Which do you think will go higher THIS TIME?

[DarkGravity]

Two rockets, nearly identical. Given enough fuel to reach about 25km. (same amount for both). The lower stage is IN-LINE in one, and OFFSET in the other. Which will go higher?

- - - Updated - - -

I'll wait for a few replies before posting the results of the test.

[Levelord]

It doesn't matter either way. The game engine does not recognize parts behind other parts for drag occlusion. It only applies for parts obstructing engine thrust.

[DarkGravity]

Not sure I follow. What I'm hearing seems self-contradictory. Are you saying that in the Inline model, the lower stage <edit> IS ignored, but in the offset model, (since it isn't "behind other parts") it will add to drag? If so, then it matters, yes? Maybe I misunderstood you.

[r4pt0r]

It doesn't matter either way. The game engine does not recognize parts behind other parts for drag occlusion. It only applies for parts obstructing engine thrust.

wait what? I thought the new aero was smarter than that?

[DarkGravity]

THE OFFICIAL RESULT: Inline was better by 4% which was within the error of this little test. (n=2) This could be interpreted to mean that if given a choice, the offset offers no advantage. I will point out however that the offset version is slightly shorter than the inline version.

[Cebi]

It doesn't matter either way. The game engine does not recognize parts behind other parts for drag occlusion. It only applies for parts obstructing engine thrust.

I think that it is not true. http://forum.kerbalspaceprogram.com/threads/119108-Overhauls-for-1-0

Edit:Sorry i did not understand your statement at first. My bad. Game indeed does not recognize this if parts are simply behind something with space between.

Edited May 23, 2015 by Cebi

[Levelord]

wait what? I thought the new aero was smarter than that?

The game only recognizes nosecones when they are attached to attachment points. If they are not placed on attachment points, they are considered radially mounted and will induce drag.

I've made a thread about it here: http://forum.kerbalspaceprogram.com/threads/122574-Radial-mounted-parts-on-atmospheric-flight-performance

[Xannari Ferrows]

My line of thinking was that the inline built rocket would fly farther, as you do not have any grade-level drag slowing you down. You still have angular drag, but it is the same for both builds.

[DarkGravity]

The game only recognizes nosecones when they are attached to attachment points.

I assume you mean the top (maybe bottom) central axis attachment point, yes? Because all my nose cones were attached to attachments points as I understand the concept. Bright small green sphere saying "attach here!" Only one was on the central axis, though.

If they are not placed on attachment points, they are considered radially mounted and will induce drag.

So if the nose cone is placed on the central axis, it doesn't induce drag? I don't think I'm hearing the concepts you're trying to convey.

Edited May 23, 2015 by DarkGravity

[Waterlimon]

All parts have some drag, and the only two things affecting the drag are:

-Parts directly connected to it using nodes (radial attachment doesnt count)

-Full containment (cargo bays, fairings, and other built in containers)

So when determining the drag a part creates, you can ignore all other parts except ones directly connected through nodes, and containers.

(as far as I know)

For the connected parts, it probably just ignores drag for the surfaces covered by attached parts, so when you add a nosecone, what you are doing is trying to completely remove drag from the surface you attach it to (and assume the nosecone itself has less drag, resulting in net win)

I dont know specifics when it comes to partial coverage - maybe the game knows the exact area of all attachment surfaces and calculates drag from the noncovered area, or maybe it just applies full drag, or maybe it has some nonsensical system that has nothing to do reality and just happens to give somewhat correct results...

Edited May 23, 2015 by Waterlimon

[DarkGravity]

So when determining the drag a part creates, you can ignore all other parts except ones directly connected through nodes, and containers.

So if you connect 2 battery packs to the top of a MK1 command pod, it doesn't matter if you apply them in line or on opposite sides? That doesn't sound right. I always put the second one right below the first one, so it can let the first one block the air.

[LostOblivion]

Doesn't matter. As far as I know, surface attached parts induce the same amount of drag no matter where they are placed, even if they are clipped inside the ship...

[Michaelo90]

Can you try removing the bottom nosecones from the top radials and the top nosecones from the bottom radials, put decouplers on the top radials and attach the bottom radials to that? I think that would reduce the drag.

[Levelord]

I assume you mean the top (maybe bottom) central axis attachment point, yes? Because all my nose cones were attached to attachments points as I understand the concept. Bright small green sphere saying "attach here!" Only one was on the central axis, though.

So if the nose cone is placed on the central axis, it doesn't induce drag? I don't think I'm hearing the concepts you're trying to convey.

The nosecones reduce drag when they are attached to an attachment node of a stack.

When you start adding rockets to the sides, they have their own drag calculations because the game considers it another vessel adjacent to the original vessel.

Lets for example say that the central stack, the middle rocket has a drag value of 1 and has the drag reduced by 0.5 with a nose cone. That alone now has a drag of 0.5.

When you add 4 boosters to the side, all with their own nose cones, the drag for each of them are also added to the whole craft. The central stack (0.5) plus the other 4 stacks that are considered separate entities (0.5 x 4) which is = 2.5 units of drag as a whole. The game applies the 2.5 drag on the ship. It doesn't matter what orientation or rotation you apply to the boosters to make them look like they are behind something, they will still induce that amount of drag.

Now, if you were to instead combine all the boosters under the central stack using a 4-way adapter, the drag now becomes a total of 0.5 because they are part of the central stack joined by attachment nodes. The rocket should fly faster and higher.

I know this is completely weird and unintuitive, but that's how the game engine currently sees drag.

[Jouni]

Now, if you were to instead combine all the boosters under the central stack using a 4-way adapter, the drag now becomes a total of 0.5 because they are part of the central stack joined by attachment nodes. The rocket should fly faster and higher.

Occlusion checking also takes the surface area into account.

I built a rocket with four stacks of a 1.25 m probe core, an FL-T800 fuel tank, and an LV-T45 engine. Then I launched vertically with SAS on, and measured the apoapsis after escaping the atmosphere.

There was no significant difference between four pointy 1.25 m nosecones and a single pointy 2.5 m nosecone. The rocket with a standard 2.5 m nosecone climbed to a lower altitude, while the rocket with an 1.25 m nosecone covering the top node of a quad-coupler faced the most drag.

[WinkAllKerb'']

2, 3, and node size then ? (and some kind of default coef applied somewhere in the code related to node size ? this or may be the rescale factor, scale thing in 3d tools related standardization when creating a part somehow messed up at some point in the calculations ? (almost everyone use value pretty close 1 to 1,25 most often but a few parts here and there got some very different rescale value, not sure anyway might worth a look)

Edited May 23, 2015 by WinkAllKerb''

[DarkGravity]

Pretty sizeable differences in final AP. Maybe even significant. Interesting data, to be sure. If you do any more tests, maybe get rid of half of the fuel or more, so the test spends a larger % of its time in atmosphere.

As for my rocket, I have nosecones on the underside of the upper stage stacks based on data from someone else's very recent thread. He found big changes in drag when the trailing surfaces where smoothed.

- - - Updated - - -

Care to guess which (if either) of these two did better?

Answer: no difference at all. Less than 0.1% anyway. However, if I stood the batteries on end (to maximize air braking), that did decrease max altitude.

Edited May 23, 2015 by DarkGravity

[sj176]

Interesting little experiments. Good to test the aero, but probably won't make a difference most of the time.

[DarkGravity]

Probably not for those who know the game well already. But stuff like this helps me to get to know the game more than you might guess.

[SAI Peregrinus]

Testing in FAR, they both flip out and explode. The one with the line of batteries flips out a bit sooner for me, but I probably didn't place them perfectly in a line, thus causing asymmetric drag.

[DarkGravity]

If you want to test them in FAR, maybe try putting better fins and turning down the thrust.

[SAI Peregrinus]

Oh, of course. That was at 25% thrust. Trying that at 100% thrust would probably make them explode on the pad.

One of the reasons to test in FAR is because it models strut connectors nicely so they behave as expected.