Time to learn space planes... What's wrong with mine?

[MitchS]

Hi everyone. I'm an experienced KSP player and understand most of the science behind the rocket-building aspects of the game, and have hundreds of happy hours landing all over the Kerbol system in stuff that rode tall pointy things to space. Spaceplanes, however, have always eluded me.

Well, I've decided to tackle that challenge, and want your expert advice on the way. I've watched some tutorials, noted design features in other people's successful ships, and have tried quite a few iterations without any success. I don't have the RAPIERs unlocked yet in this career save, but I kind of want to solve the SSTO challenge without them. Feels like I'll know the art better if I do it that way. 

I have a ship I'm happy with, and it's struggling to get sub-orbital. I can reach about 25km and 1,100m/s with my air breathers (why can't I get it higher or faster?), and the nukes just don't cut it to take it the rest of the way. I'm conscientious of effecient ascent profiles, but I may not be doing it right. 

I climb at about 45 degrees until around 10,000m, and then gently lower my nose to 10 degrees and ride the air breathers until they quit--when they reach about 5-10% thrust, I engage the nukes. 

Here's a picture of my spaceplane. Not all the fuel tanks are full, and it's aerodynamically sound both wet and dry. Help me, oh mighty wizards of the SSTO world.

(also, how do I embed imgur pictures into a post...?) http://imgur.com/FaQggff

[edit by Snark: put image inline]

Edited December 15, 2016 by Snark
Fix imgur link to be inline image

[Firemetal]

Ok. Switch out those whiplashes for rapiers and add a rocket fuselage instead of the pre-coolers. The shock cones already give way too much air so you don't need those. The Nukes on their own cannot get you to an orbit. You need to use the rapiers rocket mode to raise your AP and Velocity sufficiently. So keep adding more oxidizer until that happens. Apart from that, looks good!

Fire

[bewing]

Heh! I disagree with Firemetal completely. You are doing the right thing by learning on whiplashes, the nukes are fine, and this is a very cute spaceplane that will certainly get you to orbit with a few tweaks.

Several points, in no particular order:

The shock cones are great. But if you have them you don't need the precoolers. Replace the precoolers with mk1 liquid fuel tanks.

MK2 parts have very high drag compared to MK1 parts. So this guy is going to go slower than an equivalent MK1 design, which makes things easier on reentry. But that is why you max out at 1100 m/s airbreathing.

Whiplashes become deadweight as soon as they go above 25km. That is their max operating altitude, period. So it's often a good idea to minimize the count -- provided the reduced number is still enough to get you supersonic.

If you climb too steeply, then your engines are doing most of the work. Part of the point of having wings is to let them do their share of the lifting -- especially low in the atmosphere. You don't need to open up the throttle until you hit 5km, minimum. Before that, climb at 20 m/s and take your time.

Which brings me to the point of saying that you probably need more wing. The BigS wings carry fuel (they are free fuel tanks!) and give the same lift as other wings.

Spaceplanes take a good long time to get to orbit. Don't rush it. Be patient.

You should completely level out at about 16km for your "speed run". Then let the nose climb by itself with SAS.

You can light the nukes at 18km to 20km, along with your whiplashes. That will give you a good shove in that region.

The nukes only really start working nicely around 29km, so you kinda need to "hop" from 20km to 29km. But then you need to mostly level out again in an MK2 design. (An MK1 design needs to get to 42km before leveling out because of heat.) Just so long as you are accelerating and climbing above 29km, you're going to make it -- providied you brought along enough fuel.

 

PS. imgur: when you have the image posted, right click on the pic and say "open image in new tab" or whatever. This will give you another link that starts i.imgur.com -- that one is embeddable.

 

 

Edited December 15, 2016 by bewing

[MitchS]

Great! Thank you for the guidance. I really appreciate how patient SSTO guys are with the rest of us. 

I switched the pre-coolers for fuel tanks, and topped off all the other tanks on the ship. Is that too much? Should I trim some weight SOMEWHERE? Intuitively, I'd say yes, but I'm out of my expertise here.

I followed the ascent you recommended--let the wings do most of the work, leveled off at 16km (didn't get much faster than before though... 1,200m/s or so) and then punched the nukes at 20km. I tried to let the nose rise, but it never got above 5 degrees or so... I seem to have very little pitch authority above the atmosphere, too, and couldn't give any effective nose-up input without aerodynamic instability--uncommanded yaw and extreme drag from a very high AoA. Nukes don't vector, so I'm not sure how to combat this. I was stuck within 5s of my Ap at 1,500m/s and 30km, and with my speed creeping up very slowly, and then I began to descend on that trajectory and started to slow down from the thicker atmosphere's drag below 25km. 

I don't have FAR installed, by the way. You know, 1.2 and all. That said, do I need to consider area ruling at all?

Still no dice. Here's a screenshot of my edited ship. 

Edited December 15, 2016 by MitchS
edit: More questions. :(

[bewing]

46 minutes ago, MitchS said:

Still no dice. Here's a screenshot of my edited ship. 

Maybe a couple of advanced reaction wheels in your nacelles. And perhaps trade the hypersonic canards for tailfins -- that'll give you more control authority in the atmosphere and above it.

Note: You probably want to maintain an AoA of about 10 degrees when you are above 29km. You choose the AoA to keep your prograde marker above the horizon.

As far as fuel goes: the more fuel the better, really -- because once you are in orbit, you'll want to go cruising around.

-- Your plane looks like it's right on the edge of being a fully functional design. 1200 m/s on the whiplashes is a pretty good number, BTW. With the rapiers you'd already be in orbit ... that's how close you are.

 

Edited December 15, 2016 by bewing

[KevinW42]

If you're having problems with yaw it's probably because of how close your CoM is to the rear of the plane. It might help with stability to mount some wing strakes vertically with the rear end extending past the trailing edge of the wing.

[AeroGav]

6 hours ago, MitchS said:

Great! Thank you for the guidance. I really appreciate how patient SSTO guys are with the rest of us. 

Hi Mitch, 

you're making life harder for yourself with the mk2 fuselage.   It creates three times the drag of mk1.  They can withstand more heat, but you're not able to take advantage  because you are using the pointy cockpit , which suffer really bad heating from being right at the front.

A mk1 inline cockpit is actually less prone to overheating than a mk2 pointy one.

An oxidizer-free SSTO such as you are creating needs a good lift:drag ratio - ie. plenty of wing area, low trim drag and clean lines.

Best lift :drag ratio is found with the nose at 5 degrees above prograde or just above.  

At 1200 m/s you're still some way short of orbital velocity, and orbital freefall is only supporting 20-30% of your plane's weight.  Lift needs to make up the difference or you fall back down.    Good lift:drag ratio means you can get the lift you need without the drag overwhelming the  limited output of the nuke engines.     Plenty of wing area means you don't need to raise the nose more than 5 degrees above prograde, where things get draggy.   And it also means you climb quickly into thinner air , where that barn-door fuselage makes less drag.     As you get closer to orbital velocity, the lift requirement will decrease.   

However,  you can see from the above graph the Whiplash looses power quickly as you start to exceed 900m/s.  By 1150 m/s it's only making half as much thrust as it was at 900.    And when you go above 17km, power falls off quicker than drag too.  I normally limit myself to one whiplash per 30 tons, because bringing more doesn't allow you to go much faster and the dead weight (1.8 tons) hurts your delta V.              

Also, it sounds like you have some pitch stability issues when the fuel is burned off.   You have  9.6 tons of engines at the back , and a 2.3 ton cockpit at the front.  Most of the fuel is at the front, which helps to balance things, until it's gone.

What i'd do 

1) have a single whiplash at the rear fuselage

2) swap to an inline mk2 cockpit

3) mount the two nukes at the sides of the fuselage, then empty all the fuel tanks and check your CG.  Slide the nukes forward until the plane is balanced with no fuel.

4) add the fuel back.  Because it is no longer as tail heavy, you might need to remove some fuel from the front of the ship/add some towards the back

5) add even more wing area and enough canard to hold 5 degree nose up easily.

6) consider putting cones on your nuke engines for lower drag like i do in this vid.  It won't make a huge difference given the huge mk2 fuselage drag but it's a fairly simple mod.

 

steps 3-4 are easier if you use the RCS Build Aid mod, which gives you a dry CoM marker.

 

 

Edited December 15, 2016 by AeroGav

[GoSlash27]

MitchS,

 I'd recommend ditching one of the Whiplashes. All that thrust is just adding weight and burning fuel faster. It won't help you achieve a higher top speed in air breathing mode.

One whiplash and 2 LV-Ns should suffice,

Best,
-Slashy

 

 

[Clipperride]

Great thread!  I to have been having space plane issues since the 1.2 update and I'm fascinated by the above helpful advice.  It's inspired me to have another go at building a passenger carrier.

If I could hijack the thread a little, how does the change in the speed of Mach 3 with increasing altitude effect the Whiplash's power curve?

Edited December 15, 2016 by Clipperride

[lodger]

11 hours ago, MitchS said:

b at about 45 degrees until around 10,000m, and then gently lower my nose to 10 degrees and ride the air breathers until they quit--when they reach about 5-10% thrust, I engage the nukes. 
 

Ty to ascent with an AoA not higher than 10 degrees. I had the problem of getting into a higher orbit with a similar design and the problem was that during ascend I used way too much pitch for a Mk2 based design. As @bewing pointed out in this post, Mk2 and Mk3 designs should be flown close to prograde, 10 degree AoA at most.

Edited December 15, 2016 by lodger

[AeroGav]

@MitchS

OK,   I tried to recreate your ship with my suggested changes.   As always, the hardest thing for me is getting it to balance with empty fuel tanks.    Using mk1 liquid fuel tanks for the nuke nacelles gave me too much fuel capacity up front,  making it nose heavy when full/tail heavy when dry,  so i went back to pre-coolers.     And it turns out the subsonic circular intake is actually a good low drag nosecone, though they gather very little air at supersonic speeds.  The heat tolerance of them is not great but they survive.   Could have used slanted nose cone instead for the nukes but i prefer the look i have here.

http://pastebin.com/x3t3MFi9

(to download this craft, click GET on the above link, then rename it from a .TXT to a .CRAFT file and put in your SPH folder)

flight profile -

Subsonic climb to 10km, when speed starts to exceed 250/260 , press prograde and accelerate through mach 1 while nosing over into a shallow dive.  At 400 m/s , remove prograde / SAS and allow it to pitch up naturally into a climb.    When the climb angle reaches 15 degrees,  stop the nose rising any further by locking Stability Assist.        At 14.5km,  use prograde again to level off.   Fly level at 17km for your speedrun.    Should be able to get nearly 1100 m/s.        When the acceleration starts to fade,  press SPACE to start the nukes and maintain nose at 5 degrees above prograde until orbit.

To me, that's a disappointing level of performance as it hasn't got interplanetary delta v nor does it have any passengers or cargo.  To get more, you need to angle the wings so the fuselage flies at 0 AoA,  use endgame techs like big S wings and strakes, and RAPIERs.   Also, using mk2 fuselage parts for fuel tanks is a terrible idea, especially when like here the oxidizer tankage goes unused.    Mk1 gets you more capacity for same drag.    Far better if you can keep all your fuel in big S wings, strakes or mk1 tanks, keep the mk2 stuff for Kerbals to sit in and for cargo.

sandbox interplanetary mk2 with passengers , docking adapter (has a small cargo too, but mostly occupied with reaction wheels etc).

https://kerbalx.com/AeroGav/ASES-SP2-Xkos-Dual-Nukes

 career mk2 i built with same tech level as your ship - has  2 x whiplash, 2 x nuke and 1 x aerospike - about 3500 dv in orbit plus two nice cargo bays

https://kerbalx.com/AeroGav/X949-Xanadu-B

51 minutes ago, Clipperride said:

 

If I could hijack the thread a little, how does the change in the speed of Mach 3 with increasing altitude effect the Whiplash's power curve?

The config file has a table of mach numbers vs thrust multipliers,   so i guess that peak thrust speed (mach 3) is a lower absolute velocity at altitude than at sea level.   Bear in mind that up to 17km, air density falls more quickly than engine power on the whiplash engine, so provided you got enough lift to keep AoA below 5 degrees, it's much easier to reach mach 3 at 17km than down on the deck.     Much less heat, risk of overstress, fuel wasted to drag  too.

[GoSlash27]

27 minutes ago, lodger said:

Ty to ascent with an AoA not higher than 10 degrees. I had the problem of getting into a higher orbit with a similar design and the problem was that during ascend I used way too much pitch for a Mk2 based design. As @bewing pointed out in this post, Mk2 and Mk3 designs should be flown close to prograde, 10 degree AoA at most.

To this end... You should have some static incidence in the wings (the leading edge higher than the trailing edge) so that they will produce lift without any AoA whatsoever. This presents minimal parasitic drag.

Best,
-Slashy

[zolotiyeruki]

7 minutes ago, GoSlash27 said:

To this end... You should have some static incidence in the wings (the leading edge higher than the trailing edge) so that they will produce lift without any AoA whatsoever. This presents minimal parasitic drag.

This, a thousand times this.  Any high-speed, high-altitude atmosphere craft I build has this, and it makes a HUGE difference because of how the KSP drag model works.  Give your wings a 5 degree upward 

Unless it has recently changed, KSP's drag model for wings includes only lift-induced drag.  By putting built-in AoI on your wings, you can point your nose directly prograde and minimize the parasitic drag from your fuselage, while still getting plenty of lift.  With the way KSP's aero works, a 5 degree AoI on a 15-meter-long fuselage will *double* the drag.

[AeroGav]

1 hour ago, zolotiyeruki said:

 

Unless it has recently changed, KSP's drag model for wings includes only lift-induced drag.  By putting built-in AoI on your wings, you can point your nose directly prograde and minimize the parasitic drag from your fuselage, while still getting plenty of lift.  With the way KSP's aero works, a 5 degree AoI on a 15-meter-long fuselage will *double* the drag.

I haven't been pushing him to do this because i regard it as an "advanced" technique, someone else i was trying to help couldn't balance their CoL/CoM with angled wings, though to be fair Mitch is starting from a higher level so he'd probably be able to handle it.

I did some tests with wing segments.    They do have some parasite drag, and the optimum lift drag ratio for the wing panel itself occurs at around 7 degrees.   However, the peak L/D ratio of 9.5 to 1  at mach 1.7 is about three times better than anything i've seen on a mk2 spaceplane, which means that most of the drag must be coming from the fuselage  even when the nose is pointing directly prograde and despite the fact that i tend to use 2 or 3 times as much wing on my spaceplanes as other folks do .

[MitchS]

Thank you, everyone!

I'm actually a military pilot by trade, so increasing the angle of incidence makes a lot of sense to me--the only reason I haven't tried it on my SSTOs is because I wasn't sure KSP modeled that correctly. What a relief. If there are any other "advanced techniques" you think my SSTOs would benefit from, let me know! I just don't know how the KSP aero model really actually works (magic, I assume) so I don't know which real-life engineering solutions behave correctly in the game.

On that note, area ruling only matters with FAR, correct? I know stock aero works in part-based magic, and FAR aero works in craft-based magic, but that's as far as my understanding goes. Any enlightenment folks could pass along for getting the most out of each model would be appreciated.

I'll make those modifications when I get home this evening, and I'll update you all on the results.

Edited December 15, 2016 by MitchS
edit: grammar

[GoSlash27]

MitchS,

 Stock KSP doesn't deal with area ruling or turbulence. It's just plain drag coefficient and occlusion. Greatly oversimplified explanation, but if it looks clean it is clean.

 The principle that I always stress in spaceplane design is to focus on reducing drag instead of adding thrust.

Another tip that Isn't "advanced", but many aren't aware of: Any radially attached parts expose their full frontal area even if they're clipped into something else. Try to keep the "greebling" to a minimum and don't attach stuff you don't actually need.

Best,
-Slashy

 

[Palaceviking]

27 minutes ago, MitchS said:

Thank you, everyone!

I'm actually a military pilot by trade, so increasing the angle of incidence makes a lot of sense to me--the only reason I haven't tried it on my SSTOs is because I wasn't sure KSP modeled that correctly. What a relief. If there are any other "advanced techniques" you think my SSTOs would benefit from, let me know! I just don't know how the KSP aero model really actually works (magic, I assume) so I don't know which real-life engineering solutions behave correctly in the game.

On that note, area ruling only matters with FAR, correct? I know stock aero works in part-based magic, and FAR aero works in craft-based magic, but that's as far as my understanding goes. Any enlightenment folks could pass along for getting the most out of each model would be appreciated.

I'll make those modifications when I get home this evening, and I'll update you all on the results.

I'm quite impressed by the ksp aero model and after a  lot  of faffing with aircraft I've seen loads of real life techniques work really well, if you think it should work-try it

[foamyesque]

5 hours ago, GoSlash27 said:

MitchS,

 Stock KSP doesn't deal with area ruling or turbulence. It's just plain drag coefficient and occlusion. Greatly oversimplified explanation, but if it looks clean it is clean.

 The principle that I always stress in spaceplane design is to focus on reducing drag instead of adding thrust.

Another tip that Isn't "advanced", but many aren't aware of: Any radially attached parts expose their full frontal area even if they're clipped into something else. Try to keep the "greebling" to a minimum and don't attach stuff you don't actually need.

Best,
-Slashy

 

It's also important to note that orientation matters for radial items, something I discovered in the course of setting up RCS units. If you must mount something radially, mount it edge-on to the airflow.

The priorities I look for in designing a spaceplane are, in order:

1. Minimizing longitudinal CoM shift;

2. Enough wing area to support an unloaded gliding landing at 50-60m/s (this is often surprisingly little);

3. Minimizing uncommanded pitch (torque from offset engines or a CoL of the main wing (especially when angled) fore or aft of the CoL);

4. Minimizing drag from widgets.

 

It also helps if your plane is designed with relaxed stability, because then you need to make minimal control corrections to keep your pitch trimmed correctly, which significantly reduces your drag and, if your have gimballing turned on on your airbreathers, your cosine losses.

As far as controllability in the upper atmosphere goes, you'll want more SAS torque. RCS is also an option for really large machines, but for something this small the additional drag can be really nasty. I don't much like how magical KSP's reaction wheels are, but you may as well take advantage of them; inline stabilizers on the nacelles would probably fix your issues there. A Mk2 drone core is also an option and would push your cockpit further forward to reduce your CoM shift.

Also, speaking of the CoM shift, you want those nukes up in the midbody area; they're big and heavy, and at the back they're hurting your yaw stability and going to cause your plane to become progressively less stable as fuel burns away. The whiplashes are lighter and have gimballing; put them at the back if you're going to put anything there.

I prefer the standard canards over the advanced canards, in general; the advanced canards are somewhat quicker to respond, but the standard ones have both more wing area to work with (hence more torque for the same deflection, as well as a bigger shift in your CoL position, making that torque more effective) and have a wider range, which makes control through high-angle re-entry flight better. Another possible option are naked Big-S elevons, which have the highest deflection range and response rate of any of the control surfaces, and even more control area, but they look kind of silly. The tailfins (any of them) are not good choices for canards IMO, despite @bewing's advice; for canards I want the maximum possible control surface to fixed wing area I can get, so they have the least impact on my overall CoL controls-neutral and give me the most possible variation controls-engaged.

Edited December 15, 2016 by foamyesque

[MitchS]

Hey everyone! I've had my first success! Here's the album for the screenshots. (getting fancy with imgur...) http://imgur.com/gallery/fvFYs

I don't think I need two shock cones now that I have one whiplash, so I'm going to substitute those for the tapered Mk1 LF tanks, and cap them off with small nosecones. I'll put a pre-cooler in front of the whiplash for my intake air. Is that enough? I also want to add a bit more wing, and that longitudinal stability I was talking about because it doesn't behave how I'd like during and after reentry. I was sure to balance the fuel to give myself a good CoM/CoL position for reentry, but I got irritated with the flapping around when I got down to flying speeds at 20km. 

Thank you everyone SO much for your help!! I'm interested in optimizing this ship a little more, then giving it a small cargo bay and payload. Then I'm going to move on to the RAPIERs and more substantial payloads. As for the optimization on this design, what input do you have? Did I fly it correctly? What sticks out as inefficient to you? The v4 redesign is almost done. Screenshots in a second.

[zolotiyeruki]

When it comes to getting CoM/CoL right, here's something I do:  build the plane, then empty all fuel tanks.  Balance the plane in this state!  Then add the fuel back in.  Usually, adding fuel will push the CoM forward.  If it's too much of a shift, then rearrange things and try the same procedure again.  If the CoM moves less than 5% of the craft's length, that's good enough for me.  You can also set the fuel flow priority so that the plane comes into balance as it burns fuel.

[MitchS]

v4. The "changelog" is in the picture description.

[MitchS]

http://imgur.com/a/yeWI8

v4 was an improvement! I really liked the new canard design (Thanks @foamyesque) and the added wing area really helped. I've never seen my apoapsis increasing while my throttle is at idle cruising to space. Hahaha. Lift!

I'm going to start modifying it to have a small cargo bay to use some of that extra delta v that I have left over. Stand by for more questions. 

[bewing]

Congrats. Nicely done.

[MitchS]

Mk. 2 is largely the same, just with a cargo bay up front containing a simple 2t probe.

- Added payload bay
- Rebalanced some fuel--the precooler is further forward than the mk1 fuel tank to shift CoM further back 
- Slid the whole wing assemblies forward to offset the mass of the payload bay
- Adjusted canards
- Moved landing gear closer to new CoM... rotation speed was nearly 100m/s in the old position



The CoM is much harder to manage here. With the payload, it's too far forward, and SAS struggles to keep my desired attitude. Without the payload and low on fuel, the CoM slides way too close to CoL and I worry that it's not going to be aerodynamically stable. Is the solution here to just manually balance fuel before reentering? What about scooting the runway CoM back a little? This was a hard "happy medium" for me to find.

Also, why can't I alt+WASD trim anymore? Did that keybind get relocated for 1.2?

Edited December 16, 2016 by MitchS

[foamyesque]

Payload bays should, ideally, be on your CoM, to eliminate precisely that issue.