Moar Spaceplanes

[capi3101]

Morning, y'all.

So, this is bugging me this morning - I know for rockets that there are guidelines you should shoot for for launch (4550 delta V, TWR 1.6 optimal but good as low as 1.2 if necessary). I'm curious if there are similar guidelines I need to be looking out for for spaceplanes. So I'll ask the question outright: about how much lift and how much thrust do you need per tonne of spaceplane? I ask because my first spaceplane, the Auk I, is still the only successful design I've ever built; my attempts at building anything more useful have been unsuccessful so far.

So how about it?

Edited January 22, 2014 by capi3101

[federicoaa]

mmm Moar Wings maybe?

[JumpsterG]

I'm curious for some "rule of thumb" numbers as well. I've made a few very successful designs but the build process is still very much trial and error.

I think it's possible to work out a ratio for mass to lift required to takeoff/land. I should check my designs and note their takeoff mass to lift ratios.

dV is confusing for spaceplanes. I fear this is dependent on many factors (number of intakes, how quickly you get up to altitude, how efficient your flight profile is). Reported dVs (using Kerbal Engineer Redux) between 2200 and 3200 (measuring air-breathing engines and liquid fuel) do seem to get me to orbit on my SSTO spaceplanes. I would be careful taking the dV reading, since things like rocket engines and oxidizer might get thrown into the calculation by accident. (On that note, anyone test how engineer treats Rapier engines?)

TWR can be pretty low for air-breathing engines, even below 1 will take off and maneuver thanks to lift. I don't know what is optimal though. If going for an SSTO, there's a competing optimization problem between escaping thick, low-altitude air quickly with many engines vs. carrying less mass and burning fuel more slowly with fewer. I've been rolling pretty consistently at 1.4 TWR on launch for my designs, but I have not specifically tested the low or high ends of the spectrum for efficiency.

[irongut]

On that note, anyone test how engineer treats Rapier engines?

Based on some tests I did yesterday it calculates the air breathing & closed modes separately. By default you see stats for air breathing mode, switch your engines over and you'll see closed mode stats. (This can be done in VAB.)

[JumpsterG]

Based on some tests I did yesterday it calculates the air breathing & closed modes separately. By default you see stats for air breathing mode, switch your engines over and you'll see closed mode stats. (This can be done in VAB.)

Thanks for mentioning! It is trivial to check, but I am only active on the forums in between code stints at work, so I couldn't go into the VAB right away. My 2200 to 3200 number is with air-breathing mode, for anyone interested.

[numerobis]

My rules of thumb for Kerbin SSTO ascent:

1 turbojet per 15t (that gives TWR ~.75 on liftoff)

1 Mk2 fuselage or equivalent per turbojet (160 units of liquid fuel)

1 intake per tonne

1 small control surface or wing connector per tonne, or one swept or delta wing per 3 tonnes, mix & match.

That builds you a plane that takes off at about 60-70 m/s and can fly to a 75x40 km orbit.

I am most certain about the intakes and turbojets. I'm moderately certain about the fuel (you get roughly 6 km/s for that fuel, which is about what you need). I've only recently started actually calculating the lift, so that's the one I'm least sure about.

You can use fewer engines if you have more fuel and wings, which is a net reduction in dry mass.

Edited January 20, 2014 by numerobis

[capi3101]

Okay. So a couple of questions for you, numerobis: which "intakes" do you generally use, and when you're referring to wings, do you mean pairs of wings/control surfaces or individual wing/control surface parts? For example, lessay I've got a ten tonne spaceplane - are you saying I have sufficient lift with a pair of swept wings, two small control surfaces and a twin vertical stabilizer, or do I need to double up on parts there? Do vertical stabilizers count towards generating lift, or should I count them as part of the "dead mass" in this case?

Also, canards - yes or no? (Talking about wings ahead of the main wings, not necessarily the specific parts labelled as "canards").

Your guidelines have already shown me some of the existing fundamental flaws with my spaceplane designs to date, so I do very much appreciate the help so far.

Edited January 21, 2014 by capi3101

[capi3101]

Are you SURE about one intake per tonne? I've heard that three intakes per Turbojet provide sufficient air, and so far that's been my experience (by "intakes" in this case I'm referring to the Ram Air Intake).

So far it looks like my big problem by far is insufficient lift; I analyzed the parts on the Auk IV this morning. 25.735 tonnes raw. 2 Turbojets, check. 2 Mk2 Fuselage tanks, check. 6 Ram Air Intakes - see above. Lifting surfaces are 2 standard canards, 2 standard control surfaces, 2 wing connectors and four swept wings. She's also go two AV-R8s but those are the vertical stabilizers so I'm assuming they produce no lift, so I've got enough lift for an eighteen tonne craft - and the Auk IV is heavier. So I'll try swapping out the standard control surface for two small control surfaces and adding another pair of swept wings to it - that should only add .06 tonnes of mass to the craft overall and give me the remaining lift I need, right?

I should give my other designs a more thorough analysis too - the Auk IV's the one I really want to start using though.

Can I also get some pointers on ascent profile? I've been using the video tutorial off the Drawing Board as my guide but I've seen some criticism of it.

[Claw]

I tend to use fewer intakes myself, generally aiming for around 3 per engine. Your design will certainly use more, but I find you need to really add a lot extra. (Like going from 6 to 24 with your design.) If you use MechJeb to control throttles, it's more important to have more intakes because MechJeb is a little less agressive with the throttles than what I find possible by hand.

Not using MechJeb means you'll have asymmetric thrust to deal with.

You are correct, do not factor in your vertical tail into your lift calculations. Also angled surfaces will throw off you lift ratings.

Ksp seems to like canards, but they can lead to stability problems at higher AoA (because of some KSP mechanization). However, they are yseful at getting the nose up if you are having a problem with that.

There are a lot of tutorials out there, but a lot of your profile depends on your plane build. I find that it helps to build a lot of speed in the 20's and eek it up above 32km. Seems like once you get it above there, you can start milking the TurboJets for their last bits of thrust to really push your orbit circularization out. Asymmetric thrust and draggy airplanes will hurt you up here. (i.e. if you have a lot of angeled surfaces.)

[Silverchain]

Are you SURE about one intake per tonne?

For a recent 5 ton SSTO I used 1 ram and 8 radial intakes. With that amount of intakes I could get the apoapsis above 200km, at which point the orbital insertion burn uses only a trivial amount of fuel. That's slightly less than numerobis' suggested ratio as two radials are less effective than one ram, but pretty close...

[Claw]

For a recent 5 ton SSTO I used 1 ram and 8 radial intakes. With that amount of intakes I could get the apoapsis above 200km, at which point the orbital insertion burn uses only a trivial amount of fuel. That's slightly less than numerobis' suggested ratio as two radials are less effective than one ram, but pretty close...

So as an example of design philosophy, you can follow this link and see one of the tutorials I built. (It's an aero tutorial) With the addition of a couple small rockets, it can achieve about a 200 km orbit and only uses two radial intakes. And I have an 11 ton variant with 1 TurboJet and 3 radials that gets up around 130 km.

http://forum.kerbalspaceprogram.com/threads/65638-Basic-Airplane-Space-Plane-Aero-Tutorial

Just depends on what you want and how you fly. I have been working on wing lift rating rules of thumb, but it's still a work in progress.

[numerobis]

I tend to fly with a bit less than my guidelines; they're just a starting point.

1 ram intake per tonne gets you to a 40x75 orbit, from which it's 30 m/s to circularize, so you can circularize off an ion engine or RCS. With fewer intakes, you need more rocket thrust and fuel. It's a tradeoff.

Always use the ram air intakes. The others have less area, which is what determines how effective they are at sucking up air. The ram intake has 0.1 area, the circular one has 0.08, the radial has 0.04; so the ram is getting 25% more air than the circular and 2.5x the air of the radial, even though it's the same mass.

Lift surfaces: indeed, vertical lift surfaces don't help you fly upwards, you only count horizontal ones. The numbers I gave allow you to get level flight at 15 degree angle of attack at 70 m/s on the surface, with no support from thrust. You can get away with fewer surfaces if you pitch up and use the engines to fight gravity. And yes, if the guideline says you need 7 wings, you're going to need to figure out how to place an odd number of wings, or deviate from the guideline (I'd recommend 6 in that case).

I was keeping the recommendation to reasonable-looking planes. If you don't mind silly planes, angle a small control surface in the VAB by a large number of degrees. A stupid little plane for you: cockpit, Mk1 fuselage, turbojet, radial intake; total 3.5 tonnes. Put two small control surfaces on it, tilted down 90 degrees. Throttle up, stage, and watch your plane lift off on its own: at 35 m/s those two little control surface can lift your plane.

[capi3101]

So what exactly is the advantage in going from 3 ram intakes per turbojet to one intake per tonne? I'm assuming it ensures that you have enough air intake to keep the engine running at full throttle all the way up, is that right?

Some of my designs would look like a solid wall of intakes if I put that many on them...

I ask because I have a successful spaceplane design (the Auk Ia); it's 11.815 tonnes raw and utilizes 2 Circular Intakes and 4 Radials. I was shocked to discover the Circulars when I went through the part list this morning - I'll be swapping those out for Rams for sure. Granted, that's more than three intakes, but still less than 1 intake per tonne.

Based on the data, it looks kinda like there aren't many situations where you'd want to utilize the radial intakes at all - they're heavier (0.1 versus 0.01 of the other two) and apparently they suck in less air, despite what it says on the wiki.

I too have utilized Keptin's pictorial tutorial. It's a good early design guide. I even came across yours the other day, Claw.

I have encountered asymmetric thrust at this point with the Auk IV; it was managable by throttling back. Neither of the engines flamed out but I didn't make orbit with it either. Bottom line: it needs moar lift. Running theme with most of the Auk series.

[Silverchain]

Always use the ram air intakes.

They are very much harder to mount.

[numerobis]

So what exactly is the advantage in going from 3 ram intakes per turbojet to one intake per tonne? I'm assuming it ensures that you have enough air intake to keep the engine running at full throttle all the way up, is that right?

I'm not sure what of your question I haven't yet answered, so I'll recommend you the test yourself: build a 15-tonne plane with three ram intakes, and one with 15 intakes, and compare how they behave.

[capi3101]

Very well; I'll give that a try then.

[Claw]

So what exactly is the advantage in going from 3 ram intakes per turbojet to one intake per tonne? I'm assuming it ensures that you have enough air intake to keep the engine running at full throttle all the way up, is that right?

You will get more performance, but exactly how much I can't give you an exact number (although you've given me another experiment to try out). I can't say "you will maintain full throttle at all times" because it depends on so many things.

Like numerobis said, your best bet is probably to mount them and see how it does. What I'm trying to say is you don't HAVE to have that many intakes, but when in doubt more is better (usually). If you have a really heavy design with only one engine, then you may want to consider more intakes. A lighter design would need less. I tend to do 3 to 6 intakes but I only load the planes down with 11-12 tones per engine. Numerobis loads more weight and uses more intakes.

[capi3101]

I suppose that intakes don't weigh a great deal; Ram Intakes only weigh ten kilograms, so it'd take a hundred of them before you added a full extra tonne of mass. Setting them up to toggle in the action groups might get hairy.

Anyways, I have something to try out this evening - I'll let y'all know how things go.

[Ravenchant]

1 intake per tonne

Whoa... which intakes are we talking about? If you mean rams, then this is way more than neccessary. My last satellite lifter weighed 20 tons and got into orbit with 4 ram intakes (2 per turbojet) easily.

...to be fair though, a lot of that tonnage was oxidiser

[Claw]

Whoa... which intakes are we talking about? If you mean rams, then this is way more than neccessary. My last satellite lifter weighed 20 tons and got into orbit with 4 ram intakes (2 per turbojet) easily.

...to be fair though, a lot of that tonnage was oxidiser

It certainly depends a lot on how you want to build and how you fly. I'd be interested in what other considerations you had in your design.

[Claw]

I suppose that intakes don't weigh a great deal; Ram Intakes only weigh ten kilograms, so it'd take a hundred of them before you added a full extra tonne of mass. Setting them up to toggle in the action groups might get hairy.

Anyways, I have something to try out this evening - I'll let y'all know how things go.

Sorry, I didn't mean to sound like I was blowing you off or making you do something. Just suggesting to give it a try and see if it works for you.

If you want ideas of how to put on intakes without just plopping them all over the place, here is an "inlet stick" I built as a sub assembly to slap onto an airplane if I wanted it. This model is built with five intakes. You could, of course, turn on symmetry and have ten intakes slapped on pretty quick.

If you look at how it's built, I put it on an I-beam so that I could easy remove/move/replace them during wing assembly. I added the nose cone on back to hide the structure and make it look a little cleaner. This build isn't optimized for part count if that is a consideration.

[Claw]

Oh, I did find some of my notes on when I was testing various combinations of this stick inlet.

Inlets - Apoapsis

2 - 33.4 km

4 - 37.4 km

6 - 39.2 km

8 - 39.5 km

10 - 39.6 km

I think I did this climbing out at 45 degrees nose high, but can't remember exactly. So this isn't really an optimal test for getting to orbit, but it gave me the basis for my 3 to 6 inlets per engine. I was planning on doing more inlet experiments after wings.

Incidentally, the test airplane weighed about 6t which means numerobis' rule of thumb held for this particular case.

Edit: Sorry for the spam posts!

[capi3101]

I'd like to know how to build that Intake stick. Do you have a step-by-step procedure?

Numerobis rules of thumb work! Tweaked the Auk IV an extra pair of wings, ditched its large control surfaces in favor of two small control surfaces, and added a combination of Ram Intakes and Radials (I know radials are a bad idea, but it was the only way I knew at the time to get 26 of the stupid things on there). Got it up to orbital velocity by 38k, didn't have to kick in the nukes until a full minute later (just after 40k). Had the Ap where I wanted it at 48k.

Thanks for y'all's help once again.

[Claw]

Congratulations on getting to space! I'm glad numerobis' rules helped you out.

I ran (quite a few) tests with intakes while I was doing some wing testing. Here's the nuts and bolts of what I found with a 9.2t craft. With more intakes, it weighs slightly more but I didn't spend time getting that zeroed out. All with 1 TurboJet...

1) 1x Ram Intake got me to a 34 x -250 km orbit. Probably not the best run, but I was getting tired.

2) 3x Ram Intakes = 40 x 117 km orbit. Was able to run full throttle till 39 km. Final flameout was around 60 km.

3) 11x Ram Intakes = 40 x 215 km orbit. Was able to run full throttle till 43 km. Final flameout was around 61 km.

I used the intake stick for test run 3.

So you can do it with 3 intakes, or 1 per ton. Just depends on what you want to do. A heavier craft would probably benefit having more intakes.

Here's how to build it. Let me know if you want pictures. Perhaps I'll add this to my intake tutorial.

1) Get an I-beam pocket edition and place it somewhere level on your fuselage while you're working. Press D, W to rotate it and lay it down before placing it. It should be pointing forward. (You're holding the aft end.)

2) Place a cubic octagonal strut on the I-beam. Press W to lay it down before placing it. Again, it will be pointing forward.

3) Place a ram intake on the cubic strut. The back end of the intake should be connected to the front end of the strut. You might have to wiggle the intake around to get the clipping right. (You can put the intake on the back of the strut also, but then you can't place the nosecone on the back.)

4) Make a copy of the strut/intake combo. Hold left alt while you left click on the cubic strut (not the I-beam).

5) Slide the strut/intake combo back on the I-Beam until it flips the correct direction. This is about as far forward as you can place it.

6) Repeat the copy/place (steps 4/5) until you have five intakes. You might be able to fit a sixth.

Drop it in your subassembly folder if you want. When you want to move it, grab it by the I-Beam at the back. Cap it off with a nosecone (placed backward) if you want. Because it uses an I-Beam, you can place the intake stick down first, then slide a wing root over the top of it to cover it up.

Good luck!

Edited January 22, 2014 by Claw

[capi3101]

I went back and checked out derpnerd's Advanced SSTO video tutorial last night (still bugs me when folks use "SSTO" and "spaceplane" synonymously - I still use single stage rockets all the time), and he suggested 10 to 15 intakes per engine; I missed that the first time I watched the video. Given a single engine for approximately 15 tonnes, that does work out to about 1 intake per tonne.

So...the take away from all this. The way I'm reading it, 2.6 tonnes of equipment will lift up to 15 tonnes, for 12.4 tonnes net payload. I get that 2.6 figure from this set of equipment -

1 Mk2 Fuselage (1 tonne)

16 Ram Intakes (0.16 tonnes)

16 Cubic Octagonal Struts (no mass)

1 Turbojet Engine (1.2 tonnes)

4 Swept Wings (0.2)

4 Small Control Surfaces (.04)

Took the time this morning to analyze the ratios of lift to mass for the various aerodynamic parts - here's what I came up with:

AV-R8 = .4/.02 = 20 - supports 1 tonne

Standard Canard, Adv Canard, Std. Control Surface = .7/.04 = 17.5 - supports 1 tonne

Small Control Surface = 0.5/.01 = 50 - supports 1 tonne

Delta-Deluxe = 0.7/.02 = 35 - supports 1 tonne

AV-T1 = .3/.05 = 6 - supports 0.5 tonne

Tail Fin = .3/.02 = 15 - supports 0.5 tonne

Delta Wing = 1.9/.07 = 27.14 - supports 3 tonnes

Swept Wing = 1.6 / .05 = 32 - supports 3 tonnes

Structural Wing/Wing Connector = 1/.05 = 20 - supports 1 tonne

It looks like the Small Control Surface offers the most bang for your buck...so I've got to ask, has anybody ever tried using just a bunch of Small Control Surfaces for their main wings? That'd be an interesting thing to see.

While I'm posting, let me ask y'all about vertical stabilizers and landing gear - I know you're supposed to double up on the fantail and keep it vertical. My question is - which part is best for that purpose? I've been using AV-R8s for the most part but the Auk III utilized Standard Canards; does it really matter which part you use?

And gear - my question is: how much mass will a landing gear support before it starts to buckle? I suppose a better question is how many gear bays would you recommend per tonne of craft?

The gear and the tailfins would cut into your payload, that's why I ask. The rest of the payload then becomes the fuselage and any interplanetary/interlunar equipment you want to haul around, considerations you'd make just like a rocket.

EDIT: Claw, thanks for the procedure. I'll give it a shot and let you know how it works out.

Edited January 22, 2014 by capi3101