SSTO Spaceplane Design Changes with Realistic Aerodynamic Update

[Kicked to the Kerb]

Firstly, as I am new posting to the forum, (though I have been playing since Early Access), I was not sure where to put this. I have read a lot of threads surrounding this topic but none that explain clearly the differences since the Aerodynamics were updated.

So... After getting to the stage where I felt comfortable building rockets of all shapes and sizes, and landing them on various bodies in Sandbox mode, (at that time the only option), I moved onto SSTO spaceplanes design. After MANY hours of trial and error and inevitable explosions, (YAY KSP ), I finally became comfortable with designing those too. Anything - from a Plane/Rover 1 man Hybrid (drone control optional), through a 10 man "Ker-bus" with space for a deployable probe, to a SSTO/nuclear hybrid which could not only reach orbit but get all the way to Laythe, all with the ability to get into orbit without ejecting any fuel tanks - were the norm for me.

However I have now hit a major snag in my SP designs. For some reason since the Aerodynamics model was updated, (a great move imho), I have struggled to reach orbit. I usually design my SSTO Spaceplanes around a RAPIER Engine or 2, for their multi-purpose functionality, but have designed many without (Using rockets when the Turbo RAMJETs run out of altitude). However using either method I have currently not been able to replicate my previous success, and I am at a loss.

The main problem seems to be the "jump" from traveling anywhere between 850-1400 m/s up to the beyond 2000 m/s required for orbit, whilst retaining enough fuel, and if I add tanks, that only increases weight/drag, lowering the speed I can get up to on the jet engines, and increasing the time taken getting there. I just seem to be running out of fuel too quickly, and cannot find the right balance.

I feel like I have gone from knowing exactly what a design will do before I fly it, to expecting some unknown failure to befall me...

Any help would be MUCH appreciated (and I apologise if this is in the wrong section!)

Many thanks,

KTTK

[cybersol]

Exact right place. My recommended spaceplane ascent profile I just posted here. For the design part, GoSlashy has a good guide in the tutorials section here.

General rules are that you will need more engines than before to quickly punch through transonic drag. Because engines now lose thrust with altitude and at very high mach speeds, you won't be able to linger the plane in atmosphere all the way to orbit. You will need to get what you can get from the air breathers, and then switch to rockets sooner and get out of there. All in all you will need something like 1600-1800 dv after switching to rockets for stable LKO.

[Kicked to the Kerb]

Thank you for the links cybersol, I will check them out right away. That makes sense about the altitude and speed changes to the engines, though I perfected the technique of "lingering" so I'll now need to learn a new skill (story of my Kerbals' life) That's quite a lot of dv once you've already made the effort to get up to high speeds/altitudes, but tbh the added realism is a welcome addition. I'll check out the linked threads and have a go at hitting orbit. Wish me luck!

KTTK

[Coga19000]

As regards shape, if your planes roughly follow the very basic laws of aerodynamics, your set.

Now, for the engines, it's more complicated. SSTOs are almost impossible with Basic Jets, but Turbojets should still work fine. Air hogging also no longer works. And even though RAPIERs still make the job easier, Nerv designs are actually better for spaceplanes (since they now consume the same fuel as airbreathing engines)

[Mic_n]

Yup.. the big difference isn't really in the aero, it's in the engines.. they now have both altitude and velocity curves.. the rapiers are now more than a dual-mode engine.. You basically have a turbofan, a turboramjet (think SR71), and then the rapier in air-breathing mode is something like a mild scramjet.. they each work best in their own areas and then peter out.. the rapiers work at the fastest speeds and highest altitudes, but it's still nothing like what you could airhog in the old days.

[Kicked to the Kerb]

Wow, thanks for the swift responses! The Kerbal Community is awesome!

I had no idea about the RAPIERs being technically 4 engines in one, I thought it was just 2! That changes a lot in terms of design, I'm gonna have a mess around with them and see what I can stick together. I was also toying with the idea of a nerv/Turbo ramjet hybrid, but wanted to perfect my simpler designs first so I can build with confidence, you make a good point though Coga19000, I'll give that a whirl too.

I've never had a problem with the aerodynamic design challenge as my dad was a pilot, but it sounds like I need to be stacking more engines or at the least more dv to make the leap from Upper Atmo to lower orbit. I will post my results soon!

Thanks again for the help, you've all given me great input and food for thought!

KTTK

[Chaos_Klaus]

I had no idea about the RAPIERs being technically 4 engines in one

No, I think he meant that the rapier is the third part in the succession of engines. The basic jet engine is resempling a real world turbofan while the turbor ramjet is just what it says it is. The Rapier in air breathing mode resembles a scramjet.

[RuBisCO]

Here is what I have learned from making the Xylol

1. Wings, pre-coolers and Mk1 Jet fuel tanks wick heat nicely from intakes and engines.

2. Turbojets help: Rapiers perform poorly at low altitudes. Turbojets can make up for that.

3. Dive! When lifting heavy loads I go 30° to 10 km, then gently dive at 0° to break though the sound barrier, once past 600 m/s I begin climbing back up to 20° pitch.

4. Close intakes, I close the per-coolers until speed exceeds 1 km/s, I think this improves performance but honestly I have not noticed, other though live by opening and closing intakes at the right times.

5. Make the wings the right way! Because of the size of the Xylol the wings needed to be made out of multiple parts and assembled in a particular way to maximize stability. The Xylol's wings are made out of 4 sets of wings, each one attached to a different segment of fuselage, I call these sub-wings. The sub-wings are strutted together such that every part of a sub-wing has at least one strut connecting it to another sub-wing.

I learned the hard way that if I made a single wing with just one part connected to the fuselage and all the other parts connected to it, even with equal numbers of struts connecting the wing to ever segment of fuselage it is unstable. Each strut adds weight and drag but if the wings flex and deform too much your craft is going to have phantom yaw, roll and pitch, loss speed from snaking, worming and dolphining, even if you get MJ to try to make the thing stable enough to fly parts are going to break or overheat, and generally your craft will never make it to orbit, so trade offs.

[Mic_n]

No, I think he meant that the rapier is the third part in the succession of engines. The basic jet engine is resempling a real world turbofan while the turbor ramjet is just what it says it is. The Rapier in air breathing mode resembles a scramjet.

Yup.. sorry about the ambiguity there. The Rapier is still a 2-in-1.. I was just describing the 'progression' from turbofan to turboramjet to the rapier's airbreathing mode. The Rapier really doesn't work at all well until it gets up to speed. When it does though, it delivers more thrust at higher speed and higher altitude than the turboramjet (which in turn outperforms the turbofan).

Nerv/Turbos are certainly intriguing, but it's a difficult one to get right... the turbos aren't much good beyond around 20km and 1000m/s (plus or minus a bit), which still leaves a lot of acceleration and a lot of atmosphere around for the nukes to push through, and they have a lousy TWR to do it with. The more you add, the more weight and the more drag (very significant since you're in the middle atmosphere still).

I haven't played around with the combo much in the 0.9+, but I haven't managed to get it right yet. After having run even turbojet/ion engine combos previously, it's a big change.

[Kicked to the Kerb]

Yup.. sorry about the ambiguity there. The Rapier is still a 2-in-1.. I was just describing the 'progression' from turbofan to turboramjet to the rapier's airbreathing mode. The Rapier really doesn't work at all well until it gets up to speed. When it does though, it delivers more thrust at higher speed and higher altitude than the turboramjet (which in turn outperforms the turbofan).

Oh I see, yes I understand where I got confused, it's not just your info, I double checked on the KSP Wiki about RAPIERs and it talks about it "being the first engine with multiple engines incorporated (4 to be exact)", or something along those lines. I thought it was talking about 4 engine MODES not 4 physical engine outlets (The 4 leaf clover shape of the outlets at the back), hence the confusion!

I think my main error has been my flight profile, I didn't realise they had even built in the transonic! Way to go devs!!!

I read all the links, and I think the main tips I have taken are that I need a higher TWR, I need to make the "jump" to rockets sooner i.e. no "lingering" at the edge of atmosphere, and most importantly, a dive to breach the transonic at around 10km, as many have suggested. I'm gonna go and mess around, I'll post results soon.

Once again thank you all for your feedback. Though I've been reading the forums for a long time now, I usually choose not to get involved, (more responsibility), but I am so glad I did, and it doesn't seem like a hassle at all.

I guess if we're all flying little green men with a silly sense of humour into space, I should have expected a kickass community!

KTTK

- - - Updated - - -

Make the wings the right way! Because of the size of the Xylol the wings needed to be made out of multiple parts and assembled in a particular way to maximize stability. The Xylol's wings are made out of 4 sets of wings, each one attached to a different segment of fuselage, I call these sub-wings. The sub-wings are strutted together such that every part of a sub-wing has at least one strut connecting it to another sub-wing.

I have recently started designing my wings a little like this. Before that I would "layer" them. What I mean by that is I'd have Deltas as the main wing, and for extra lift I'd take the small Deltas, attach them to the fuselage above and below the main Delta (One either side of the main wing) and angle them so they all met at the same point at the wingtips (not parallel) I guess you could visualise it like this : /|\ (only it would be on its side and the ends of the wings touch).

This was an EXTREMELY useful method of adding extra lift without compromising the width or length of the spacecraft, though I would imagine the new Aerodynamics would affect this setup, as the flow of air would be partially obstructed. Has anybody tried this method since the new model?

KTTK

Edited June 7, 2015 by Kicked to the Kerb

[kStrout]

An interesting thing I have realized is that due to the new areo system, having very swept back wings helps tremendously, even if they are tiny. I recently made a 8 kerbal SSTO where the only wings were the large wing strakes.

[Mic_n]

Personally I like to make the dive from a little higher up... the turboramjets don't really start spooling up until about 12km, the rapiers a little higher. Climbing a little higher initially can let you get into a higher thrust envelope to begin with, meaning your dive doesn't need to be as large and you're not having to climb back out of it at an altitude where your engines aren't really working so well.

Really depends on the plane a little, too.. if you've got a high TWR thing you might not even need to dive at all..but if you can keep on climbing through to about 15-17km before dropping into the dive (need only be a few degrees), you'll probably find your speed picking up considerably as you climb and then zooming up through the dive.

I haven't really crunched the numbers though, but it feels better to me.. but like I said, it can vary from craft to craft.

[dronkit]

Hey great posts guys!

ANother question: squary intakes, shock cone intakes, fuselage intake, structural intake, pre-cooler... which are best? (say, for flying higher and faster and eventually reaching space)

Adding moar intakes, up to what point will it help?

Is it true that a RAPIER will be better than a turbo jet even within the atmosphere and in air-brathing mode?

In 0.9 I used to reach almost orbit or go around Kerbin by ballistic suborbital shots with just a turbo jet and one tank, now I can't pass 16k or 300 m/s

Edited June 7, 2015 by dronkit

[cybersol]

ANother question: squary intakes, shock cone intakes, fuselage intake, structural intake, pre-cooler... which are best? (say, for flying higher and faster and eventually reaching space)

Is it true that a RAPIER will be better than a turbo jet even within the atmosphere and in air-brathing mode?

In 0.9 I used to reach almost orbit or go around Kerbin by ballistic suborbital shots with just a turbo jet and one tank, now I can't pass 16k or 300 m/s

Squary (RAM), pre-coolers, and shock cones are all decent. You want enough intake area per engine for any of them though. 0.01 intake area for turbojets and 0.016 for rapiers seems about right for me.

Turbojets are better at punching through transonic with peak thrust near 13km, rapiers are better for top speed and ceiling with a thrust peak around 18km.

A turbojet with a single tank should easily break the sound barrier unless you have too much wing and drag. However, you'll still be about 1800 dv from orbit when they give out.

[ajburges]

An interesting thing I have realized is that due to the new areo system, having very swept back wings helps tremendously, even if they are tiny. I recently made a 8 kerbal SSTO where the only wings were the large wing strakes.

This is the classic case of how lift contributes to stability.

Lift surfaces behind CoM respond to slipstreaming with a force that produces torque to align the craft to prograde. Surfaces ahead of the CoM apply torque that would increase slipstreaming. The further a surface is the more pronounced the effect.

The CoL indicator shows the sum of lift values in the vertical. It provides an easy conceptualization of the sum of your lift effects. Too far back and you may lack control authority to leave a prograde force. Too far forward and you may lack the authority to maintain a prograde course. Even worse if CoL is ahead of CoM, the craft will always try and diverge from a prograde heading!

On the numbers...

Turbo jets have a fairly consistent curve there is no special region to break through the trans-sonic barrier on lower TWR craft. Peak power vs drag occurs near 16 km but the economics of fuel consumption means 17-18 km is a better acceleration altitude. Power drop vs speed occur at 1 km/s so you won't see much beyond that without over designing.

RAPIERS see optimal thrust vs drag at 20 km but there is another, local optimum near 8 km. Because they see greater Mach thrust scaling then turbojets, they tend to use the local optimum to break the sound barrier. They also see power drop off at 1.3 km/s.

Basic jets are abysmal in comparison. No increased thrust in the trans-sonic and only about double peak thrust (vs 6-9x peak) at 570 m/s. They also start final power decile at 8 km. Base TWR is also bottom of the pack.

Edited June 8, 2015 by ajburges

[Kicked to the Kerb]

Sorry for the delayed reply, I've been a little busy the last few days. I did however have a chance to test a few of my older designs with the recommended flight changes, and WOW what a difference it makes. As far as I can tell, the only real limiting factor on my previous attempts was stupidly ignoring how much effect the transonic would have on flight, which is actually a huge amount! On my first flight (in a single RAPIER craft making the dive at about 18km) I got a 72km apoapsis and a 56km periapsis (which of course is not quite a stable orbit due to licking the atmosphere and orbital degradation), and I wasn't really paying much attention to fuel consumption or angle of attack (apart from the dive), I was just experimenting with punching through the transonic. I'm sure if I actually try to improve my flight profile on the way up I'd easily make it to a stable orbit with fuel to spare.

Also, the way drag is now implemented makes a massive difference, as I had some vertical fuel tanks attached to one design (for VTOL purposes), and there's almost no way I can see that getting into orbit without severely overbuilding (and probably losing the VTOL capability as the plane would be too heavy to lift by then). So really, breaking the sound barrier seemed to be the main withholding factor in getting a Spaceplane into orbit. I am going to experiment more with newer builds though as this is what I gleaned from about an hour of testing 3 or 4 designs which is not enough for conclusive evidence.

FYI I don't use any mods at the moment (I used to but my game broke in one update and I have stayed with stock since) though I think some of them are amazing. (Just for those commenters who do use them and want to give accurate feedback.)

Lift surfaces behind CoM respond to slipstreaming with a force that produces torque to align the craft to prograde. Surfaces ahead of the CoM apply torque that would increase slipstreaming. The further a surface is the more pronounced the effect.

The CoL indicator shows the sum of lift values in the vertical. It provides an easy conceptualization of the sum of your lift effects. Too far back and you may lack control authority to leave a prograde force. Too far forward and you may lack the authority to maintain a prograde course. Even worse if CoL is ahead of CoM, the craft will always try and diverge from a prograde heading!

On the numbers...

Turbo jets have a fairly consistent curve there is no special region to break through the trans-sonic barrier on lower TWR craft. Peak power vs drag occurs near 16 km but the economics of fuel consumption means 17-18 km is a better acceleration altitude. Power drop vs speed occur at 1 km/s so you won't see much beyond that without over designing.

RAPIERS see optimal thrust vs drag at 20 km but there is another, local optimum near 8 km. Because they see greater Mach thrust scaling then turbojets, they tend to use the local optimum to break the sound barrier. They also see power drop off at 1.3 km/s.

Basic jets are abysmal in comparison. No increased thrust in the trans-sonic and only about double peak thrust (vs 6-9x peak) at 570 m/s. They also start final power decile at 8 km. Base TWR is also bottom of the pack.

Great info here. The CoM and CoL I feel comfortable with, but this is incredibly well explained and clear. The more specific info and numbers actually helped a lot!

Thanks again to you all, and I'll check back in once I've had a chance to space out.

KTTK

Edited June 9, 2015 by Kicked to the Kerb

[Kicked to the Kerb]

Before that I would "layer" them. What I mean by that is I'd have Deltas as the main wing, and for extra lift I'd take the small Deltas, attach them to the fuselage above and below the main Delta (One either side of the main wing) and angle them so they all met at the same point at the wingtips (not parallel) I guess you could visualise it like this : /|\ (only it would be on its side and the ends of the wings touch).

This was an EXTREMELY useful method of adding extra lift without compromising the width or length of the spacecraft, though I would imagine the new Aerodynamics would affect this setup, as the flow of air would be partially obstructed. Has anybody tried this method since the new model?

By the way has anybody tried said "layering" method I used to use since the new Aerodynamics? If so, what were the results?

KTTK

[ajburges]

By the way has anybody tried said "layering" method I used to use since the new Aerodynamics? If so, what were the results?

KTTK

That will still work in stock, but it probably won't help in FAR. NuStock occlusion rules are really simple. In summary: only attachment nodes and cargo bays occlude. Everything else is independently calculated and summed. That's why radial attachments make drag skyrocket.

As a consequence, all supersonic VTOL craft will want to use cargo/service bays to store inactive lift engines, land on the tail (research the Goblin), or incorporate a mode shift (there was a noval tilt moter design on in the craft exchange recently). The drag of any perpendicular stack is crazy.

I personally don't layer wings since I find the clipping distasteful in flight. Plus the struts needed for large compound wings add a ton of drag.

Hmm, there's a fun idea: hide wings in cargo bays that you can open for better subsonic lift. Breaks willing suspension of disbelief for me though.

[Kicked to the Kerb]

NuStock occlusion rules are really simple. In summary: only attachment nodes and cargo bays occlude. Everything else is independently calculated and summed. That's why radial attachments make drag skyrocket.

I saw a "Sir" Scott Manley video where he used attachment nodes to make a VTOL (structural hardpoints), but it was several versions ago, and every time I've tried It won't let me attach to one side of it. Did they fix it or am I missing something?

As for cargo bays, in that case they need a MK1 cargo bay to fill the gap imho I've used the MkII cargo bay for VTOL but not SSTO capable (yet) with the Aero update.

I hear what you're saying about suspension of disbelief, just to clarify though, I'm not talking about overlapping wings "inside" each other, just making them so that they come to a point at the end. Not sure if that was clear, it's hard to explain without visual reference (which I don't have privelages for yet)

Agreed about hiding wings in a cargo bay, that would be "cheating". IRL they wouldn't get the flow of air to make them aerodynamic in the 1st place.

Edit : Oh wait, surely it's just the foremost component(s) that drag is calculated for, as would be IRL?

Edited June 9, 2015 by Kicked to the Kerb

[ajburges]

I hear what you're saying about suspension of disbelief, just to clarify though, I'm not talking about overlapping wings "inside" each other, just making them so that they come to a point at the end. Not sure if that was clear, it's hard to explain without visual reference (which I don't have privelages for yet)

Agreed about hiding wings in a cargo bay, that would be "cheating". IRL they wouldn't get the flow of air to make them aerodynamic in the 1st place.

I find that even if you don't mean them to clip, that style of wing can still visually clip under certain load profiles.

Edit : Oh wait, surely it's just the foremost component(s) that drag is calculated for, as would be IRL?

You have drag on all exposed surfaces IRL. You get drag on the sides because no surface is frictionless and you get drag in the back from vortices. This is approximated in connection node occlusion. Only the non-oclluded "front" and "rear" portions of a part are added to drag; all parts experience full drag from "sides"

You can see this in the aero overlay if you clip the camera into an occluded part. It has drag vectors even on a prograde course.

[Mic_n]

only attachment nodes and cargo bays occlude.

Hrm... out of "loophole!" curiosity, do those 'attachment nodes' include the rocomax (and "not rocomax") hubs? If the physics just ignores "node attached" items for drag, then those 6-way adapters can mount four new stacks perpendicular to the existing stack orientation...

[Kicked to the Kerb]

You have drag on all exposed surfaces IRL. You get drag on the sides because no surface is frictionless and you get drag in the back from vortices. This is approximated in connection node occlusion. Only the non-oclluded "front" and "rear" portions of a part are added to drag; all parts experience full drag from "sides"

Yes this is what I meant. I didn't explain clearly, but by foremost I meant external surfaces, i.e. Not the "inside" edges of a part. My fault, I worded it badly.

Thanks for the posts ajburges, always very informative and desciptive!

KTTK

Edit : It just hit me, the word I was looking for was "outermost"!

Edited June 11, 2015 by Kicked to the Kerb