How to *properly* reduce Drag?

[Adenosine Triphospate]

So I've been flying a few spaceplane designs, some which have worked and others that only make it to the end of the runway. As my designs get better, there is one thing I suspect that I have not improved at, and that's minimising drag.

What I understand is:

Nosecones reduce drag, even on faces not on the airflow, due to how the game handles surfaces, therefore having open nodes is a big no no. Don't use too many wings, angle wings slightly so the plane faces prograde in level flight. Offsetting radial parts into the body doesn't actually help, "physicsless" parts actually add drag, but to the body its attached to.

What I would like to know is other, albeit smaller things, such as struts, should I minimize their use? Does a slanted nosecone actually perform better radially? What type of nosecone is best in what circumstance? Do slanted wings perform better (regarding drag)? Any tips in general?

[Rocket In My Pocket]

4 hours ago, Adenosine Triphospate said:

What I would like to know is other, albeit smaller things, such as struts, should I minimize their use? Does a slanted nosecone actually perform better radially? What type of nosecone is best in what circumstance? Do slanted wings perform better (regarding drag)? Any tips in general?

Everything you stated is correct, so I'll focus on your questions.

Struts: Very draggy. Use autostruts instead. (Enable advanced tweakables in main game options menu.)

Slanted nosecone: Nope, game doesn't really care where you place it.

Best nosecone: I think there is a thread somewhere around here that compared them all, suffice to say the one with the best "stats" is the best one. (I think shock cone intake came out as least draggy?)

Slanted wings: No, KSP doesn't simulate airflow, just total lift/drag numbers and angle of attack. You can go Mach 5 in a huge bi plane if you want, it'll work just as well as a sleek SR-71 copy cat.

General tips: Basically you got most of it in your post. Open nodes are a big no-no. Less is more generally as far as wings go.

Edited May 5, 2019 by Rocket In My Pocket

[Laie]

Do tests. You know, for science?

This vessel was launched with and without struts. Fuel in the tank was adjusted to maintain similar mass in either case. Launch, wait until the SRB runs out, then check the F3 menu for highest airspeed achieved.

Result: 636m/s with struts, 638m/s without. However, the "without" vessel also was slightly more lightweight, nearly 1% of dry mass.

More experiments would be necessary to determine if the slight speed gain was only due to the difference in mass. But I think I can already say with high confidence that struts are not "very draggy". (paging @Rocket In My Pocket)

Really, I encourage you to do some tests yourself and not just believe everything that is said on these forums. Not everything is as important as it's made out to be. In the case of struts: regardless of drag, they inflate part count and weigh 50kg apiece. So, don't use them if autostruts will do. However, you don't have to avoid them like the plague.

[Rocket In My Pocket]

25 minutes ago, Laie said:

But I think I can already say with high confidence that struts are not "very draggy". (paging @Rocket In My Pocket)

It's possible that something has changed, not that I ever saw a patch note about it, but struts have been proven to be very draggy in the past.

 

Edited May 5, 2019 by Rocket In My Pocket

[Geonovast]

3 hours ago, Rocket In My Pocket said:

Open nodes are a big no-no

I don't believe that's true.

I just did a really quick test.  Simple rocket, pointy nose cone, probe core, couple fuel tanks, small fins and an aerospike.  Launched it straight up with SAS on stability and no touching the controls.

First test was bone stock.  Second had Superfluous Nodes installed, which adds a small node to the tip of the nose cone.

Drag numbers appeared the same during the launches.  The Highest Speed Achieved and Highest Altitude Achieved were very close, with the second launch (the one with the open node at the tip) actually getting about 12 meters higher and 1 m/s faster.

 

I've seen the open node problem mentioned both ways.  All I can really draw from that is that is used to be a problem, and is no longer.  What the issue is really is flat, blunt surfaces, which by the design of the game happens to almost always have an open node - when the node itself isn't the problem.

Edited May 5, 2019 by Geonovast

[Rocket In My Pocket]

Hmm...just tried putting about a dozen struts on one wing of a very simple plane but not the other, the thing can barely fly now, as the left wing is pretty much always in a state of stall.

Seems like they still produce a fairly significant amount of drag?

Open nodes, I'd agree is something of a "rule of thumb." It's not exactly what's going on behind the scenes, but it's easy to understand for most players. Although here's a post from @bewing about a bug with open nodes in cargo bays where putting nose cones on them reduces drag? Seems to imply that open nodes (or at least some open nodes?) produce drag under normal circumstances? Unless that is/was a bug as well?

@Geonovast Although I do agree, it must be something deeper than just "open nodes" as I use your extra nodes mod as well, and haven't noticed a sudden performance drop in my craft or anything. (Which actually just reminded me to update it, I was still using the early "test" version lol.)

Edited May 5, 2019 by Rocket In My Pocket

[Aeroboi]

While struts add drag It seems the cross angle matters so placing it as much vertical and as short as possible should help (which is best anyway)
I also heard but I am not sure that the parent attachment part of the strut connector influences the drag. If that part is a nosecone or a wing piece then drag would have been less.

I think this is proper, lage cillindrical tubes create drag vortexes and increase as you fly faster. For that matter, using no struts and attaching mk3 fuel tanks together and they'll hang as if they were loose. Using autostruts when necessary it's like bolting the fuel tanks on, and only the strut connector is used when necessary.

Also, the shock cone intake seems to do best for minimizing drag at super/hypersonic speed. Use it coupled with the 2.5 to 1.25m adapter and you will have the most aerodynamic shape. If you have Making history for 1.875m the FL-C1000 fuel tank seems to be the best shape.

Furthermore, what people always forget to mention is the amounts of stacks used. If you have a 1.25m fuselage but you require 2 engines it's better to use a adapter on the back then to create a second fuel tank.
If you have a space plane with 3 fuel tanks (1 center and 2 parallel) and you require more then 3 engines you'll also use adapters.

Also, drag also depends on the rear tapered ends. If it's only engines on the rear attachment nodes it is influenced by the engines, some give more drag. When using the R.A.P.I.E.R.S. you can use the reversed nosecone trick to minimize drag further.

Furthermore, while wing incidence helps you want to calibrate it so that the prograde reticle is at the horizon when you are at the crucial 400m/s mark. While some just take a simple ballpark of 3 degrees while 4 or 5 might be better. That means you'll break past 400m/s easier and also use less fuel.

[Adenosine Triphospate]

3 hours ago, Rocket In My Pocket said:

Struts: Very draggy. Use autostruts instead. (Enable advanced tweakables in main game options menu.)

Funny, I didn't know that autostrut was a stock thing before this post! I had always assumed it was some kind of mod when it was brought up, never knowing that you could just enable it like that.

[Adenosine Triphospate]

14 minutes ago, Aeroboi said:

Using autostruts when necessary it's like bolting the fuel tanks on, and only the strut connector is used when necessary.

Thanks for your reply! Do you mean that the strut connector performs differently to autostrut, in terms of the structural integrity it offers? I know that you can select different parts to strut to in auto strut, so would this have the same effect as a strut connector to the said part?

[Rocket In My Pocket]

1 hour ago, Adenosine Triphospate said:

Thanks for your reply! Do you mean that the strut connector performs differently to autostrut, in terms of the structural integrity it offers? I know that you can select different parts to strut to in auto strut, so would this have the same effect as a strut connector to the said part?

Well with auto strut you pick between "modes", which work for most cases.

Heaviest: Self explanatory, the heaviest current part. (Use with caution, as the heaviest part can change.)

Root: The root part of the ship. (Typically your command pod.)

Grandparent: The part that the part it's attached to is attached to. (A bit confusing I know.) 

The real struts let you manually connect thing A to thing B exactly how you want, so they still have their uses.

Edited May 5, 2019 by Rocket In My Pocket

[Aeroboi]

1 hour ago, Adenosine Triphospate said:

Thanks for your reply! Do you mean that the strut connector performs differently to autostrut, in terms of the structural integrity it offers? I know that you can select different parts to strut to in auto strut, so would this have the same effect as a strut connector to the said part?

What I mean is that a strut part isn't a dedicated method by which side boosters are supported in real life. There aren't strut lines hanging of the top of a rocket to hold up the boosters. Instead mechanisms or bolts are used to seperate a booster.
Attaching fuel tanks directly to a center tank should simulate whole structural integrity because a rocket with 2 boosters should be integral without lines helding it together. Therefore autostrut is the proper method to discard ever needing the strut connector.

Only when it's necessary use the strut connector, at least that's my opinion. Some people may find them look cool.

[bewing]

5 hours ago, Rocket In My Pocket said:

Although here's a post from @bewing about a bug with open nodes in cargo bays where putting nose cones on them reduces drag?

Yeah, that bug got fixed several versions ago. I tried to reproduce it recently and couldnt.

[OHara]

The drag from struts was removed in version 1.2 (so test-reports from before October 2016 no longer apply). 
The release-notes say "Fix drag from struts and fuel lines" but the change was to add a line "DRAG_CUBE none=True" in the configuration file, and I have never been able to measure any drag from struts since then.  I think the developers had what they considered 'correct' drag from struts, but users complained, so they turned it off.  If I take that line out and restart KSP, then I see drag from each strut that is more than a linearRCS, about half that of a 4-way RCS block.

The trouble with struts is their 50kg mass (per strut, regardless of length) as Laie says above.

Open nodes don't directly give drag, but are often a missed opportunity to reduce drag.
KSP's drag model treats each part individually, with each feeling the full airflow, except when faces are node-attached to another face, the mating area is removed from the drag computation.