Some Questions about Semi-Physicsful Part Drag

[Right]

This is my starting point.

the drag area (but not drag coefficient) is added to the first physicsful parent part.

Interesting. My first take away would be that, all else constant, having a parent part with the lowest drag coefficient is ideal.

How does the game add these values? Does it just add the volume of the physics-less part to the volume of the parent part? One possible answer:

Not exactly, it adds drag factor, but drag is still based on the cross section of the non-physicsless parent part which is multiplied by the drag factor of the part which had the strut drag factor added to it (or something like that, I don't have the exact formula).

If the parent part is completely drag occluded (inline, 0 AoA), does the it become less occluded due to "growth" when adding limited-physics parts? Along similar lines, if the parent part is over-occluded (1.25m root part with a 2.5m nosecone), will the root part's occlusion remain 100% up to the 2.5m area?

Finally, if the parent part is a wing, my understanding is that wings don't use drag boxes (InternalDragModel = True). So wouldn't limited-physics parts on wings increase their lift area?

Edited October 12, 2015 by Right

[Val]

I tried to answer some of your questions.

My findings are, physicsless parts:

• add the same drag no-matter the stack occlusion status of the parent.
  
• add more drag to wings, than they do stack attached parts. Unexpected!
  

The pictures I based my findings on.

Javascript is disabled. View full album

[Arugela]

Are there still physics-less parts? I thought they made them all physics based now.

[NathanKell]

Physicsless is a misnomer, yeah. Now it's "full physics" and "limited physics". In particular PhysicsSignificance=1 (limited physics) parts don't have rigidbodies so don't matter for wobbliness, and have drag and mass added to the parent, but drag is still calculated and mass still matters.

[Right]

Point taken, hence forth called limit-physics.

Nicely done Val!

I'm not terribly shocked that over-occluding doesn't work. Occlusion is probably calculated by the drag area in the part database file rather than that value combined with semi-physics parts. Sounds like it would be an easy change, but it would inevitably lead to un-aerodynamic looking craft being very aerodynamic.

My own tests are showing that placing limited-physics parts on parts known to have low drag coefficients doesn't make a big difference. Sometimes its even worse. I think we could be mistaken about the parent part's drag coefficient being unchanged.

Interesting bit: if you make a limit-physics part your root part, it gains full physics.