Drag, cone, and... intake?

[Jestersage 293]

In order to optimize my SSTO, I am reading up on the KSP drag-cone research put forth by fourfa, aerogav, and others:

https://forum.kerbalspaceprogram.com/index.php?/topic/183551-rapier-engine-drag-in-161-with-tailcone-attachments/
https://kerbalx.com/fourfa/Rapier-tailcone-test
https://kerbalx.com/fourfa/Rapier-tailcone-test-2
https://forum.kerbalspaceprogram.com/index.php?/topic/155779-best-hyper-sonic-nose-cone/
https://www.reddit.com/r/KerbalAcademy/comments/5eh8yh/the_least_draggy_nose_cone_and_ideal_shell_shapes/

The thing is, many of these info are 2~3 years old, so many versions ago.

So my question is:

1. Are there advantage of using advanced nose cone over aerodynamic cone, as tests seems inconclusive about the drag reduction (especially based on those done by fourfa)
2. No one have done an NCS+small nose cone. How is the drag-stat for combination of such cone?
3. When is the shock-intake small drag outweight (no pun intended) its high mass for SSTO?

Edited December 2, 2019 by Jestersage

[Signo 295]

"Opinions" follow:

1) less dead weight with ncs + 0.625 cone + precooler

2) that Is my style choice, works pretty well.

3) It is subtle: when you feel you need a little bit more air intake when flying away from prograde the shock cone will do. The precoolers are good for a less acrobatic flight.

*) End cones on rapiers or any rocket engine you may have are really important. 

**) Control surfaces are draggy. Canards are very draggy.

 

[IronMaiden 8]

The "tail" cone is only useful if you can't punch through the sound barrier. Once you get going mach 1.3 or so, that rear facing surface causes less and less drag the faster you go, and your engine makes more and more thrust. Depending on your angle of attack they may even make things worse. Save the mass unless you absolutely need it to reach mach 1.3, however in that case I'd recommend other ways to reduce drag that don't involve adding dead mass to your spaceplane.

 

1. No, the aerodynamic cone is superior to the advanced cone in every conceivable way, especially the further you are from 0° AoA. Unless you need fuel crossfeed through your nose cone for some reason.

2. Difficult to say since you're adding fuel it's not just dead mass unless your plane already has a mass ratio over 5. So it would depend on a lot of different factors. I will say that the longer your tail cone, the more sensitive it is to angle of attack. The more you expose those x and z surfaces to the airflow, the more drag you get. This is the main reason the aerodynamic cone is so much better than the "advanced" one and why tail cones aren't particularly effective in the first place. I love the NCS + small nose cone for its aesthetics though. They're typically the front of my mk1 spaceplane designs. It offers good aerodynamics and good heat resistance in addition to the extra fuel.

3. Depends how exploitative you want to be  Intakes don't care if they're backwards as far as air intake goes, might as well put your intakes on the back of your RAPIERs. Since shock cone intakes are the best in the game by far, it's always worth it to have at least one. Keep in mind that you only need to have one shock cone per ~10 RAPIERs, so you could attach them all inline. But if you're going to do that you might as well use an open ended fairing.

 

Ultimately it's up to you how far you want to go in exploiting the aerodynamic model. Using tail cones is one of the more pointless and potentially negative ones though.

[Jestersage 293]

42 minutes ago, IronMaiden said:

The "tail" cone is only useful if you can't punch through the sound barrier. Once you get going mach 1.3 or so, that rear facing surface causes less and less drag the faster you go, and your engine makes more and more thrust. Depending on your angle of attack they may even make things worse. Save the mass unless you absolutely need it to reach mach 1.3, however in that case I'd recommend other ways to reduce drag that don't involve adding dead mass to your spaceplane.

 

1. No, the aerodynamic cone is superior to the advanced cone in every conceivable way, especially the further you are from 0° AoA. Unless you need fuel crossfeed through your nose cone for some reason.

2. Difficult to say since you're adding fuel it's not just dead mass unless your plane already has a mass ratio over 5. So it would depend on a lot of different factors. I will say that the longer your tail cone, the more sensitive it is to angle of attack. The more you expose those x and z surfaces to the airflow, the more drag you get. This is the main reason the aerodynamic cone is so much better than the "advanced" one and why tail cones aren't particularly effective in the first place. I love the NCS + small nose cone for its aesthetics though. They're typically the front of my mk1 spaceplane designs. It offers good aerodynamics and good heat resistance in addition to the extra fuel.

3. Depends how exploitative you want to be  Intakes don't care if they're backwards as far as air intake goes, might as well put your intakes on the back of your RAPIERs. Since shock cone intakes are the best in the game by far, it's always worth it to have at least one. Keep in mind that you only need to have one shock cone per ~10 RAPIERs, so you could attach them all inline. But if you're going to do that you might as well use an open ended fairing.

 

Ultimately it's up to you how far you want to go in exploiting the aerodynamic model. Using tail cones is one of the more pointless and potentially negative ones though.

Not the actual tail cone (which is long and big), but the short version (that's the advanced nosed cone).

As you can guess, I am still trying to optimize my LF-only SSTO.

The main reason is that I really dislike using fairing as permanent part, but I would like a forward facing nosecone that can stand up to heat. At mach 5, the nosecone is at 1800 deg

Edited December 4, 2019 by Jestersage

[IronMaiden 8]

2 minutes ago, Jestersage said:

Not the actual tail cone (which is long and big), but the short version (that's the advanced nosed cone).

As you can guess, I am still trying to optimize my LF-only SSTO.

The main reason is that I really dislike using fairing as permanent part, but I would like something that can stand up to heat.

Right, I just meant something you put backwards as the final piece of your fuselage/engine pod, a cone on the "tail." They're usually not worth it.

I assume this is your "Ascensor" craft? What are you looking to improve? More payload? More Δv? Quicker ascent? A nose cone that doesn't blow up? Like I said, in the current aerodynamic model, the "aerodynamic cone" is superior to the "advanced cone" in every way, including temp tolerance, and IMO would even look much better on this craft

[Signo 295]

I tried empirically with the old "many of the same" method when refining my own LF only SSTOs and I simply noticed that the overall drag coefficent during flight and the d/v I had left in LKO was better with the "opinions" I stated above, even if by a small margin. Roll of self explicative pics follows.

Edited December 4, 2019 by Signo

[IronMaiden 8]

Just to show you're looking in the wrong direction if you want to make substantial improvements to your design...

I downloaded your craft "Ascensor" and made a few changes:

• Removed Whiplash and two Nervs.
• Removed canted Tail Fins, replaced with single static Delta Wing.
• Removed backward facing aerodynamic cones and quad adapter.
• Removed "advanced" nose cone from the front, replaced with aerodynamic nose cone.
• Added small ailerons to tips of wings.
• Adjusted spring strength and damper settings on landing gear to reducing bouncing on runway, moved rear gear forward to accommodate my CoM displacement.
• Moved RAPIER pods forward to bring CoM further forward and ensure it doesn't move between placing and removing orange tank in payload bay.
• Increased payload to 40.5t.
• Finally, and most importantly, adjusted wing angle of incidence to move CoL much further forward when AoA is close to 0° and provide substantial lift at that attack. This allows for a very shallow climb with minimal control surface deflection which vastly improves performance.

Craft File

The result is 4.5t more payload to orbit with nearly 8t fewer engines, ending up with over 50% more Δv.

Flight Instructions:

• Set pitch trim to 1 tick below 1/2 up. Nose will become airborne around 70m/s, pull up around 100m/s to takeoff.
• Slowly trim down to keep climb rate below 1m/s until angle of attack becomes negative, should happen around mach 1.3.
• Let it nose up to around 15°, continue to trim down to keep it from going over 20°.
• As you pass 10km trim down more aggressively to bring it to level flight between 15-18km.
• Lock in surface prograde when climb rate drops below 20m/s, do no further pitch adjustments until positioning for final circularization burn.
• Let it accelerate and climb to ~1600m/s at 20km.
• Active Nervs during the final phugoid before 1600m/s when climb rate drops below 50m/s
• Just keep it on prograde hold til you reach desired apoapsis... and pray your cockpit doesn't explode
• Should just about circularize around 30km before pushing apo above atmo.

[Signo 295]

Flying my "experiments" I noticed that drag is just a "state of mind" if you fly "lightly"

You can really find it only if you go deep in your "inner self".

 

Edited December 4, 2019 by Signo

[AeroGav 1,042]

Intakes don't work properly if facing backwards in fact - as you get faster,  they normally bring in more air,  but facing the wrong way,  they intake less which is bad news as that's just when engine demand starts to go up.

I do normally still put cones on the back of all my engines because busting mach 1 is the hardest part,  the less jet engines you need for that, the more delta v or payload fraction you can bring as those things are heavy. 

However,  the difference in drag between aerodynamic nose cone vs shock cone are very small compared with the drag the mk3 fuselage parts are creating.

This is the area you should focus on in my opinion.    Ways to get the fuselage drag down

1. add incidence to the lifting surfaces,  as IronMaiden has done, so the craft will fly on prograde.  5 degrees is optimal angle of attack for lifting surfaces at supersonic/hypersonic speed - which coincidentally is what happens when you hold SHIFT-S during part attachment.

2. try not to use any fuselage fuel tanks.   Seriously , liquid fuellers don't need them.    If you swap all your wing parts over for big S ones (including your vertical stabilizer), you probably have enough capacity to get to orbit without them.   Check out my Andromeda freighter,  which can put an Orange tank in orbit.

https://kerbalx.com/AeroGav/Andromeda

In practice,  you will probably still need some kind of fuselage "trim tank" to get the CoM where you want it to be,  or to make sure it doesn't shift around as the fuel burns off.  But ,  it might not need to be a full mk3 one.   Maybe just a mk1 lf tank up front, behind the intake cone?

3. That mk3 "Space Shuttle" engine mounting plate can be very draggy if you don't use it right.    Your craft notes say it was based off my "Lusitania" ship..   hopefully you copied the attachment method that thing used  ?   Remember folks, it has a center 2.5m part attachment node in the middle of the three "vector engine" mounts,  and you get an eye watering flat plate drag penalty if you don't put something on this.    I usually fit a 2.5m tri-adapter on there, for three more mk1 sized engines.   I clip that adapter inside so it looks less like the airplane is suffering a prolapse,  then offset the engines horizontally so it looks like they are attached to the sides of the rear fuselage.

 

I can't download your ship as I don't have "making history" x-pac.    

Two other changes i'd probably make if it were mine - i'd swap half the rapier for panther,  this arrangement gives slightly more thrust for busting mach 1,  weighs 2 tons less,  at the cost of a negligible decrease in top speed in level flight, airbreathing mode.   There's not much in it though.

I'd probably want some other cockpit arrangement, for "quality of life" if nothing else.  That mk3 cockpit gets so hot as it is at the front.   The nose cones etc on its front node don't shield much of it from heating as they are much smaller diameter part.    You could fit some kind of mk1 or mk2 inline cockpit,  via a series of mk3 to whatever size adapters (would do double duty as your trim tank).   Or  put a mk1 inline cockpit at the front of the cargo bay, and offset it upwards so only the bubble canopy pokes out the top.  That way you can still fly it from the in cockpit view,  but the cockpit is still shielded from the drag and heat.      This is the method i used in "Monstrosity 225L"

[Jestersage 293]

3 hours ago, IronMaiden said:

Just to show you're looking in the wrong direction if you want to make substantial improvements to your design...

I downloaded your craft "Ascensor" and made a few changes:

• Removed Whiplash and two Nervs.
• Removed canted Tail Fins, replaced with single static Delta Wing.
• Removed backward facing aerodynamic cones and quad adapter.
• Removed "advanced" nose cone from the front, replaced with aerodynamic nose cone.
• Added small ailerons to tips of wings.
• Adjusted spring strength and damper settings on landing gear to reducing bouncing on runway, moved rear gear forward to accommodate my CoM displacement.
• Moved RAPIER pods forward to bring CoM further forward and ensure it doesn't move between placing and removing orange tank in payload bay.
• Increased payload to 40.5t.
• Finally, and most importantly, adjusted wing angle of incidence to move CoL much further forward when AoA is close to 0° and provide substantial lift at that attack. This allows for a very shallow climb with minimal control surface deflection which vastly improves performance.

Thank you. I will load it in KSP later. (I was experimenting on the wing angle, but did it in a way that actually slowed it down)

While I can understand most of the changes, what I am curious is about the removal of aerodynamic cones and quad adapter (namely the cones), as those are added on the research from AeroGav and fourfa, where they put it at the end to reduce drag. So why is removing okay in this case?

 

 

Quote

That mk3 "Space Shuttle" engine mounting plate can be very draggy if you don't use it right.    Your craft notes say it was based off my "Lusitania" ship..   hopefully you copied the attachment method that thing used  ?   Remember folks, it has a center 2.5m part attachment node in the middle of the three "vector engine" mounts,  and you get an eye watering flat plate drag penalty if you don't put something on this.    I usually fit a 2.5m tri-adapter on there, for three more mk1 sized engines.   I clip that adapter inside so it looks less like the airplane is suffering a prolapse,  then offset the engines horizontally so it looks like they are attached to the sides of the rear fuselage.

Yes. In fact, the beta version, with 4 NERVs, are identical: a quad adaptor onto the center 2.5m for the nervs, then top 1.25 node have the whiplash, the side nodes having the pre-coolers and rapiers.

The Current version (due to nonoptimization, and a preference for 0.4 TRW once in orbit) used 6 NERV while the 4 NERVs and Whiplash are mounted the same: the two side nodes have the 2 NERVs, and then the rapiers are attached to the nodes and offset as needed.

Also, I thought Mk2 have event greater drag then Mk3?

Edited December 4, 2019 by Jestersage

[AeroGav 1,042]

5 minutes ago, Jestersage said:

While I can understand most of the changes, what I am curious is about the removal of aerodynamic cones and quad adapter (namely the cones), as those are added on the research from AeroGav and fourfa, where they put it at the end to reduce drag. So why is removing okay in this case?

probably because of the wing incidence angle allows flying on prograde lock,  so the engine parts and the empty nodes are now at zero angle of attack, reducing their drag, allowing you to get away with not having them.   I'd still keep them personally, it's not like you're going to notice the dry mass reduction with the weight of all those nervs on such a vessel

[Signo 295]

Wings, even if they provide lift as a counterbalance, are among the most draggy parts you can carry. You really need to find your own balance on that.

 

[AeroGav 1,042]

4 hours ago, Jestersage said:

Also, I thought Mk2 have event greater drag then Mk3?

No,  mk2 has about 2 and a half to three times of a mk1 part of the same length.   This is terrible for fuel tanks as they contain no more fuel than a mk1 part of the same length.     For passengers,  they are not so bad as you get twice as many kerbals per module,  but the kerbals to drag ratio is still worse, especially given that you'll now need mk2 to mk1 adapters at front and back of stack,  which contain oxidizer tanks you are not using.

 

[AeroGav 1,042]

1 hour ago, Signo said:

Wings, even if they provide lift as a counterbalance, are among the most draggy parts you can carry. You really need to find your own balance on that.

 

Don't trust the red arrows,  they are not to scale.   Use ALT+f12  > physics > aerodynamics and tick the "show aero data in action menus" checkbox.

I did a quick test -

Those strakes on the back of the main wing have 0.44 drag,  if you used 4 of them to have same capacity as the fuselage tank ,  you'd get 1.76 drag, which is slightly higher than the mk1 tank.    To an extent,  you can compensate for this by busting mach 1 at a higher altitude, since the larger wings enable you to fly at low angle of attack in thinner air,  but beyond a certain point thinner air makes engine thrust fall away faster than drag,  so yes,  you get better airbreathing performance with small wings.

 

However,  when you are on NERV power,  the lift from the larger wings quickly takes you up in to the thinner air above 30km,  where fuselage drag and heat are greatly reduced.     I find myself applying nose down input at 20km for the speedrun but once i've lit the nukes and pitch for optimum lift/drag,   it zooms upstairs , levels off and starts to gain speed quickly.

 

 

[Signo 295]

So a stack of 7 Mk1 tanks has got the same drag of, lemme guess, a pair of wings?

[AeroGav 1,042]

13 minutes ago, Signo said:

So a stack of 7 Mk1 tanks has got the same drag of, lemme guess, a pair of wings?

???   0.44 drag on a big S strake,  lift rating of 1

Last time i checked, all wing parts have same lift to drag ratio,  so a pair of big S wings (lift rating 10 for the pair ) would be 4.40 kn drag at the speed , altitude and angle of attack in that screenshot.        Equivalent to 3.18 mk1 tanks,  not 7.

Given how heavy NERVs are,  I'm inclined to make all my fuel tanks Big S parts to make this part of the ascent easier (get out of the lower atmosphere faster, once you pitch up from the speed run).   This does slow down the air breathing part of the climb,  when you can't just fly higher because your engines need air to breathe.   On the other hand it makes landing speeds so slow you're guaranteed to survive no matter where where you land and enables some of the dumb stunts i've done (mountain lake to minmus challenge on the Puffin,  or takeoff/land from the VAB roof while fully fuelled)

[Signo 295]

If you have a bare Mk1 tank yes. If you have a stack of tanks with all the tricks we mentioned above the stack will carry a lot more of fuel with a negligible increase in drag.

[AeroGav 1,042]

1 minute ago, Signo said:

If you have a bare Mk1 tank yes. If you have a stack of tanks with all the tricks we mentioned above the stack will carry a lot more of fuel with a negligible increase in drag.

The tank i used in my screenshot above was in the middle of a stack, it had the shock cone at the front and a cockpit behind it.     You could reduce the drag of the tank by mounting the wings at 5 degrees incidence angle and flying the fuselage on prograde, but that's not a fair comparison as you wont fly on prograde hold for much of the airbreathing part of the flight,  you'll probably be at some negative fuselage aoa.

More wings = less drag during the nerv part of the flight,   less wings = less drag & faster acceleration on airbreathers

since nervs are so heavy and underpowered,  the main driver of my designs is how it flies in the hypersonic part of the flight.     Getting through mach 1 is the next most important part,  and you can improve this by fitting less wings than i normally do,  but i don't like to do that because it makes the nerv part so much worse.   You can still easily break mach 1 even with a flying wing design.  That said, my designs are criticised for taking a long time to reach 1500 m/s in the speedrun,  with less wings and drag,  and flying lower where there is more intake air and thrust, i imagine your ssto can get to 1500 quickly,  even on a single jet engine.

[Signo 295]

So you are suggesting there is the need to find some kind of "balance"? 

[Jestersage 293]

4 hours ago, AeroGav said:

No,  mk2 has about 2 and a half to three times of a mk1 part of the same length.   This is terrible for fuel tanks as they contain no more fuel than a mk1 part of the same length.     For passengers,  they are not so bad as you get twice as many kerbals per module,  but the kerbals to drag ratio is still worse, especially given that you'll now need mk2 to mk1 adapters at front and back of stack,  which contain oxidizer tanks you are not using.

 

But you are talking about mk1 (1.25m parts) instead of the mk3 parts...

Anyway, I will look at IronMaiden's design. Will probably keep the 6 NERV design, but modify the wings, which i actually have it in pipeline already with multiple possiblity, because it's not a craft for "I can reach the orbit", but "it can do useful work".

So my end goal is to match even closer to the M19, that is:

• Low part count
• at 80~100k, 28 degree, can lift 40t and at least 4 Kerbals, excluding components that may be added.
• at above mission parameter, with possibility of lower payload mass, have about total 500 m/s for transfer to spacestation (~300 m/s on average)
• no oxidizer tank if possible (I actually made a version using LFO-tank and vernor, which gives better in orbit performance)
• Easy to fly (ascent)
• Easy to maneuvur in orbit - hence, 0.4 TRW (originally 0.6 TRW)

Edited December 4, 2019 by Jestersage

[IronMaiden 8]

6 minutes ago, Jestersage said:

Thank you. I will load it in KSP later. (I was experimenting on the wing angle, but did it in a way that actually slowed it down)

While I can understand most of the changes, what I am curious is about the removal of aerodynamic cones and quad adapter (namely the cones), as those are added on the research from AeroGav and fourfa, where they put it at the end to reduce drag. So why is removing okay in this case?

Your plane has no problems breaking the sound barrier without them so they're not really needed. I'm all for reducing parts and mass where possible. Not that the mass reduction will make much difference here, but the drag improvement is equally negligible. The faster you go the less drag is created by rear facing surfaces, so the effect will continue to become more and more negligible as you climb and accelerate. The most critical part of your plane's ascent is between 1600-1900m/s and 20-30km. When you're up that high and going that fast it buys you nothing. And if you're making large angle of attack maneuvers, like you need to without any wing incidence, they can actually cause increased drag due to adding more x and z surface area (pointy is bad unless your point is flying exactly into the wind).