Spaceplane: high-speed deadlock, unrecoverable?

[Laie]

So, after using jets for a first stage on a rocket, I went one step further: trading jet engines for wings and a rolling start from the runway. My first spaceplane!

However, both with jet-powered rockets and spaceplanes, I tend to run into trouble at ~25km altitude. I need to point up and go up at all times. Nose at 45°, actual climb 10°. When the turbojets really kick in and I get faster, I have to work hard to keep the vessel pointed up. If I ever let the nose down for a moment, the drag will straighten my ship towards prograde, leading to less climb -> more speed -> more drag, pulling my nose down even further. The situation appears to be unrecoverable.

Depending on miniscule differences in my trajectory, sheer momentum may carry me another 8km higher after flameout, or merely 3km. There appears to be an ideal trajectory; hugging it as closely as possible will pay off big time, but crossing the line means that I won't go to space.

Does it have to be that way, or is this a design flaw? And is that situation really unrecoverable?

Edited May 16, 2014 by Laie

[Voculus]

If your plane wants to flip rear-end forward, it sounds like your center of mass is sitting forward of your center of lift. It could also be that your center of thrust is not lined up with your CoM. At high altitudes, control surfaces become much less useful, and any force applied off center, will sling the craft about that misalignment. Clear as mud?

[Laie]

Clear as mud?

Yes, but not relevant to my problem. The nose comes down, yes, but the plane does not flip. It only gets "straightened up" -- don't know how that's called. If I leave all controls and let nature have its way, it will turn into the wind like a weathervane. The faster I go, the more difficult it becomes to turn any other way.

[Rickenbacker]

Sounds like a good design then, if it automatically wants to minimize drag (i.e. point the way it's going). You might try to add some more thrust to it, if you can't get the climb rate you want.

[phoenix_ca]

http://forum.kerbalspaceprogram.com/threads/52080-Basic-Aircraft-Design-Explained-Simply-With-Pictures

Read it. See if you're doing something that's a bad idea.

Then come back if it doesn't help. With pictures. Pictures of the craft in SPH with CoM, CoL, and CoT indicators can help. Also knowing whether or not you're using FAR.

[Claw]

So, after using jets for a first stage on a rocket, I went one step further: trading jet engines for wings and a rolling start from the runway. My first spaceplane!

However, both with jet-powered rockets and spaceplanes, I tend to run into trouble at ~25km altitude. I need to point up and go up at all times. Nose at 45°, actual climb 10°. When the turbojets really kick in and I get faster, I have to work hard to keep the vessel pointed up. If I ever let the nose down for a moment, the drag will straighten my ship towards prograde, leading to less climb -> more speed -> more drag, pulling my nose down even further. The situation appears to be unrecoverable.

Depending on miniscule differences in my trajectory, sheer momentum may carry me another 8km higher after flameout, or merely 3km. There appears to be an ideal trajectory; hugging it as closely as possible will pay off big time, but crossing the line means that I won't go to space.

Does it have to be that way, or is this a design flaw? And is that situation really unrecoverable?

Well, it would help if we had a picture before I make a broad statement like "you has a design flaw." But it does sound like your plane might be missing something. Having 35 deg AoA can be caused by several things. It's really the fact that you have to have such a high AoA that's contributing to the "if I let my nose down for a moment, then drag will straighten my ship." It may also be a Center of Mass (CoM)/Center of Lift (CoL) relationship problem. Or it could be...(a whole bunch of other things).

So instead of listing all the possible causes which might make your plane act how you're describing, I'll list my basic design rules of thumb for space planes.

These are my basic design rules that act as a place to start. You can adjust each one to get the look and performance you want.

1) 1 TurboJet per 12 tons of craft. Up to a maximum of 15 tons. There's not really a minimum.

2) ~3 ram intakes per TurboJet. Minimum recommended is 1 per TurboJet. Maximum is as many as you want, but once you go past 3, you have to double the number of intakes to see much effect. (so 6, then 12, then 20+)

3) ~150ish units of liquid fuel per TurboJet. Minimum of 100 units (as you get more experience).

4) 0.5 to 1.0 lift rating per ton. I usually aim for around 0.75 per ton to start off. You can use a variety of wing parts to get there. Minimum is 0.5 units/ton. Recommended Max is 1.0 units/ton because after that you don't gain much benefit and suffer from drag.

The way these work is that if you lower the numbers in one rule, you should raise the numbers in another.

So for example, a 12t craft would be: 1 TurboJet, 3 Intakes, 150 fuel, 9 units of lift. You could also increase the number of intakes and reduce the lift and get 1 TurboJet, 6 Intakes, 150 fuel, 6 units of lift. Or if you wanted it to fly a bit further: 1 TurboJet, 3 Intakes, 200 fuel, 12 units of lift.

These rules of thumb don't have an exact exchange rate. (Such as 50 units of fuel requires 3 units of lift.) It's just a place to start your design, then gives you some guidance on how to adjust as you build.

Phoenix ninja'd me (because I'm slow this morning), but Keptin's article that Pheonix posted is also really good.

If you feel up to posting a picture, you will probably get more specific advice.

Good luck!!

[Laie]

1) 1 TurboJet per 12 tons of craft. Up to a maximum of 15 tons. There's not really a minimum.

2) etc.

My craft is a bit on the heavy side. I'm not at home (hence no pictures), but IIRC we're talking about one engine per 15t and ~0.45 lift-to-weight. It can barely take off, but at high altitudes it climbs well... has to, in fact, because of my weathervane problem.

Phoenix ninja'd me (because I'm slow this morning), but Keptin's article that Pheonix posted is also really good.

I had a good look at that before I started building, but it didn't occur to me to refer to it for troubleshooting. Thanks, both of you!

My center of lift is both behind and above the CoM, that might well be it. Maybe too far behind as well.

Putting this tentatively on "answered", at least until I can try.

[Taki117]

Generally, the way I fly spaceplanes (I use FAR so YMMV) is to ascend to 10km as fast as possible (Usually about a 45 deg pitch angle) then at 10km pitch down to about 20 degree pitch angle and accelerate until my time to apopsis drop by .1 sec (I use Mechjeb for orbital info) then kick in the rockets and shut down the turbo jets and begin pitching up (If I do it right I don't have to get beyond a 30 degree pitch angle, but 45 is not uncommon) and then raise my apoapsis to about 80km, watching my time to apoapsis and pointing prograde at 1 min 30sec to apoapsis (By this time I'm going about Mach 4-5 at an altitude of about 40-50km) then cut the engines and just coast to the edge of the atmosphere where I circularize.

If you are having issues with your aircrat pitching forward check your CoT vs CoM relationship, IT could be that your CoT is higher than your CoM and that is forcing your nose down, and at higher speeds your control surfaces can't compensate for that like they can at low speed.

[O-Doc]

One word: canards.

[zarakon]

If you are having issues with your aircrat pitching forward check your CoT vs CoM relationship, IT could be that your CoT is higher than your CoM and that is forcing your nose down, and at higher speeds your control surfaces can't compensate for that like they can at low speed.

That's what I would suspect. Either that or you just don't have enough thrust

[xcorps]

1) 1 TurboJet per 12 tons of craft. Up to a maximum of 15 tons. There's not really a minimum.

2) ~3 ram intakes per TurboJet. Minimum recommended is 1 per TurboJet. Maximum is as many as you want, but once you go past 3, you have to double the number of intakes to see much effect. (so 6, then 12, then 20+)

3) ~150ish units of liquid fuel per TurboJet. Minimum of 100 units (as you get more experience).

4) 0.5 to 1.0 lift rating per ton. I usually aim for around 0.75 per ton to start off. You can use a variety of wing parts to get there. Minimum is 0.5 units/ton. Recommended Max is 1.0 units/ton because after that you don't gain much benefit and suffer from drag.

The way these work is that if you lower the numbers in one rule, you should raise the numbers in another.

So for example, a 12t craft would be: 1 TurboJet, 3 Intakes, 150 fuel, 9 units of lift. You could also increase the number of intakes and reduce the lift and get 1 TurboJet, 6 Intakes, 150 fuel, 6 units of lift. Or if you wanted it to fly a bit further: 1 TurboJet, 3 Intakes, 200 fuel, 12 units of lift.

These rules of thumb don't have an exact exchange rate. (Such as 50 units of fuel requires 3 units of lift.) It's just a place to start your design, then gives you some guidance on how to adjust as you build.

RE rule 1: Are you saying that 12 tons should be the max weight on the entire design? Or is that just for the plane, payload not taken into account?

Maybe I'm thinking this all wrong. I'm playing with a design that has 12 tons of fuel parts, thinking I'll add a payload to it. Should I design the payload then the plane?

[Claw]

RE rule 1: Are you saying that 12 tons should be the max weight on the entire design? Or is that just for the plane, payload not taken into account?

Maybe I'm thinking this all wrong. I'm playing with a design that has 12 tons of fuel parts, thinking I'll add a payload to it. Should I design the payload then the plane?

By rule 1, the design all totaled up together, should be about 1 TurboJet per 12 tons. That includes fuel, wings, payload, etc.

What I will say is that for overall design philosophy, if you want to build a plane that carries payloads to space you have to decide how big the max payload will be. Then design your airplane around that. So if you want to carry 18 tons into space, you'll need at least 2 turbojets to start off with. After you add all your engines, wings, fuel, structure, etc, you would want it to weigh roughly 24 tons. If the design starts getting up to the 30 tons with all that stuff, then you need to think about adding an extra engine or increasing the fuel/wings.

I'm not sure if that's clear. Hopefully it answers your question.

[Redshift OTF]

Without pictures I would guess you don't have enough lift which means you have to angle the plane up a lot to compensate. Also you probably have the Centre of Lift too far back meaning you are nose heavy. You may have your control surfaces to close to your Centre of Mass meaning they are ineffective.

[Claw]

My craft is a bit on the heavy side. I'm not at home (hence no pictures), but IIRC we're talking about one engine per 15t and ~0.45 lift-to-weight. It can barely take off, but at high altitudes it climbs well... has to, in fact, because of my weathervane problem.

I had a good look at that before I started building, but it didn't occur to me to refer to it for troubleshooting. Thanks, both of you!

My center of lift is both behind and above the CoM, that might well be it. Maybe too far behind as well.

Putting this tentatively on "answered", at least until I can try.

There's nothing wrong with leaving it unanswered for a bit.

You do want the CoL above and behind, however... You typically only want it very slightly above. In line up/down is actually fine also. You don't want the CoL tooo far behind because your craft will "lawn dart." Meaning that it will constantly want to pitch nose down. I would say that based on your descriptions, this is not THE problem, but could contribute. Bad lawn dart behavior occurs just about throughout the flight envelope. So you would likely notice it on takeoff.

If your craft is at 15 tons and 0.45 lift/ton, then it's probably under powered/under lift. If you don't want to add an engine, you need to add some lift.

Also, I "generally" find that if you takeoff and your AoA is at 10 degrees or more down low, you'll probably have issues at altitude.

Generally, the way I fly spaceplanes (I use FAR so YMMV) is to ascend to 10km as fast as possible (Usually about a 45 deg pitch angle) then at 10km pitch down to about 20 degree pitch angle and accelerate until my time to apopsis drop by .1 sec (I use Mechjeb for orbital info) then kick in the rockets and shut down the turbo jets and begin pitching up (If I do it right I don't have to get beyond a 30 degree pitch angle, but 45 is not uncommon) and then raise my apoapsis to about 80km, watching my time to apoapsis and pointing prograde at 1 min 30sec to apoapsis (By this time I'm going about Mach 4-5 at an altitude of about 40-50km) then cut the engines and just coast to the edge of the atmosphere where I circularize.

This is pretty close to how I fly my profiles as well. A bit different, but it's a good basic profile.

If you are having issues with your aircrat pitching forward check your CoT vs CoM relationship, IT could be that your CoT is higher than your CoM and that is forcing your nose down, and at higher speeds your control surfaces can't compensate for that like they can at low speed.

At face value, I would disagree with this statement because in the OP, the description at altitude leads me to believe the plane is flying at 30+ AoA. You've already passed the point where CoT vs. CoM is the root cause. This is, of course, based on what I'm reading here but pictures are worth a thousand posts.

One word: canards.

I would also disagree with this as a "complete" fix. It will increase your lift/ton so it may help. Unfortunately, if adding canards doesn't immediately fix the problem it will make it much worse. That's because there's a bug with KSP that causes the lift on flight control surfaces to increase with AoA all the way to 90 degrees. This effectively can cause your CoL to shift forward at high AoA which causes the plane to flip backward and go out of control.

That's what I would suspect. Either that or you just don't have enough thrust

This would be one of my guesses based on what I'm hearing. Not enough thrust OR not enough wings.

All that being said, CoT vs CoM could also contribute, as can the placement of control surfaces. But in my experience, a plane that is at 35 AoA in the heart of the TurboJet envelope has other problems. Not to mention that at that AoA, you've really cut back on your air intake, further hurting your thrust production.

Like I said earlier, pictures or a craft file would help a ton.

[Laie]

There's nothing wrong with leaving it unanswered for a bit.

[...]

Yeah, but you all answered as best as you could given the lack of information, and I was out for the weekend. No point in leaving it an "open" question.

Here's a picture:

This is just at the start of trouble. Despite maximum pitch, the nose will slowly come down, eventually killing the climb rate. After a few more experiments, I can say with high confidence that the problem is lack of wings more than anything else. The flyer in the picture isn't quite the same vessel as the one that spawned this thread; here the wings are tilted by 10° while originally they were straight. Gave me a much better climb rate early on, but performed even worse at altitude.

That said, it is still possible to get into orbit by applying assymetric thrust. Switching off two of the upper engines will

1. allow me to point upwards,
   
2. restoring climb rate,
   
3. while still gaining speed
   

However, do not *ever* allow the climb rate to drop into the negative. Recovering even from a slightly negative climb will cost you as much time and fuel as the getting to 18000m in the first place.

I've since experimented with slightly smaller vessels. Starting with only 3.5 orange tanks worth of fuel (~160t rather than ~250t), I can still take as much fuel into orbit as with this big boy.

[Random Tank]

Yeah, I used to get this problem too; it's caused by 1 of 2 things: Lack of lift, or CoL too far behind the CoM

You can tell this straight away by if you have a small SSTO (25 tonnes or less) that won't get off the runway until you fly off the end; it will work after this fine, but when you get to higher altitudes it will be flying horizontal even when pitched up to 45 degrees, and will drop to horizontal as soon as you drop the controls.

With the Col distance, it follows the rule stated in keptin's guide; making the CoL closer to the CoM will increase agility but reduce stability (good for fighters), having it further away is vice versa (which is good for cruisey type planes and jetliners); a Spaceplane needs to compromise between the 2, so I usually find that having the CoL just touching the back of the CoM is pretty much perfect.

[SSSPutnik]

@25 km your engines are probably not generating much thrust due to lack of air?

Your speed for that altitude is woefully low.

[MarvinKitFox]

Despite maximum pitch, the nose will slowly come down, eventually killing the climb rate.

After a few more experiments, I can say with high confidence that the problem is lack of wings more than anything else.

That said, it is still possible to get into orbit by applying assymetric thrust. Switching off two of the upper engines will

1. allow me to point upwards,
   
2. restoring climb rate,
   
3. while still gaining speed
   

AHA!

If switching off two of your top engines actually improves the situation, then I can say with 100% confidence:

You problem is Thrust not aligning with Mass. Your engines are definitely above the CoM.

Possibly, this only shows once your fueltanks are partially emptied?

And no, pre-angling your wings is a very bad idea..

What I am seeing on your design:

1) not enough control surfaces. Not nearly. Either get some more control, or an industrial-sized SAS.

2) non-ideal wing choice. Those wings make for nice low-speed gliders, but aren't the best for hypersonic flight.

3) WAY too few air scoops, unless you are hiding a few dozen of them somewhere.

If you want to go orbital, you want thrust to last until >40km alt. That means 3+, prefer 6+ ramscoops *per engine*

4) Not visible in the design, but obvious from your description... Off-centre engines.

[O-Doc]

I stand corrected. Moar canards, moar elevons, more ailerons, moar rudders, moar wings, moar reaction wheels, moar intakes, moar jets!

Or,

Less payload...

[Laie]

AHA!

If switching off two of your top engines actually improves the situation, then I can say with 100% confidence:

You problem is Thrust not aligning with Mass. Your engines are definitely above the CoM.

How do you think I get to 15km without noticing asymmetric thrust?

http://ksp.schnobs.de/TankerII -- Notice CoM, COL, CoT.

Also notice how CoM moves while I drain the tanks. It's running on the big tanks alone; at launch, these should lack 30% oxidizer: the excess fuel is more than enough for the jets. Actual use on a typical flight is like 2/3rds of that -- so this is the amount you need to consider for atmospheric flight.

It's lagging badly on my i5. Depending on your rig, you may not want to take a ride. It really flys well at low altitude... but the 10° wing tilt make the stall in the critical phase even worse. I need to power down the first engines log before they're oxygen deprived.

As for intakes, they're hiding in plain sight. See the black area in the center? That's not a shadow.

@25 km your engines are probably not generating much thrust due to lack of air?

Your speed for that altitude is woefully low.

That's at 15, not 25km. If that matters? But more generally, all my designs (even the newer, lighter ones) seem to be constrained by terminal velocity. Trying to maximize the time I spend with all engines running at almost level flight will make me go slightly faster in the end, but the speed gain does not justify the fuel consumption. I find that the most efficient flight path in terms of total fuel consumed has me climbing at ~30-40m/s at 20km, then keep pointing in that direction until most engines have gone out. But my designs probably still have too little wing area: at altitude, flight turns into wing-assisted rocketeering.

Regarding the "newer, lighter ones":

This one worked reasonably well; I had to pump around fuel to move the COM a bit forward, though. Around 20km the pitch was again at maximum, yet if I wanted it to point further up it did so willingly. Turning by SAS maybe? The next iteration hat the CoM/CoL fixed, but I also clipped every fourth wing. As a result, I had the same attitude problem as with the big one, being pulled nose-forward at high speeds (or what I consider to be high speed). That's why I believe lack of wings to be the reason of that issue.

[allmhuran]

Insufficient lift is a possible reason, but one thing that hasn't been mentioned much in this thread is drag.

Generally, in order to fly straight and level, you have to have your nose pointed slightly upwards, giving your wings some angle of attack with which to provide lift.

Now, imagine you had a deployed parachute on the back of the plane. The parachute will want to be as far backwards as possible... which means it will be trying to pull the plane into a horizontal position. But now you'll start losing altitude, since you actually need to be angled up to fly level.

OK, so next let's look at the situation of trying to get suborbital on jets. You're at 30-odd km altitude, going super fast. You want to push higher, but your wings aren't providing much lift anymore. This means you need to increase your angle of attack... ie, you need your nose to be pitched up a bit.

Now let's imagine you have a whole bunch of engines and air intakes on the back of the jet. Parts with high mass have high drag, and air intakes also have high drag. So you have all of this drag affecting the back of the aircraft, much like the parachute example. This drag at the rear wants to level your aircraft out so that your velocity vector is aligned with your nose, in other words it will try to pull your nose down.

More wings helps because it means you don't need to have as high of an angle of attack, which means your nose is already closer to your velocity vector, which means that the impact of this drag is smaller.

You can also fix this by spreading air intakes out a little more. In the picture above, you have a huge amount of drag being generated by all of those intakes at the rear of the aircraft. If you took a couple and moved them to the front, then when pitched up those ones would want to pitch you up further, while the ones at the back try to pitch you down, evening out the drag forces and giving you better pitch authority from control inputs.

[Laie]

The intakes acting like the feathers on an arrow? Good point.

But after a few quick tests, I say "no worries". Four intakes per ton, placed on one side, will make a vessel uncontrollable; two intakes/ton will make themselves felt, one intake/ton is barely noticable. This beast is well beyond that.

With pWings (http://forum.kerbalspaceprogram.com/threads/29862-0-23-Procedural-Dynamics-Procedural-Wing-0-7) the wingspan now is over 3x the runway width, and it's still a biplane, but it works reasonably well. All lag was gone, too, until I slapped on even more intakes.

[allmhuran]

Where did you perform the tests though? The drag effect of intakes is very different at sea level/150m/s than it is at 30,000m / 2000 m/s.

[Random Tank]

More wings helps because it means you don't need to have as high of an angle of attack, which means your nose is already closer to your velocity vector, which means that the impact of this drag is smaller.

You can also fix this by just rotating your wings up a few degrees in the SPH to give you this angle of attack; this saves you having to pitch the plane up when wanting to fly level, which will not only bestow the above advantage, but will also not impair any additional drag.

With pWings (http://forum.kerbalspaceprogram.com/threads/29862-0-23-Procedural-Dynamics-Procedural-Wing-0-7) the wingspan now is over 3x the runway width, and it's still a biplane, but it works reasonably well. All lag was gone, too, until I slapped on even more intakes.

Have you tried making your pWings deeper? Delta style wings are much better at high altitudes then swept; as they have additional surface area.

[Laie]

Command Pod, mostly-drained tank, engine, intakes mounted on one side using cubic struts. Lauched straight up.

Things are worst at 12km, then it gets better again. However you need some RCS to even make it that far. Hence the command pod rather than a probe core. I figured some RCS to be the most minimalistic approach, control surfaces would probably have had a much stronger effect. Didn't even try the latter, though.