My newbish spaceplane efforts: suggestions for improvement?

[Mister Spock]

I spent much of today making spaceplane after spaceplane, and I had a great time with it. At first I was scared of trying to land, but by the end I was landing on the KSC runway, rough terrain, islands, you name it. I now see why people say the stock KSP aerodynamics are like pea soup: my multi-ton jet glides to a perfect landing every time, with all engines off. Anyway, I didn't achieve my goal, which was to carry a large payload into space and then land back on Kerbin. I have four specific questions, followed by pictures of various designs, as to which please feel free to comment.

1. I had trouble marrying engines to fuel. Does every engine have to be directly behind its fuel sources? It seems maybe jet fuel is pumped anywhere on the plane, but rocket fuel has to be directly over the receiving engine?

2. I also had trouble with the center of mass shifting as fuel was consumed. Radial fuel tanks would seem to be the answer, but when I tried these, I got more unstable flight.

3. I made hilarious attempts to lift an 11t payload into space. I got as far as 30,000m once. Is 11t out of the question for a spaceplane? What is a reasonable payload? I actually was able to strap the payload on okay; my trouble was keeping my engines fueled and my aircraft balanced. (See questions 1 and 2 above.)

4. I've read older posts suggesting there's no cargo bay in stock. Is this still true? I don't have one yet, but I haven't unlocked the whole tree.

5. Finally, I hope this is the appropriate forum to ask for advice to improve my designs. My starter designs were single-engine jets, just to get me used to flying and landing. Even with the Basic Jet Engine, these work great for exploring Kerbin. Here's an example:

Then I tried turbojets plus the spike rocket. This plane didn't make it to space; it sucked down fuel fast. Then again, I may have given up on it too quickly.

Next I fired up the RAPIER, first in a single-engine configuration, to see how far it would get. It got pretty far, though I don't think I had enough air intake. So I optimistically strapped on a big ol' 8 ton payload. That barely made it off the runway. Here's a look:

Finally, I tried strapping on three RAPIERS, in various configurations, with fuel tanks inline at first, then arranged radially. Most of these designs were unstable. The most successful is here, but the two outer engines lost power once the inline jet fuel was exhausted, and I went into a spin. (Managed to recover and make a nice landing, though.)

Any comments or suggestions? Thanks in advance.

Edited August 2, 2014 by Mister Spock
Minor edit.

[Sirrobert]

You can't just strap a payload ontop of a plane and expect it to go anywhere.

You have to make sure the Center of Mass is inline with the Center of Lift(the big yellow orb, and the blue orb)

[Wanderfound]

I spent much of today making spaceplane after spaceplane, and I had a great time with it. At first I was scared of trying to land, but by the end I was landing on the KSC runway, rough terrain, islands, you name it. I now see why people say the stock KSP aerodynamics are like pea soup: my multi-ton jet glides to a perfect landing every time, with all engines off. Anyway, I didn't achieve my goal, which was to carry a large payload into space and then land back on Kerbin. I have four specific questions, followed by pictures of various designs, as to which please feel free to comment.

1. I had trouble marrying engines to fuel. Does every engine have to be directly behind its fuel sources? It seems maybe jet fuel is pumped anywhere on the plane, but rocket fuel has to be directly over the receiving engine?

2. I also had trouble with the center of mass shifting as fuel was consumed. Radial fuel tanks would seem to be the answer, but when I tried these, I got more unstable flight.

3. I made hilarious attempts to lift an 11t payload into space. I got as far as 30,000m once. Is 11t out of the question for a spaceplane? What is a reasonable payload? I actually was able to strap the payload on okay; my trouble was keeping my engines fueled and my aircraft balanced. (See questions 1 and 2 above.)

4. I've read older posts suggesting there's no cargo bay in stock. Is this still true? I don't have one yet, but I haven't unlocked the whole tree.

5. Finally, I hope this is the appropriate forum to ask for advice to improve my designs. My starter designs were single-engine jets, just to get me used to flying and landing. Even with the Basic Jet Engine, these work great for exploring Kerbin.

1) Fuel needs to be in the stack above the engine (in rocket terms, so for planes that'd be the fuselage in front of the engine) or connected by fuel lines (those yellow pipe things). Fuel lines are very useful for aircraft; by arranging them carefully, you can make the tanks drain in the order you want them to (engines drain the "furthest" tank first, and you can manipulate this by chaining your tanks into the order you want them to drain). You can use this to keep the plane balanced regardless of the fuel load if you do it right.

2) See (1). The other trick is to try and keep your fuel tanks clustered about the CoM as tightly as possible. Avoid putting tanks near the nose of the plane. Ideally, you want to keep CoM steady, but if it has to shift you're usually better off letting it go forwards rather than backwards.

3) A well done heavy lift spaceplane can toss more than its own weight into orbit; for the big ones, this can be over 100 ton. But it's not an easy thing to design. Get the basics down before you try to summit Everest. Also, as mentioned above, it's not just a matter of strapping a load on the back; you need to keep the plane balanced. Having the CoM above the CoL is almost as bad as having it behind CoL. Either of these will make the plane handle very poorly.

4) No, there is no stock cargo bay. However, there are an assortment of mod packs that include them, and if you're creative then it isn't too hard to construct one from stock parts. See http://forum.kerbalspaceprogram.com/threads/88628-Kerbodyne-D7-Heavy-X5-a-heavy-lift-SSTO-spaceplane for an example of an all-stock SSTO cargo spaceplane with an open-air bay. If you look around in the Spacecraft Exchange you'll find assorted examples of people constructing sealed bays from structural panels and wing pieces.

5) Yup, all cool. There are a large group of spaceplaners here who are always happy to critique designs (and most of 'em will even do it politely...).

Finally: I strongly recommend that you give FAR or NEAR a try. Spaceplanes are much more fun when they actually fly like planes.

Edited August 2, 2014 by Wanderfound

[MarvinKitFox]

You are on the right track, but you have many miles to go yet.

Try to have your craft have aligned CoM(centre of mass) and CoT(Contre of thrust) at all times.

When full of fuel. When empty. When carrying its payload. *at all times*

This will require you to rethink your design philosophy. Mounting cargo on top is mechanically easy, but aerodynamically unsound.

Remember that you have the option of, for example, having engine groups switch on and off via action keys.

You also want your centre of drag aligned with your centre of mass, and *behind it* if at all possible.

As VAB doesn't show centre of drag, this can be tricky.

You need a solid enough plane to withstand the rigors of launch and reentry... You seem good on that, your designs are butt-ugly but very solid.

You need enough thrust to both fly and gain some real speed.

Realistically, this means turbojets or rapiers.

Try for 1 turbojet per 8 tons of plane, *minimum*. That is plane weight including engines, fuel and payload all together.

You need enough intake air to feed your turbojets.

For a fuel-efficient spaceplane, this means 4 to 6 ram inlets per turbojet.

Note *ram* inlets, the others are not good at high altitude and speed.

Those radial scoops you seem to like are very nice for subsonic flight, but they just don't work right for the role of hypersonic orbiter.

Consider what I've done with your design.

Basically, I've twinned the hull, to allow a twin-boom fuselage, with a space inbetween them for the payload.

This fixes *all* of the balance and stability issues, and provides a clean means of dropping the cargo, via a docking port.

Slapped on some more wing, to allow easier takeoff.

I threw a pile of ramscoops on the *rear* wing, to provide air yet keep drag to the rear.

This guy can operate in two modes..

Either go to 41km at 2150m/s, drop the payload which then flies of (that's whats in my screenshots)

or at that same alt, engage the nuke to thrust to orbit, seeing as KSP is not very forgiving of me trying to fly two different craft in different directions whilst in atmosphere. Silly, really!

Note that the plane is mostly empty air. ***ONLY*** the second-from-front jet fuselage contains jetfuel, 150 units is *ample* to get to the desired alt+speed.

If we go on to orbit, we loan the payload's engine to do so, but not its fuel. Thats why the two rear pods are fueltanks (usually empty). Put about 40 units of fuel+lox in each, and it is sufficient to orbit.

And still, this really is the same plane you built.

Edited August 2, 2014 by MarvinKitFox

[GoSlash27]

-I always design my aircraft around the fuel tank so that the center of the fuel tank is also the center of mass. Balance everything else around that like a see-saw; same torque moment (mass*distance)in front of the tank as behind.

-Wing loading is important. You want enough wings on it to maintain level flight at NLT 30 Km altitude. Otherwise you never get to a point where your terminal velocity can reach orbital velocity. Your most efficient lift-to-drag should happen at around 23 degrees angle of attack.

-build your plane to be dynamically stable; center of mass a little ahead of center of lift so it wants to nose down. Wings sloping uphill (dihedral) so it naturally wants to fly wings level, etc.

-KSP generates 2 types of drag; parasitic and induced, and they happen at your CoM and CoL. At low altitudes and speeds most of your drag is induced and works at your CoM. At high altitudes and speeds it's reversed. Your CoL should be in the longitudinal axis so that when this happens your aircraft won't spin out of control.

Best,

-Slashy

[Jimbimbibble]

B9 has a lot of great spaceplane parts. You can get 50+ tons into orbit all inside a pretty looking cargo bay.

[Mister Spock]

Thank you for all the very useful replies! I've found rocket design pretty intuitive, but spaceplane design seems much more challenging. So I appreciate the help.

MarvinKitFox and Wanderfound: I hadn't thought of creating a twin-boom fuselage. Marvin's plane, and the one linked by Wanderfound, both have this feature, like the old P-38 Lightning. Looks great, and a good solution for carrying payload on the same plane as the center of mass.

MarvinKitFox, how does the jet fuel get from your center fuselage to the engines in the rear side fuselages? I can't tell if there are fuel lines connecting them. I assume the fuel in the rear tanks is rocket fuel?

Slashy, I'll work more on my wings. Yeah, mine have no dihedral angle at all -- straight parallel to the ground. I'll try tilting them up a tad.

Also, I had trouble attaching the standard control surfaces to wings -- they seem to want to go on at crazy angles, not parallel to the ground as one would expect. I had no trouble attaching small control surfaces, though. Maybe the standard surface is designed for a larger wing than the one I was using?

[GoSlash27]

Also, I had trouble attaching the standard control surfaces to wings -- they seem to want to go on at crazy angles, not parallel to the ground as one would expect. I had no trouble attaching small control surfaces, though. Maybe the standard surface is designed for a larger wing than the one I was using?

I have better luck with turning off the angle snap and using symmetry mode. Also, holding shift while rotating a part will rotate it in 5 degree increments instead of 90.

Best,

-Slashy

[Wanderfound]

Also, I had trouble attaching the standard control surfaces to wings -- they seem to want to go on at crazy angles, not parallel to the ground as one would expect. I had no trouble attaching small control surfaces, though. Maybe the standard surface is designed for a larger wing than the one I was using?

The trick: stick one wing vertically on top of your plane (i.e. 90 degrees straight up). Put the control surfaces on; they'll usually still be misaligned to start, but it should be by exactly 90 or 180 degrees off, so it's easy to rotate to correct alignment.

Once the control surfaces are placed, take the wing off the top of the plane, toggle symmetry on and place them where you actually want them.

[Wanderfound]

One thing to be careful of with the twin boom design: the nature of KSP means that you have to build them like a tuning fork. Despite appearances, the only thing holding the back of the D7 together is a lot of struts. Brace carefully.

[allmhuran]

One thing to be careful of with the twin boom design: the nature of KSP means that you have to build them like a tuning fork. Despite appearances, the only thing holding the back of the D7 together is a lot of struts. Brace carefully.

Yes indeed, maintaining rigidity can be a challenge. If I'm building a spaceplane shaped like this, with twin hulls to carry cargo in the middle, I generally use structural girders as the fork prongs. They are quite rigid (joints don't flex much) and structurally strong.

And, like GoSlash, I generally put the fuel at the centre of mass. In fact, I'll go so far as to only use odd numbers of fuel tanks, and then put a fuel line that drains from the middle one (at the CoM). In this configuration fuel will draw evenly from the front and rear into the middle. I then also usually mount the engines on the wings, not at the back, because it's easier to pipe fuel from the central tank to the engines when the engines are also near the centre. Tack on a bunch of wings and a bunch of RAM air intakes (towards the rear to keep the drag behind the CoM... but not TOO far towards the rear or the plane will resist pitching), and you don't need much rocketry at all to push a spaceplane the last few kilometers into orbit.

Taking all of those ideas to the absolute extreme results in something like what you see in the video below (this link will take you to the appropriate time to see the aircraft in question), which notably only used four rockomax 24-77 engines for the final push to orbit.

[Voculus]

Allmhuran, that was one impressive ship! Nice vid, and funny, too.

[Laie]

1. I had trouble marrying engines to fuel. Does every engine have to be directly behind its fuel sources? It seems maybe jet fuel is pumped anywhere on the plane, but rocket fuel has to be directly over the receiving engine?

If you can ask such questions, you're ready for THE TRUTH. Look here, particularly post #8: http://forum.kerbalspaceprogram.com/threads/64362-Fuel-Flow-Rules-%280-23-5%29?p=880734&viewfull=1#post880734

Also check out the part about how air flow is distributed among engines.

[Mister Spock]

Thanks for the further replies. Yeah, I noticed lots of struts on the back of the D7. I will follow suit.

[Wanderfound]

The D7 struttage is a bit on the extreme side; the heavy lift nature of it meant that I initially had issues with rapid unplanned disassembly and landing gear failures during takeoff. Smaller planes don't need to be quite as spiderwebby.

But, hey, struts are cheap. Err on the side of excess.

[Mister Spock]

Laie: Thanks for the link. I think I'd read some of Kasuha's findings before, particularly his comment that air intake is processed depending on when the intake part was added in the Space Plane Hangar (or VAB). That is, that the order of construction influences air flow. I gather from your link that some of the same is true of fuel flow?

Allmhuran: Awesome video!

I like the idea of fuel tanks in the middle of the plane, with fuel lines feeding the middle tank. But what if one needs both jet and rocket fuel? The two different types of fuel presumably can't be pumped into each others' containers, can they?

[Wanderfound]

Laie: Thanks for the link. I think I'd read some of Kasuha's findings before, particularly his comment that air intake is processed depending on when the intake part was added in the Space Plane Hangar (or VAB). That is, that the order of construction influences air flow. I gather from your link that some of the same is true of fuel flow?

Allmhuran: Awesome video!

I like the idea of fuel tanks in the middle of the plane, with fuel lines feeding the middle tank. But what if one needs both jet and rocket fuel? The two different types of fuel presumably can't be pumped into each others' containers, can they?

Into, no, through, yes. Not a problem. "Jet fuel" is just rocket fuel minus oxidiser.

[Wanderfound]

(snip Whackjob vid)

Kudos for the Great Escape soundtrack.

[allmhuran]

I like the idea of fuel tanks in the middle of the plane, with fuel lines feeding the middle tank. But what if one needs both jet and rocket fuel? The two different types of fuel presumably can't be pumped into each others'

containers, can they?

Fuel and oxidizer will both flow through any part that is fuel crossfeed capable, which is most parts (excludes decouplers, ibeams structural panels and a few others). In the craft pictured below, I have one FLT800 fuel tank in the middle carrying fuel and oxidizer, and one jet fuel tank at either end of that FLT800.

The engines are attached to structural fuselage pieces on the wings to also be near the CoM, which is just behind the middle of the FLT800 tank. When the craft is completely out of fuel the CoM shifts very slightly rearwards. If I had the engines at the back, then it would shift much further rearwards.

Instead of structural fuselages to mount the engines you could use more fuel tanks here, or engine nacelles, if you wanted more atmospheric range, but you definitely don't need it (see how much fuel is left in the second image).

I quickly built this to make the fuel lines very visible so it's obvious what is going on, normally I hide the fuel lines for aesthetics.

The fuel path to the jet engines is from the jet fuel tanks, into the FLT800, into the wing, and then into the structural fuselage feeding the engine.

The fuel path to the small rocket engines is the same. For the oxidizer path it starts at the FLT800 since jet fuel tanks contain no oxidizer, other than that it's the same.

In the second image, you will see that in flight the jet fuel tanks are draining evenly from the front and rear jet fuel tank.

Notice also the extra RAM air intakes towards the back of the craft. With 10 RAM air intakes total (5 per engine) I can happily run at full throttle at 26000m, gaining much speed. As I gain more altitude I would slowly decrease throttle to prevent flameout, and would probably get to around 2300m/s at about 40000m (at which point my apoapsis would be well into space) before switching over to the little rockets and finishing off normal orbit insertion.

Edited August 3, 2014 by allmhuran

[Mister Spock]

Allmhuran, thanks for that extremely helpful diagram. I'm embarrassed to admit that I didn't understand what "crossfeed capable" meant until I read your post. Now I understand why I see "not crossfeed capable" on decouplers and such.

Anyway, that's a really enlightening diagram. I think I understand why you need the fuel lines running radially from the central tank to the wings: by default, KSP engines look for fuel sources "above" them, not to the side, yes?

But I don't see fuel lines running from the front fuel tank to the center tank, or from the rear fuel tank to the center tank. Are such fuel lines unnecessary, or do you have hidden fuel lines leading from the rear tank to the middle tank, and from the forward tank to the middle tank? (In my experiments with asparagus staging, I've had to run fuel lines from "outer" tanks to the central tank to ensure that fuel drains from the outer ranks first.)

Thanks again for a great post -- and a great video.

Edited August 3, 2014 by Mister Spock
Glitch removal.

[allmhuran]

KSP will, by default, draw from whichever fuel tank is the "furthest away" from the thing using the fuel, and it will go through any crossfeed-capable parts to get it.

So I don't need fuel lines from the jet fuel tanks to the FLT800. The jet engines will go like so (so, starting from the jet engines)

"Nom nom nom where's mah fuel?

There's a fuselage piece attached to me... but it has no fuel.

Oh look, a fuel line, I wonder where that leads.

Into a wing... OK, wings are crossfeed capable, where does that lead?

To another fuel line! Let's go up there.

AHA! An FLT800 fuel tank! This has fuel that I could use. But let's see if there's any further away....

Right.. a jet fuel tank is one piece further away. And there's another jet fuel tank on the other end too, just as far away as the first. OK, I'll drink those two tanks first, then move on to the FLT800"

So, just like you dont need fuel lines between node-attached (end to end) fuel tanks when building a rocket, you also don't need them to get fuel to flow from the jet fuel tanks through the FLT800.

Just for completeness, it's worth noting that the little rocket engines will do exactly the same thing looking for fuel. Whey they look for oxidizer they will reach the FLT800 and stop there, because jet fuel tanks don't contain oxidizer. What this means is that if you still have fuel in the jet fuel tanks, the rocket engines will use fuel from those first, while draining oxidizer from the FLT800. But that doesn't mean your jet engines don't have any fuel left... they will happily draw fuel from the FLT800 without using up the oxidizer.

Edited August 3, 2014 by allmhuran

[Mister Spock]

Thanks, allmhuran -- now I get it! Now I'm ready to go design a better plane!

So I'm curious -- what percentage of your launches (or those of other posters in this thread) are spaceplane launches? Can planes do the lion's share of the lifting work of a space agency?

Thanks again for taking the time to answer my questions.

[Wanderfound]

Yes, you can run a pure spaceplane program. It isn't too hard to design something that can lift 100 ton or better, and anything bigger can be assembled in orbit.

I only use rockets when I'm in a hurry. Spaceplanes are cheaper to fly and more interesting to build. Rockets are faster and simpler.

Edited August 3, 2014 by Wanderfound

[Mister Spock]

Spaceplanes are also more fun to fly, often. I've spent the day buzzing around Kerbin. "Flight over Kerbin" contracts are generally a lot easier with spaceplanes than rockets.

A further question: I find myself using the little cube struts to attach ram scoops. And I've used a plate, or a long thin stick, to attach the rear landing gear. Are these sort of kludges unusual? I kinda wish there were a ram-intake-attacher part.

Along the same lines, my small control surfaces seem to hover about six inches over the wing. They work; they just look strange. Maybe I'll post a pic later to show you what I mean. Gotta run right now.

[Wanderfound]

Spaceplanes are also more fun to fly, often. I've spent the day buzzing around Kerbin. "Flight over Kerbin" contracts are generally a lot easier with spaceplanes than rockets.

A further question: I find myself using the little cube struts to attach ram scoops. And I've used a plate, or a long thin stick, to attach the rear landing gear. Are these sort of kludges unusual? I kinda wish there were a ram-intake-attacher part.

Along the same lines, my small control surfaces seem to hover about six inches over the wing. They work; they just look strange. Maybe I'll post a pic later to show you what I mean. Gotta run right now.

Those sorts of kludges are common, but they're not necessary. Have a look at this one for an example without them: http://forum.kerbalspaceprogram.com/threads/89092-Kerbodyne-Scattershot-a-simple-and-easy-to-fly-beginner-s-SSTO-spaceplane

There's no reason why you can't build a good SSTO spaceplane without resorting to such tricks. Airhogging, stacked wings, excessive part clipping; these are all just things that let you get away with otherwise poor design. There's nothing wrong with doing that if it's fun for you, but it's not the only way.

There is a ram-scoop attacher part, BTW: the nacelles. It's what they're made for; intake on front, engine on back. You don't need five hundred ramscoops to go to space.

Edited August 3, 2014 by Wanderfound