Spaceplane Orbital Insertion Spin of DOOM!!!

[ThatSpacePlaneGuy]

So I'm in the final phases of testing a heavy-lift SSTO spaceplane (it takes off without disintigrating, flies reasonably well, and has awesome enough TWR to get to orbit...theoretically). Once I hit 1700 m/s at around 20km up, I adjust my ascent trajectory with Mechjeb to point the nose about 20-25 degrees up.

The plane pitches up VERY violently and spins out of control, well before the engines flame out.

When flying level, my CoL is slightly behind my CoM, so it's nice and stable. However, when pitching up about 20-25 degrees, the CoL moves forward, which causes the sudden and violent orbital insertion spin of doom. At more normal speeds and altitudes, Mechjeb can use the canards and elevators to compensate without much trouble.

How do I prevent/reduce the CoL's forward movement as the plane pitches up?

Wings are large and delta-ish, for low supersonic drag and plenty of room between the CoM and elevators. I would really rather not decrease wing size, as that would severely limit the amount of stuff the ship can carry. (Design goal is 1 orange tank and able to take off before the runway's end, and no stalling issues when transiting from low-altitude sharp ascent to supercruise.)

EDIT:

Canards removed & wings resculpted. Screenshots:

Notice here, everything is great.

When pitched up...not so much.

Note that the fins near the front of the fuselage are mapped to provide yaw authority, not pitch. Without them, the plane has some yaw-stability problems.

2ND EDIT:

Problem has been fixed. The spaceplane lacked reaction wheels (I thought they were installed), and suffered from a relative air intake deficit (less than 1:1) which limited maximum air-breathing altitude and forced the orbital insertion window much lower than normal. Both design flaws have been fixed, and the aircraft now meets or exceeds all of my design criteria.

Edited December 4, 2013 by ThatSpacePlaneGuy

[Sirrobert]

Try pitching up slower? If you suddenly pitch up your CoL is going to be below the CoM (since you are still going forward, rather than up)

[700NitroXpress]

Your center of mass is off because you're burning fuel, always burn from the rear instead of the front and it'll solve this problem. Lock the tanks by right clicking on them. Also, fly it without Mech Jeb.

[CaptinKornflake]

So what your saying is that when you barely pitch up at all it keeps going up and spins out? Don't worry that happens! You just have too much surface control in the front. If you could post a picture of your plane that would help, but if you have Canards or Winglets on the nose of your plane you may want to replace these with wings that don't control but give lift, having your control surfaces in the back will keep it flying forward like an arrow because of the small amounts of drag, but like an arrow adding feathers to the front will make it want to fly backwards and it won't fly straight. Hope I helped!

[peadar1987]

If you're using FAR, tilt your forward lifting surfaces upwards so they stall before the main ones, dropping the nose if you begin to flip, and (hopefully) saving the situation.

[Murph]

MechJeb can be very harsh with its control inputs if you give it a significant course change by punching numbers into Smart A.S.S. It also has a habit of oversteering sometimes, so your 25º change might temporarily go to, for example, 45º, before settling back onto the requested course. As far as I know, it has either no or very little knowledge of aerodynamic stability, so will just jam the controls hard over for large changes, even if that means a guaranteed loss of control incident. A spin out of the type you mention tends to become likely when the heading strays too far away from the current prograde vector. So, if you do your change more gradually, e.g. in 5º or 10º increments, allowing time for the prograde vector to catch up with the new heading, the problem may well go away.

It does very much depend on the design of the plane, however. Some designs can cope with the very harsh handling, others will spin out very easily.

[numerobis]

MechJeb ignores aerodynamic stability, and it sometimes commands the oddest trajectories to get to your target heading (you want to yaw left by 45 degrees? Let's pitch and roll and yaw right, then swing around, and finally end up there... what's that about the tail getting torn off?)

[ThatSpacePlaneGuy]

Mechjeb ascent guidance generally does a good job of getting the spaceplane to go where it needs to, if one uses the slider to adjust the ascent angle. It's infinitely more steady and precise than trying to use a keyboard to fly. Occasionally it requires some manual input to remove unwanted roll, but this particular spaceplane only requires a couple periodic Q or E taps during ascent. It's remarkably roll-stable.

Pitching up slow & steady works, but only up to a point. No matter how slowly I raise the nose, it still spins well before I get to prograde-20-degrees-up.

[Kasuha]

This is just a guess ... but I believe the problem appears when your nose is too way off your prograde indicator. And that would likely happen even in level flight. Your goal is not to pitch your plane's nose 20 degrees up, your goal is to pitch your trajectory 20 degrees up, and you need to guide your plane there. If it starts to spin, you just need to do it slower.

[Murph]

Well, if slow and steady doesn't work, it seems like there's only 2 ways to fix it. 1) A more gentle ascent from 20km to 32km (where the atmosphere gets so thin that it should hopefully no longer cause a spin). 2) Find a way to move the mass forward, lift back, or a combination of both. I can't see from your pictures where the fuel tanks are, and what sizes they are, but careful movement of fuel could well get the mass forward for you. Also, if those are turbojets on the top of the tail, you could move them forward, as their exhaust doesn't damage things like rocket exhausts do.

Another thing which looks a little odd, but it could just be the angles of the photos. If those are the B9 SABRE intakes on the group of 4 nacelles, I believe the angle on them is supposed to be pointing downwards, so that they are facing more into the airflow when you have a positive angle of attack. They might be fine, it just looks like some of them might be pointing in odd directions, but hard to tell.

How heavy is this thing? If you've got SABREs on there, I'm slightly surprised at having turbojets as well, and levelling out at 20km. I thought the SABREs generally kicked the turbojets out for both power and fuel efficiency. A couple of planes I've been working on use SABREs, the smaller one taking about 80–100t up to 100km orbit on a pair of SABRE Ms, and the larger one taking 200t up on quad SABRE Ms. Every plane and pilot has their preferred altitude for gaining speed, but 20km seems maybe a little low to me, for SABREs  I tend to level out at about 25km to gain speed. Perhaps gaining your speed at 25km might help with the spin out issue, due to the thinner air there?

[700NitroXpress]

Add mass to the front and your problems are solved.

[ThatSpacePlaneGuy]

Well, I finally discovered one particularly massive issue: The plane COMPLETELY lacked reaction wheels (aside from those in the cockpit, which are woefully inadequate for a plane of this size). I thought I installed a cluster of three heavy-fuselage reaction wheels near the back of the ship, but apparently that was a different version. I shuffled the fuselage parts a bit to add in three reaction wheel pieces without altering CoM/CoL relation.

Another issue (as pointed out by Murph) was my orbital transfer altitude: It was far too low. I was getting flameout-prevention thrust loss at around 20 km - where the air is too thick to execute the nose-up maneuver without losing control. As it turns out, the centerline Sabre had NO air intakes whatsoever, and the two inboard ramjets also had no air intakes - so I had a HUGE air intake deficit. I added several radial-mount air intakes, and the plane can now begin orbital insertion at ~25 km, and burn air-breathing ramjets all the way up to ~55 km.

The combination of higher orbital insertion altitude with extra intakes and reaction wheels for extra rotational authority has fixed the problem. The spaceplane is now quite a pleasure to fly.

To answer your question, those other engines are not turbojets - they are ramjets, from the mod "Taviero's Pizza and Aerospace". They perform quite well at high speed, and similar to Sabre engines, their thrust does not taper sharply to zero as you approach a vaccuum. Sequentially switching Sabres from air-breathing to rocket mode as altitude increases conserves a lot of rocket fuel during orbital insertion, and ramjets give statistically significant thrust almost all the way up to main engine cutoff.

The spaceplane weighs 184 tons fully fueled, with no cargo. Carrying an orange tank, it weighs 218 tons. (975 delta-V remaining upon establishing circularized 300 km orbit, before consuming any of the orange tank's fuel.) Since the cargo bay is right on the CoM, the flight profile of the spaceplane is NOT altered by hauling an orange tank. The thrust-to-weight ratio just goes down a bit.

In fact, thanks to careful design and shuffling parts around on the airframe, the CoM when fuel-full and CoM when fuel-empty are almost identical, so the spaceplane doesn't even fly differently upon re-entry - just lighter.

Edited December 4, 2013 by ThatSpacePlaneGuy