal craft file provided in archive: Bug-B9PW-Deploy.zip

In comparison to procedural dynamics' wings, these are far less effective wings and cause stalling on the main wings at low angles of attack. This causes jagged turns and an overall less manoeuvrable aircraft. The B9 Pwings are very versatile when it comes to making good looking wings and intricate control surfaces.
However, when it comes to their function, a plane made of predominantly stock parts can well out perform it. This was tested by replacing a previously good plane's main wings with admittedly better looking B9 Pwings. The plane with nothing different but new wings was outperformed by the previous version. TLDR: there is very frequent stalling on these wings in comparison to the other procedural wing option (procedural dynamics). Could there be a mix of function and form?

All flights used ferram aerospace.

Tried to use KSPI and this mod and "Mk2 extended KSPI integration" together and it seems it works incorrectly - "J" key does nothing. After some time spent in debugging I noticed Mk2 KSPI Integration adds radiators to wings (via MM patch). Simpliest way to fix is to change patch, of course, but I did more digging.

After little code changes to debug...
for (int i = 0; i < part.Resources.Count; ++i)
{
PartResource res = part.Resources[i];
Debug.Log ("PartResource = " + res.resourceName);
double fillPct = res.maxAmount > 0 ? res.amount / res.maxAmount : 1.0;
WingTankConfiguration tankSetup = StaticWingGlobals.wingTankConfigurations [fuelSelectedTankSetup];
Debug.Log ("Got tankSetup = " + tankSetup.GUIName);
WingTankResource resource = tankSetup.resources [res.resourceName];
Debug.Log ("Got Resource units/vol = " + resource.unitsPerVolume);
res.maxAmount = aeroStatVolume * resource.unitsPerVolume;
res.amount = res.maxAmount * fillPct;
}

I got this:

[LOG 09:41:58.386] PartResource = WasteHeat
[LOG 09:41:58.386] Got tankSetup = STR
[EXC 09:41:58.386] KeyNotFoundException: The given key was not present in the dictionary.
System.Collections.Generic.Dictionary`2[System.String,WingProcedural.WingTankResource].get_Item (System.String key)
WingProcedural.WingProcedural.FuelVolumeChanged ()
WingProcedural.WingProcedural.UpdateGeometry (Boolean updateAerodynamics)
WingProcedural.WingProcedural.RefreshGeometry ()
WingProcedural.WingProcedural.Setup ()
WingProcedural.WingProcedural.Start ()

Obviously WasteHeat can't be in STR TankSetup, so it throws exception. For sure, I reported this issue to (very old) Mk2 KSPI Integration thread, but, just in case, there can be a workaround against such case in B9 PWings too.

Nothing appears in the parts menu. The only thing I did "wrong" was rename the folder B9_Aerospace to B9_Aerospace_Wings to prevent folders clashing.

MFT disables "J" menu for B9-PW Procedural Wing part.
B9-PW Control Surfaces are working OK

Could you add a patch cfg that will disable MFT for your PWings?

When changing the width and length of the control surfaces the center of lift is changed asymmetrically.

I have the most recent version of Procedural Wings installed, but whenever I place any of the three parts or press J on them, the tweaking screen doesn't appear. What am I doing wrong?

I have it installed in the right place and I have the most recent version of Module Manager installed, as well, so that isn't the issue.

It functions for now, but needs to be moved to an asset bundle at some point

while i tried to pitch, the wings on 2 sides acting reversed, which caused serious roll without sas help
(testing under FAR)

When wings are swept back the Center of Lift will appear to move in the editor, but the wings will BEHAVE as if their Center of Lift is where they would be for straight wings of the same size.

TO TEST: Create an airraft that has neutral stability or slight (but manageable) instability due to Center of Lift being just barely ahead of Center of Mass. Give it very high wingspan and a somewhat thick chord length in its main wing to amplify the effects of any changes you make to wing sweep on craft stability, as well as high wing thickness and fueled (wet) wings to increase its mass. Then sweep the wing as far back as you can.

The Center of Lift will APPEAR to move back and make the plane more stable, but the display is lying to you. Test flying the aircraft before and after introducing wing sweep (save and exit the editor then reliad first- it may be that what is actually occuring is that Lift offset is lost during reloading). Instead of the aircraft becoming MORE stable, it will become FAR LESS stable- as the wing's Lift will continue to act as if it were not swept, but the Center of Mass will still move backwards.

EXPECTED BEHAVIOR: Introducing wing sweep should offset the Lift of the wing in the direction of wing sweep, not only in the VAB/SPH (which does occur/display correctly), but also in actual FLIGHT. This currently does not occur, at least on some computers.

Excerpt from the log file:

[LOG 18:37:48.932] PartLoader: Compiling Part 'B9_Aerospace_ProceduralWings/Parts/Aero_Wing_Procedural/wing_procedural_typeA/B9_Aero_Wing_Procedural_TypeA'
[LOG 18:37:48.949] EffectList: Created 16 effect types
[ERR 18:37:48.963] Module WingProcedural threw during OnLoad: System.NullReferenceException: Object reference not set to an instance of an object
at WingProcedural.WingProcedural.OnLoad (.ConfigNode node) [0x00000] in :0
at PartModule.Load (.ConfigNode node) [0x00000] in :0

[LOG 18:37:48.977] PartLoader: Compiling Part 'B9_Aerospace_ProceduralWings/Parts/Aero_Wing_Procedural/wing_procedural_typeB/B9_Aero_Wing_Procedural_TypeB'
[ERR 18:37:48.981] Module WingProcedural threw during OnLoad: System.NullReferenceException: Object reference not set to an instance of an object
at WingProcedural.WingProcedural.OnLoad (.ConfigNode node) [0x00000] in :0
at PartModule.Load (.ConfigNode node) [0x00000] in :0

[LOG 18:37:48.985] PartLoader: Compiling Part 'B9_Aerospace_ProceduralWings/Parts/Aero_Wing_Procedural/wing_procedural_typeC/B9_Aero_Wing_Procedural_TypeC'
[ERR 18:37:48.987] Module WingProcedural threw during OnLoad: System.NullReferenceException: Object reference not set to an instance of an object
at WingProcedural.WingProcedural.OnLoad (.ConfigNode node) [0x00000] in :0
at PartModule.Load (.ConfigNode node) [0x00000] in :0

Reverting to version 0.40.8 fix the problem.

Hi @Crzyrndm First of all I love this mod and I use it as dependency for other parts mods that I have (BDArmory - Modular Missiles) .

But I still thinking that for very small wings (mainly missile tail fins and canards) the drag values are too high.

If you have compare the smallest stock fin (basic fin) vs a PWing equivalent you will notice that the amount of drag produced by the PWing is much higher than the stock one. (Using FAR at high speed Mach 3.5)

Pitch and roll inputs are getting switched. presumably the control surfaces are using the wrong axis somewhere. Possibly in part setup not code

Can you push an update for ksp 1.3.1 for this so it's indexed on ckan, looking to release RO for 1.3.1 (then 1.4.x sometime after) and for 1.3.1 I need this to be updated as a dependency on ckan.

If I install only this mod - everything is fine. If I also add FAR, the mass doubles (in simplest case), but, depending on the setting of strenght/mass and if vessel has many Pwing parts, in some cases, mass may even become negative (stock wings behaves normally). This problem occurs only in this fork, in the original B9 Aerospace mass seems to be behaving normally with FAR.

Reproduction on fresh KSP 1.0.5:
1)Install FAR and this mod (I used CKAN)
2)Go to SPH
3)Create new vessel consisting of one light part(I use FL-T100 Fuel Tank)
4)Attach B9 PW Procedural Wing to tank
5)Right click on wing and notice curWingMass, it has to be approximately 0.75
6)See in engineer`s report window that vessel mass is 1.5 tons

Wrong mass calculations affect CoM position, but I dont know, how it affect vessel in flight...

[EXC 03:33:02.611] ArgumentOutOfRangeException: Argument is out of range.
Parameter name: index
System.Collections.Generic.List`1[UnityEngine.MeshFilter].get_Item (Int32 index)
WingProcedural.WingProcedural.UpdateMaterials ()
WingProcedural.WingProcedural.RefreshGeometry ()
WingProcedural.WingProcedural.UpdateCounterparts ()
WingProcedural.WingProcedural.Update ()

There is two similar messages in log. Vessel has 3 controllable surfaces(Pwing controls surface, NOT all-moving) - one independent and 2 mirrored, and 3 fixed surfaces, 1 independent and 2 not.

When editing a wing in symmetry mode, its counterpart on the opposite side retains its original shape and the wing much be unattached/reattached after editing. (Ignore if this is a known issue. I just today discovered the mod and this patch.)