Why the Starship needs a bellyflop maneuver?

Why not to return vertically like Falcon 9?

The term is used in articles like this one:

During such a flight, experts believe SN8 will perform an unusual “belly-flop” maneuver once it begins to fall out of the sky, using its new aerodynamic surfaces to slow its descent in a horizontal position

  • $\begingroup$ What is a "bellyflop maneuver"? The question will be more interesting and helpful to future readers of the site if obscure terms are explained. Thanks! $\endgroup$
    – uhoh
    Oct 31, 2020 at 23:11
  • $\begingroup$ I wonder if some of this is related: What would the BFS' landing on Mars be like, if it is different from this animation? $\endgroup$
    – uhoh
    Oct 31, 2020 at 23:17
  • $\begingroup$ The term is used in articles like this and I think it is self explanatory - thomasnet.com/insights/… $\endgroup$
    – Joe Jobs
    Nov 1, 2020 at 0:36
  • 2
    $\begingroup$ I don't know if you can see them, but since there are already three votes to close in six hours and only two more will close it, you might consider taking a moment and adjust your post accordingly. "Needs details or clarity (3) This question should include more details and clarify the problem." i.stack.imgur.com/OYcvc.png Many people will vote to close before ever reading down through all the comments, and comments are considered temporary and can be deleted at any time. The only thing that really counts is the content of the post itself. $\endgroup$
    – uhoh
    Nov 1, 2020 at 0:42
  • $\begingroup$ Thanks for the edit! Close vote and down vote retracted and +1 $\endgroup$
    – uhoh
    Nov 1, 2020 at 4:07

2 Answers 2


Why not to return vertically like Falcon 9?

Falcon 9's first stage does not return from orbital velocity, since it never goes to orbit. Starship is muuuuuuuuuuuuuch faster than Falcon 9's first stage. Starship not only reenters from orbital velocity, but from interplanetary velocities as well.

The only sensible way of shedding that much energy is to aerobrake. Aerobraking is free, whereas propulsive braking needs the same amount of energy as accelerating in the first place. That means you have to carry enough fuel for braking with you, which means you need to carry extra fuel to accelerate that fuel during launch, and extra fuel to accelerate the extra fuel, and extra fuel to accelerate the extra fuel for accelerating the extra fuel, and so on. And because you need to carry fuel for braking, your spaceship will be heavier than it would be when you aerobrake, so you need to shed even more energy, which means you need even more fuel … etc.

So, as soon as we have decided to use aerobraking, it becomes immediately obvious that you want to point the side with the most wind resistance into the flight path, and that is the broad side of the ship.

Note that this is an oversimplification. What we are really doing is managing energy vs. heat.

Going full broadside first gives you the maximum deceleration, but it also generates the maximum heat, and it puts more stress on your crew and vehicle (it's basically like slamming the brakes or driving head-on into an albeit slightly fluffy wall). On the other hand, you get out of the heat faster.

Whereas if you re-enter at an angle, you allow the body generate lift, and thus stay longer in the thinner parts of the atmosphere and bleed off velocity there, where there is less air resistance and thus less heating. It also generates a gentler acceleration profile for crew, cargo, and vehicle. But, the entry takes longer, and thus, while the temperatures are lower, the vehicles is heated longer.

This is a very delicate trade-off, and we will have to wait and see what it actually looks like. At last year's Starship update presentation, Elon Musk mentioned a re-entry angle of about 60°, although I am not 100% sure whether that was for Earth or Mars.

Remember, only the first stage of Falcon 9 lands. The second stage is expended, precisely because it goes so much faster and has so much more energy to shed that there is no sensible way of making it reusable. And Starship is of course the second stage of the Starship/Superheavy system, so comparing it to the first stage of F9 doesn't really make sense. Note that Superheavy will return vertically, just like F9. Starship is more comparable to the Space Transportation System's Orbiter Vehicle (aka "Space Shuttle") or a re-entering Apollo, Dragon, or Soyuz capsule, all of which also enter blunt end first. (Or, more precisely, at an angle that is not the most aerodynamic shape but actually both bleeds off velocity and generates lift.)

Starship is more comparable to the Space Shuttle than the Falcon 9 on re-rentry. The main difference being that the Space Shuttle was designed as a glider and thus was moving forward and was controlled with "conventional" airplane control surfaces (rudder, elevons, speed brake), whereas Starship is designed like a skydiver and thus falls straight down and is controlled with its four "limbs". (The community hasn't really settled on a name yet, some call them "flaps" or "body flaps", some "fins", I have heard combinations such as "flins", my personal favorite is Tim Dodd's "Eloneron". They really work unlike any control surface you typically find on an airplane because they are perpendicular to the airflow.)

That last phase, the descent phase, is what the belly-flop is really about. Whereas the actual entry may look very much like the Space Shuttle, except maybe with a somewhat higher angle, it is the next phase, the "now that I have gotten rid of most of my velocity, how do I get from high up in the air to the ground" phase, where Starship really differs from both a gliding landing like the Space Shuttle, Dreamchaser, X-37, or Buran, as well as a propulsive landing like the Falcon 9 first stage or New Shepard. It will fall straight down like an F9, but it will be oriented body down like the Shuttle.

And one of the main differences between F9 and the Starship is going to be that the Starship does not need an entry burn, even though it enters from a much higher velocity than F9 does. Because F9 falls in a fairly aerodynamic shape, its velocity is so high that it needs to fire 3 engines during entry to slow down. (Also, the exhaust from the engine creates a layer of air "cool" air around the lower part of the rocket. Yes, that sounds insane, but it turns out the exhaust of the rocket engine is actually still cooler than the air in front of the rocket.)

To put some numbers to the differences: the Falcon 9 first stage typically reaches a peak velocity of about 8000 km/h, maybe 10000 km/h if they're hitting it really hard. (Although that would likely mean the stage is expended, and thus the difference in landing techniques does not really come into play.) Orbital velocity, however, is at least 25000 km/h, maybe higher, so the velocity of Starship will be between 2 and 4 times that of a Falcon 9 first stage. The temperature of the bow shock wave goes up with the cube of the velocity, so at 2–4 times the velocity, Starship will have to deal with 8–64(!!!) times the temperature.

  • $\begingroup$ "you want to point the side with the most wind resistance into the flight path, and that is the broad side of the ship." It's not quite that simple. If you can orient the craft to generate some lift, it will prolong the entry time, causing a more gentle re-entry. That was done with both Apollo and the Shuttle. Neither of those entered broad side (bottom) first, but rather entered edge (Apollo) or nose (Shuttle) first. $\endgroup$
    – DrSheldon
    Nov 1, 2020 at 14:49
  • $\begingroup$ @DrSheldon: I will try to rephrase my answer to focus on descent rather than re-entry, because that's indeed where the difference is more obvious. I believe on Earth, the re-entry profile of Starship may very well look much more Shuttle-like, but it is the descent phase where the difference between gliding and falling really becomes obvious. Think wingsuit vs. skydiver. IIRC, the only official information we have is a 3D animation of an outdated iteration of Starship entering into the Martian atmosphere, so what the final iteration of Starship looks like on Earth is somewhat speculative. $\endgroup$ Nov 1, 2020 at 17:44
  • 1
    $\begingroup$ For practical purposes the shuttle basically did a belly-flop during re-entry also. $\endgroup$ Nov 1, 2020 at 20:54
  • 2
    $\begingroup$ @DrSheldon Shuttle did not enter nose first. It was pitched up about 45 degrees. The nose was pointing in the direction of travel, but the velocity vector was well below the nose. This is not from a NASA site, but it's pretty close: orbiterwiki.org/images/2/20/GPIS_6_01.png $\endgroup$ Nov 1, 2020 at 21:54
  • 1
    $\begingroup$ @OrganicMarble: You are of course correct. However, the point which I was trying to make -- that re-entry is not as simple as entering broad-side (bottom) first -- still holds. $\endgroup$
    – DrSheldon
    Nov 1, 2020 at 22:08

Let me explain as simple as possible.

Falcon Heavy (LEO Payload delivery):

2 x side boosters: Booster cut off at 1.8 km/s | ~70 km altitude

1 x Main / core / stage 1: MECO (Main engine cut off) at 3 km/s | ~130 km altitude

1 x second stage = max speed at ~8 km/s | 700+ km = Expended / discarded in space / Non reusable

Starship (LEO Payload delivery):

Super heavy booster / Stage 1: Cut off at ~ 2 km/s | ~70 km altitude

Starship: After payload delivery at LEO at speed ~8 km/s | 700+ km altitude = Fully reusable and will return to Earth

As you could see from difference in speed, Starship will no have enough fuel to do a reverse burn, even if it does payload capacity will be extremely low.

So to avoid this problem, Starship never does any re-entry burn which consumes half the fuel which it took to lift off, rather it uses the thick atmosphere of earth (~ 1 bar pressure) to hit it's wings and cause friction to slow it down, and then only do the landing burn for a mere <10 seconds which consumes extremely low fuel.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.