A few aero tests

[drewscriver]

This may have been discussed before, but I wanted to see what sort of penalty I might be taking for one of my crazy designs, so I did a quick test to see how KSP treats some build choices aerodynamically.

I know there was a discussion of how KSP regards open nodes, but it goes further than that. Apologies if I'm repeating something that has been covered already.

All fuel tanks were empty in my tests - the only power source was the booster. Pictures were taken at pre-launch (to show the test scenario), during ascent (as the results become apparent), and at booster burnout (to show the final result).

Test documentation gallery

Mark 1 / Mark 2 pieces. Question: is node size the predominant factor, or do different piece sizes matter?

Result: piece sizes definitely matter. The raw Mark 1 / Mark 2 interface seems to cause a lot of drag.

Mark 1 / Mark 2 adapters. If piece sizes matter, what effect do adapters have?

Result: adapters reduce drag, orientation matters

Offsets - what if pieces are offset into each other?

Result: offset does not matter

Offsets, continued - what if pieces are separated?

Result: offset does not matter

Bad news for my crazy plans, but hopefully the tests help someone.

Edited July 24, 2015 by drewscriver
fixed pics. arrgh.

[Foxster]

Piccy no worky.

[Red Iron Crown]

As I understand it, each node has a "bulkhead profile" parameter, which defines whether a part is 0.625m, 1.25m, 2.5m, 3.75m, Mk2 or Mk3. Mismatched profiles in attached nodes will have more drag than matched profiles; this is how KSP models smooth profiles being better than ones with steps in them. It completely ignores offsetting though, so you can make some silly looking things that are still aerodynamic according to the model.

Rotation, however, is not ignored. You can exploit this by mounting an appropriately-sized nose cone to the front of a stack and then rotating it 180 degrees so it clips into the stack, this will actually result in less drag than having the pointy end exposed. I don't use this method for "serious" craft though, as I feel it is exploiting the simplifications of the model.

And yes, your inlined pics need fixing.

[Starhawk]

Very strange. The pics work fine for me in Firefox, but not in Chrome.

Happy landings!

[NathanKell]

Bulkhead profiles are only used for filtering in the VAB/SPH, they are not used for drag calcs. Drag calcs only care if something is attached by a node or not; beyond that the drag cube of each part is queried to see how much area is at that joint.

[Red Iron Crown]

Bulkhead profiles are only used for filtering in the VAB/SPH, they are not used for drag calcs. Drag calcs only care if something is attached by a node or not; beyond that the drag cube of each part is queried to see how much area is at that joint.

Thank you for that clarification and correction. So the node connections compare area but not shape?

[drewscriver]

Sorry about that, pics are fixed now (as far as I can tell).

Thanks for that explanation, NathanKell, it gives me the idea for another test.

[Foxster]

Sorry about that, pics are fixed now (as far as I can tell).

Nope, still not showing. Where are you linking them from?

[selfish_meme]

I saw the pics earlier, maybe the host is up and down

[Dispatcher]

No pics; Safari

[drewscriver]

Dropbox. The photos section doesn't seem to be behaving nicely when I try to direct link to the images in the gallery. I gave up and used public links.

[Foxster]

Thanks for fixing the pictures!

Very interesting stuff. I suppose the offset thing is only to be expected, there being a limit to just how much can be calculated/modelled without burning up the CPU.

I tried something a bit like this but simpler a while ago. I found that even turning a part through 180° so it was blunt side into the wind didn't change its drag. All that mattered was the original placement and orientation.

[Pecan]

A note about the pictures - each one of them is over 1MB, more than enough for screen display; I refer you to http://forum.kerbalspaceprogram.com/threads/79658-Exploring-The-System-A-design-tutorial-campaign-0-90-Final where the pictures are generally only 10 - 20KB.

It is good netiquette to reformat graphics and other large files so they don't have such long download times and cost so much for those of us who have to pay for broadband usage. Especially useful if you're posting several in one thread.