RocketMoron

Space plane design help

Recommended Posts

Hi, novice kerbonaut here.

I've successfully designed and created multiple SSTOs of all classes that are Mun and Minus capable, thanks to many of the topics here and KSP tubers. 

Some important tips are:

1. Low drag is better than high thrust (Thank you GoSlash27 and Warzouz)

2. Control surface placement is key (Thanks to physics)

3. Centre of mass ahead of centre of lift (Just try and see what happens if not)

4. Static incidence in wings, i.e. tilt your wings to include a little angle of attack while level so that thrust in prograde can still generate positive climb without control surface drag losses.(GoSlash27 at God level IMO)

5. Try to ensure dry center of mass remains within 5% deviation from its initial position (preferably at the center for maximum aerodynamic control authority)

Tip number 5 is where I struggle with. I try optimising my vehicles by messing around with fuel placement, but can never seem to keep track centre of mass within those limits. (Sorry, will not be attaching craft files just yet).

Most of the time, this shift is unavoidable until I use radially placed fuel tanks and engines, which comes into direct contradiction of tip 1, and tip 1 always gets more preference for efficiency purposes.

So to conclude, I'm asking for tips about how my designs can incorporate that last shred of increased efficiency through smooth control.

Edited by RocketMoron

Share this post


Link to post
Share on other sites

...but as much as overdoing #5 may compromise #1, not following tip #5 may compromise #3, which is why the only SSTOs I bother with tend to be rockets and not planes.

That said, #5 shouldn't be that hard, you only really need to make sure your fuel tanks are either already centered on the CoM or that you spread the mass around your plane equally. It does get harder when you need a bunch of heavy engines at the back, though...

Quote

Most of the time, this shift is unavoidable until I use radially placed fuel tanks and engines, which comes into direct contradiction of tip 1, and tip 1 always gets more preference for efficiency purposes.

...but there's a limit on how big your plane can get if you don't build radially, otherwise you will get spaghetti just like any rockets... or rapid unscheduled disassembly due to drag and G-Forces tearing your plane apart the moment you touch the controls (or your CoM moves too much due to burnt fuel).

#4 is so simple yet so smart, I might have to try building spaceplanes again... can't believe I didn't think of that.

Quote

So to conclude, I'm asking for tips about how my designs can incorporate that last shred of increased efficiency through smooth control.

That sounds like a unicorn to me... I mean, some YouTubers are really good with spaceplanes, can't remember if it's Mark Thrimm or Marcus House off the top of my head. Anyways, my point is that it is possible, but... my experience has been that the most efficient designs are the smaller ones which have little purpose since they really can't bring much to orbit. The other options usually sacrifice something: either you have to do all kinds of crazy fancy maneuvers to actually get the plane into orbit, or you have to sacrifice some efficiency for thrust and just power your way to orbit. Unless you're a spaceplane wizard like very few...

Share this post


Link to post
Share on other sites

Thanks for the reply. 

9 hours ago, Ohm is Futile said:

...but as much as overdoing #5 may compromise #1, not following tip #5 may compromise #3, which is why the only SSTOs I bother with tend to be rockets and not planes.

Yes, #3 has to be followed. The majority of my spaceplanes are cargo lifters to a station at 200Km LKO. I've poured hours into honing my Mk2 spaceplanes for tiny probe launches and I had to move into Mk3 to avoid the cargo from getting stuck in the cargo bay. I've designed one Mk3 spaceplane that way, and it works, the only issue I face is the constant fight i put up with sas to keep the nose steady and pumping fuel around to maintain control. (Not a huge problem, but I like simplicity). 

And on another note, I haven't yet built my first rocket SSTO. I also believe those would be much simpler than spaceplanes, only that I haven't yet worked out the math yet. I do have one returnable rocket, carries an orange tank and a half to the same station. Only expenses are on 6 Kickbacks and fuel.

9 hours ago, Ohm is Futile said:

That said, #5 shouldn't be that hard, you only really need to make sure your fuel tanks are either already centered on the CoM or that you spread the mass around your plane equally. It does get harder when you need a bunch of heavy engines at the back, though...

Yes, that would be the ideal solution, but cargo bays make that a problem. One solution was using ballast equivalent to the dry mass of each half of a spaceplane, which as you go bigger becomes a huge problem on the engines. I keep my spaceplanes from 0.65 to 0.75 TWR to try and maximize payload fraction. And I tend to use only Rapiers and a Ramjet or two for large vehicles. NERVs are for celestial travel only.

9 hours ago, Ohm is Futile said:

...but there's a limit on how big your plane can get if you don't build radially, otherwise you will get spaghetti just like any rockets... or rapid unscheduled disassembly due to drag and G-Forces tearing your plane apart the moment you touch the controls (or your CoM moves too much due to burnt fuel).

#4 is so simple yet so smart, I might have to try building spaceplanes again... can't believe I didn't think of that.

Thanks for that tip. I haven't built quite that large yet, like Matt Lowne's Argus class spaceplane. I'm starting to prepare my fleet of Duna and back spaceplanes myself. But I expect building radially will become a neccessity soon.

You may thank GoSlash27 for #4, helped me loads. I've actually studied aerodynamics and a fair amount of aeronautical engineering but didn't think to apply it *facepalm*

I happen to draw a lot of inspiration from KSP tubers like Mark Thrimm and Matt Lowne. My whole tryst into spaceplanes was merely due to the long drawn out waiting times between maneuvers of craft waiting for ejection or on interplanetary trajectories. 

Edited by RocketMoron

Share this post


Link to post
Share on other sites

#5 always, always fill your wings with fuel, like flying a bus on takeoff sometimes but makes for a smooth ride overall. 

I also build from the cargo bay out-forwards and back keeping my cargo on the centre of mass and using the pod to balance the engines (normally on a long lever arm).

 

Also because I just love strakes everything about my SSTOs is always at the back- centre of mass and lift and control surfaces and consequently rcs, reaction wheels etc which shouldn't handle well but it does even though my pod is stuck out in front on it's own.

Anyway there's little more you can learn after advice from slash, he is awesome at this stuff.

Share this post


Link to post
Share on other sites

Folks,

 I appreciate the kind words, but please keep in mind that virtually everything I know about spaceplane design in this game has come either directly from others or in conjunction with their hard work. A prime example of this is the "wing incidence" thing; I stole that technique shamelessly from @Wanderfound and I certainly don't deserve any credit for it.

 I merely pass on what I learn just as you're doing here, I don't invent this stuff!

 On the subject of efficiency through finesse...

 This is especially important for underpowered designs that rely on clean aerodynamics rather than brute force. "clean" designs tend to only be clean when the nose is pointed directly prograde, which is why I stress wing incidence to make that happen at the sound barrier which is the draggiest portion of the flight. Any misalignment from prograde or maneuvering creates drag, which wastes fuel. So to that end...

- I use minimal control surfaces and turn the authority down. The feel should be pleasantly heavy but neutral so that control inputs make a nice smooth response while needing minimal active correction. If your plane oscillates in response to an input, the authority is too high.
- I don't use 1 set of control surfaces for more than 1 input. Elevons, ailerators, etc. tend to be sloppy when responding to conflicting demands.
- My rudders do not actively respond to yaw inputs. Having them active induces a crab condition with roll when SAS is engaged, which is draggy.
- My reaction wheels are disabled during climbout. Mainly due to the same yaw cross-coupling effect as the rudders.
- I prefer to have the wings mounted a little above centerline if I can, or some dihedral in the outer panels in order to make the plane want to return to level.
- I prefer to have a conventional tail arrangement instead of canards. Canards tend to be dynamically unstable in pitch, which is good for fighters... but bad for space planes.
- Remember that the center of drag (pressure center)  is a different thing than center of lift or mass and is not indicated in the VAB. If your CoM and CoL are in the back of the plane and you have a lot of fuselage hanging out in front, there's a good chance that it will be very squirrely at high speed and may not be controllable on reentry *especially* if it's a Mk2 design. For best results, your CoM and CoL should be near the longitudinal center of the plane.

 


 

Best,
-Slashy

 

Edited by GoSlash27

Share this post


Link to post
Share on other sites
On 23/12/2016 at 6:32 AM, Ohm is Futile said:

...but as much as overdoing #5 may compromise #1, not following tip #5 may compromise #3, which is why the only SSTOs I bother with tend to be rockets and not planes.

Putting the fuel tanks out on side sponsons doesn't have that much of a drag penalty provided each stack has a low drag nosecone front and back, and any size transitions are handled with adapters to make it smooth.   Mk1 size tanks are pretty low drag anyway.  Mk0 not so good, and mk2 are a disaster.   2.5m are godlike - but you'll only be interested if you're building something really large.

KSP%202016-11-22%2020-26-06-623_zpstd2no

Edit - an example of the stuff we need to do to get CoM where we want it 

20161216174338_1_zpstbdmwv9i.jpg

CoM must be in the centre of the cargo bay so it flies the same unloaded vs loaded.

Had to use a forward swept design to get the nukes far enough forward.  The mass of the cockpit alone certainly ain't enough to balance the ship if i'd put 6 tons of nerv on the tail.

I'm using two side sponsons for fuel so i have freedom to slide them fore and aft as needed.  Also, the mk2 fuel tanks have such awful aerodynamics i'd have more drag if i incorporated them into the main fuselage, even though i'd theoretically have less frontal area - ie. 4 less type B nose cones to push through the air.

With a mk2 design like this, you pretty much have to angle the wings (add incidence) to make it work.  

Please note - this is a career mode design with 500 tech cost parts.   If they were available, you absolutely would use Big S wings and strakes instead of sponsons on a lf-only ship like this. And you'd swap the whiplash for the rapier, no matter how you love it's blue afterburner.

edit  2 - check out my sandbox plane, with wing root mounted nukes.   The need to get dry CoM in the right place does create a bit of an odd look, as if someone sawed off the rear half of the fuselage.

20161223183913_1_zpskrme4ox6.jpg

This one does not have angled wings.  On a mk1, you can get away without doing that because the fuselage drag is so much less, especially when you have very low wing loading like this design which means you'll by flying at small AoA and very high up, by the time you're going fast.

Not having to add incidence  greatly simplifies construction especially if the airplane is also a canard.  

Why?

unstable_zpsje5rumlu.jpg

This is what happens if you add incidence to the main wing but not the canard.   It looks stable in SPH, but as the nose rises the main wing will start to stall before the canard and the aircraft will flip backwards - and stay there. You can see the col moving forward as the nose goes up even a small amount.

uptrim10deg_zpspxlpa1e5.jpg

If you use a slightly greater incidence on the canard than the wing, you can see benign behaviour returns - the canard stalls first, CoL moves aft as you pitch up.  However, in itself this moves CoL way forward so you'll have to slide your wing well to the rear to compensate.   A fairly major redesign.    

 

Edited by AeroGav

Share this post


Link to post
Share on other sites
On 31/12/2016 at 7:01 PM, GoSlash27 said:

 

- I use minimal control surfaces and turn the authority down. The feel should be pleasantly heavy but neutral so that control inputs make a nice smooth response while needing minimal active correction. If your plane oscillates in response to an input, the authority is too high.
 

In a spaceplane,  ailerons only get used to correct any unwanted roll that occurs and to make heading corrections of 2 or 3 degrees max.  I do normally equip fairly large ailerons but turn their authority way down to 30% or so.  At small deflection angles, they make less drag, which is good for speed, and also means you get less yaw as a byproduct of using the aileron.

Regarding pitch, if it's only able to pull 10 degree AoA nose up, that's OK provided deflection angles and drag are low in normal flight.

Quote

 My rudders do not actively respond to yaw inputs. Having them active induces a crab condition with roll when SAS is engaged, which is draggy.

That explains why when I use SAS mode my planes end up crabbing after a while.   Rolling in one direction, trying to turn, and a load of opposite rudder trying to make it not turn.   My solution was different -  stopped using Stability Assist mode.

I either fly with SAS off completely, or I make airplanes with wing incidence built in, that still make enough lift when the fuselage is on Prograde Hold.      If you're flying in Prograde hold, the SAS doesn't seem to attempt to correct roll.  But, it will use the rudders to try eliminate yaw and sideslip, which is a good thing.   If you've got dihedral on the wings,  the plane will keep itself wings level provided there is no yaw.  Otherwise it'll dutch roll into the yaw,  and annoy you.

Quote

My reaction wheels are disabled during climbout. Mainly due to the same yaw cross-coupling effect as the rudders.

My preference is to leave them on, because they don't create any drag out on the wing tip to produce the roll torque, so there is less secondary yaw produced.  Of course, they're too weak to do the job on their own, and also you got to bear in mind rapiers don't have alternators and solar panels can't be deployed in the atmosphere.    Running out of electricity at 40km can be embarrassing.

re: wing sizing.

Another shout out to the mighty CorrectCoL mod.   Not only does this make the SPH blue ball more accurate (accounting for fuselage as well as wing lift) it also lets you simulate various flight conditions and see how much AoA will be needed to fly at X altitude and Y airspeed.

For your speedrun, you typically want to simulate 1500 m/s and 22km.   If you're using angled wings, then you want there to be just the right amount of wing so that you're at 0 aoa here.   Efficiency drops fast if you're having to angle the fuselage to a negative aoa to stop the wings making too much lift.     With non angled wings,  it matters less, but efficiency drops quickly below 2degree AoA and above 5 AoA.

 

Edited by AeroGav

Share this post


Link to post
Share on other sites
On 23/12/2016 at 1:18 PM, RocketMoron said:

Thanks for the reply. 

Yes, #3 has to be followed. The majority of my spaceplanes are cargo lifters to a station at 200Km LKO. I've poured hours into honing my Mk2 spaceplanes for tiny probe launches and I had to move into Mk3 to avoid the cargo from getting stuck in the cargo bay.

 

Cargo bays are certainly convenient,  but  they add so much drag and weight you need a vastly bigger SSTO to put something in orbit than if you just bolted wings and engines on the side.

For example, using level 8 engines (poodle, nerv, whiplash) how big of a mk3 would you need to launch a small station , with a lab, hab , and cupola?

Vs. just bolting wings on the side.   Looks like i over engineered this, it was only supposed to go to low orbit !

20161020072226_1_zpsqc10uu10.jpg(2 nervs, 1 poodle, 2 whiplash)

Of course, that has the drawback of leaving some expensive spaceplane parts to gather dust in orbit for as long as you use need the station.

This takes quite a bit more design work, but in my last career game  i was able to launch my lab with 2 panthers and 3 terriers by towing it.  Gotta love autostruts !

20161108094431_1_zps6tohwxbd.jpg  by way of comparison, this mk2 cargo bay ssto with two panthers and four terriers only has room for a really small probe (and yeah, i might want to flip that bay upside down so the wings don't trap the cargo:blush: )

20161121175806_1_zpsx5gxkbqt.jpg

Edited by AeroGav

Share this post


Link to post
Share on other sites
This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.