28
$\begingroup$

In videogames (such as KSP or Spaceflight simulator) people put rockets in orbit with the following procedure:

  • Turn on the engines and move vertically for a bit.

  • Start a roll program to turn the vertical trajectory into a ballistic one.

  • Keep going until the Apoapsis of the trajectory reaches the desired height.

  • Turn off the engine and wait until the rocket (almost) reaches Apoapsis.

  • Burn prograde to increase periapsis until orbit is achieved.

I understand the reasons behind each of this manouveurs, however I'm wondering if this is how real rockets get into orbit.

Cutting off the engine and letting the rocket loose vertical speed looks counter-intuitive to me (you basically spend a lot of fuel to accelerate and then you let the rocket slow down).

So my question is: how do real rockets go in orbit? Is there somewhere an explanation of the manouveurs and trajectory the they follow to get in orbit?

$\endgroup$
  • 10
    $\begingroup$ Minor correction; the turn is a "pitch program", not a roll program. Roll is rotation around the long axis of the launcher, and does not itself change the course of the rocket. $\endgroup$ – Russell Borogove Jun 2 at 17:11
  • 3
    $\begingroup$ @DavidHammen My understanding was that the "roll program" is the roll done to line the pitch axis of the rocket up with the desired flight azimuth prior to the pitchover, so the gravity turn can be done with all pitch and no yaw, but you're the ex-ex-rocket scientist. $\endgroup$ – Russell Borogove Jun 3 at 2:17
  • 6
    $\begingroup$ @RussellBorogove - The roll program was a rotation about a single axis that put the vehicle in the right roll orientation for crew and vehicle operations, the right yaw orientation for flight azimuth, and the right pitch orientation for the start of the gravity turn. That single axis wasn't about the roll, pitch, or yaw axis. $\endgroup$ – David Hammen Jun 3 at 2:33
  • 4
    $\begingroup$ Hence its true NASA name, SAR - Single Axis Rotation. $\endgroup$ – Organic Marble Jun 3 at 2:37
  • 4
    $\begingroup$ @Skyler that sounds like a separate question that could be asked $\endgroup$ – Baldrickk Jun 3 at 15:53
2
$\begingroup$

That is how real rockets go into orbit, and for orbit you need to lose vertical speed or it would be a spiral. Orbit is about horizontal speed though not vertical.

$\endgroup$
  • $\begingroup$ vertical speed is usually lost through gaining altitude. $\endgroup$ – JCRM Jun 4 at 8:52
36
$\begingroup$

I understand the reasons behind each of this manouveurs, however I'm wondering if this is how real rockets get into orbit.

Cutting off the engine and letting the rocket loose vertical speed looks counter-intuitive to me (you basically spend a lot of fuel to accelerate and then you let the rocket slow down).

In most real launches to low Earth orbit, the burn continues from liftoff until orbital insertion, without a coasting phase. Some (like Antares) do coast between the first stage and second stage burn; the exact design of the launcher determines which approach is more efficient.

In flights to higher orbits, for example geosynchronous orbit, a coast phase is always used; this approximates a Hohmann transfer orbit.

In KSP, neither the trajectory nor the rocket design, typically, is optimal, and the planet is much smaller than Earth, so the tradeoffs are different.

$\endgroup$
  • 1
    $\begingroup$ and of course many rockets do in fact reduce thrust for part of their ascent :) $\endgroup$ – jwenting Jun 3 at 3:46
  • 2
    $\begingroup$ The reduction in thrust for part of the ascent is usually when traveling through maximum dynamic pressure "max Q" the reduction in thrust is used to reduce losses to due to the drag rise associated with breaking the sound barrier while still in the thick soupy lower atmosphere $\endgroup$ – DJ319 Jun 3 at 9:26
  • 3
    $\begingroup$ One thing simulators often don't simulate is that many rocket engines have a limited number of restarts (or may not be restartable at all!) and therefore it makes more sense to keep it running at a lower thrust than to turn it off and back on. $\endgroup$ – Skyler Jun 3 at 15:29
  • 1
    $\begingroup$ @Skyler In Kerbal's case, there are mods for that! $\endgroup$ – Harabeck Jun 3 at 16:35
  • 5
    $\begingroup$ Actually, you can install a mod that makes solar system distances and masses real in KSP. I tried it and it reduced opportunity for coasting phase. $\endgroup$ – Mołot Jun 4 at 13:01
18
$\begingroup$

It's real, although many rockets heading to low orbit are engineered such that when the apoapsis reaches the desired altitude it's time to begin the final burn.

Remember that orbit is not truly zero gravity no matter how many times Hollywood calls it that. Rather, it's free fall--craft in low orbit fall towards Earth just about as fast as they would here on Earth. It's just their velocity counterbalance this (think of spinning a rope around over your head--it's pulling outward.) This is not some magical property that exists only in orbit, though--rather, orbit is simply the speed at which the forces balance. Moving horizontally gains some benefit even when you're not at orbital velocity. Hence anything heading to orbit builds horizontal velocity as fast as it can. Hence what you have observed of tipping soon after launch. The faster the rocket is going the less you are losing to gravity and the more the rocket can tip and yet avoid falling back.

The economics of Kerbal Space Program are very wrong, engines are too cheap and fuel/fuel tanks are too expensive. This favors overpowered rockets that mean the engines have to shut down long before apoapsis. (An optimum KSP rocket generally goes for the maximum thrust consistent with making the turn.) In the real world it's the engines that are the expensive part and thus rockets are built with a lot less thrust than a typical KSP design--in many cases it is specifically calibrated to burn all the way to orbit without a shutdown. KSP also has no cost for relighting an engine--when in practice you must include an ignition system (which can require it's own chemicals--see the loss of the central core in the first Falcon Heavy launch) and when you're in free fall it also requires some sort of ullage burn. (The fuel will be floating around in the tank, trying to light the engines in that state can result in them not sucking in the right amount of fuel and oxidizer. Bad things are likely to happen. To prevent this you either include some small solid rocket boosters or else use some of your reaction control system propellant to do the job.)

$\endgroup$
  • 7
    $\begingroup$ Or in other words, it's falling to earth but missing the ground constantly. $\endgroup$ – Tschallacka Jun 3 at 8:59
  • 1
    $\begingroup$ The other issue with KSP is the planet: it's too small, with too thick of an atmosphere and too much surface gravity. This favors a launch profile with a mostly-vertical kick to get above the atmosphere, followed by a short horizontal burn to pick up orbital speed. $\endgroup$ – Mark Jun 3 at 23:04
  • 1
    $\begingroup$ To be precise, the atmosphere of Kerbin is about half as deep as Earth's, but the radius of the planet is 1/10th of Earth. So proportionality the atmosphere is 5x deeper than Earth's. $\endgroup$ – Blake Walsh Jun 3 at 23:30
  • 3
    $\begingroup$ @Tschallacka As Douglas Adams put it, the trick is to throw yourself at the ground, and miss. $\endgroup$ – anaximander Jun 4 at 14:26
  • 1
    $\begingroup$ @Mark Note that Kerbin's atmosphere favors a later tip. If you're launching from an airless body you want to tip really early--I've launched from Minmus by burning vertical for a couple of seconds, then shutting down the engine and rotating (something else the game gets wrong--you rotate way too fast) to horizontal and burning again. $\endgroup$ – Loren Pechtel Jun 4 at 22:53
33
$\begingroup$

Physics

In regards to the physics, KSP is fairly realistic, other than it not modeling n-body physics (which isn't really relevant in scope of orbiting Earth/Kerbin).

In regards to the engineering, KSP makes its parts much stronger than the real life components. The tradeoff here is that when something finally breaks in KSP, it explodes and disappears. Engines are also more powerful, and fuel is (slighly) less heavy. This is all done to "round the numbers" in favor of the player to make sure that the limited part set can get most reasonable jobs done.

In regards to the controls, KSP's control system is not as fine-grained as a real world spacecraft. Given a keyboard and mouse and moslty eyeballing the thrust gauge, there is only so much precision you can reasonably expect from a player.

In regards to the ascent profile, due to the stronger craft and less fine-grained controls, KSP spacecraft are able to ascend using sharp(er) turns than is realistically possible in real life. In KSP, a simple ascent profile is to go straight up for about 10km, and then make a 45° turn and start building horizontal speed.

In real life, a rocket won't be as easy to turn (e.g. a 45° turn) due to the stress that puts on the vehicle; most importantly it's the incoming wind from still moving straight upwards (due to inertia) that is going to start beating down on the side hull (which you expose by orienting your craft away from its prograde vector).
To avoid that, the best approach is to orient your spacecraft as close to its prograde vector as you can, because that minimizes exposure of the less aerodynamic side hull compared to the aerodynamic nose.

This means that in reality, spacecraft tend to make their 90° (from vertical to horizontal) turn slow and steady (i.e. pretty much immediately after launch). And this is actually not that hard to do: even a minor (but consistent) deviation over a long period can do most of the work for you and doesn't require active steering (other than necessary corrections due to imperfections in the initial setup).


Gameplay

In KSP's defense, there is a boundary of realism that you should avoid when it negatively impacts gameplay. The level of precision required (for both thrust and attitude adjustment) to achieve a "real" ascent profile is quite high, requires a lot of calculation, and any imperfection in the simulation would render any calculation impossible to get accurate anyway. So rather than risking the player's calculated approach being useless after all that effort, KSP opted to simplify things slightly to keep things much more intuitive.

Most rockets can be flown by the seat of your pants (and a basic grasp on atmospheric friction and that an orbit is achieved through horizontal velocity), and this is a lot more fun for most people. KSP doesn't force a player to plan their entire mission down to every control input before ever launching (NASA pretty much does, other than necessary corrections of course). KSP gives the player the freedom to plan a little, and then experience the mission (and only figure out some things such as the interplanetery maneuver at that stage, not before launching).
This is also why (unmodded) KSP doesn't implement food/oxygen logistics. Not being bound by that means that there is no real drawback to orbiting a few more times before finally doing your Hohmann transfer (e.g. waiting until the Mun is in the right spot).

NASA has years to perfect their approach and calculate every little detail about the mission (e.g. the aerodynamic impact of the angling of a single uncovered solar panel, or what a % of a degree of inclination can do to affect the ascent path), but requiring a player to do so for KSP would lead to a ridiculous amount of preplanning and testing before every achieving anything. That's not fun (except for a few die hard fans) and it's simply not a good design design (gameplay wise).

$\endgroup$
  • 3
    $\begingroup$ One difference you forgot to mention is that Kerbin's smaller radius makes the horizontal velocity needed to stay in orbit a lot lower — only around 2.3 km/s, compared to 7.8 km/s for low Earth orbit. That's one more factor that makes KSP a lot more forgiving of suboptimal launch trajectories. (There's a KSP mod called Real Solar System that pretty much does what its name says, including replacing Kerbin with a realistically sized Earth, and it indeed makes getting to orbit a lot harder.) $\endgroup$ – Ilmari Karonen Jun 3 at 11:05
  • 2
    $\begingroup$ Also, another notable difference is that KSP rockets, especially as built by beginning players, often have poor aerodynamics and relatively high drag, which makes climbing quickly out of the lower atmosphere extra important. This was particularly true in older versions of the game, which had no payload fairings and where the simplistic aerodynamics model made the atmosphere feel very "soupy". Later versions of the game have tweaked and improved the aerodynamics model a lot, but it's still far from realistic. Which is understandable; after all, KSP is still primarily a space flight simulator. $\endgroup$ – Ilmari Karonen Jun 3 at 11:19
  • 2
    $\begingroup$ @IlmariKaronen: The scale of the planet doesn't really matter that much, other than being used as a scaling factor for time. In essence, KSP could simply relax different physical constraints and scale the engine power up, and then an Earth size Kerbin would work equally well. It's all a matter of what you tweak and what you don't. KSP's developers opted for a smaller planet for more reasonable time to orbit, but kept gravity the same (regardless of the smaller planet) because it's intuitive. Or you could speed up time by the same factor and the (player) time to orbit would stay the same [...] $\endgroup$ – Flater Jun 3 at 11:24
  • 2
    $\begingroup$ [...] but then the experience of gravity would be off. The two biggest factors here (gravity and player-preferred time to orbit) leads to a contradictory nature in regards to the size of the planet. Oversimplified tweaks (e.g. massive engine output) then leads to other contradictions (i.e. massively different atmospheric pressure would be needed to maintain a sense of normalcy). Something has got to give, and since people don't really have a grasp (nor penchant) for the size of the planet, that's the best way to compromise between realism (gravity) and gameplay (time to orbit). $\endgroup$ – Flater Jun 3 at 11:27
  • 7
    $\begingroup$ "In KSP, a simple ascent profile is to go straight up for about 10km, and then make a 45° turn and start building horizontal speed" - This meta has changed a bit from earlier versions of KSP; a more modern meta is to start a gravity turn shortly after launch more like a real rocket. $\endgroup$ – Bryan Krause Jun 3 at 15:54
5
$\begingroup$

The 'coast' is a very natural part of getting to orbit. Burning close to the planet's surface saves the total amount you have to change your velocity by. The mechanics are a tad complex but the short version is you get more energy by changing your velocity when you are going faster. As you lose speed while you ascend, its best to do all of the burning you need to get out of the atmosphere as early as possible. This doesn't get you into orbit though as you will always need some burn above the atmosphere to reach an orbit. If you can get the burn done to get to your final height before you get there, you'll end up with a 'coast'.

The straight up to start with and pitching over are to do with balancing getting out of the thick part of the atmosphere as soon as possible and not having to thrust against gravity for longer than necessary (aero and gravity loses). The 'magical' part about straight up is not to do with orbital mechanics, but its often close enough to optimal.

Both of these are accurately-ish (enough to match these high-level outcomes) modeled by KSP and others.

Indeed some rockets do follow exactly that description. The reasons some real-world rockets don't is due to engineering limitations not modeled well in games. For example if you cant restart the engine you have to do it as one continuous burn. If you have too low a thrust to achieve the burn before you reach the place you'd have to do the second: again one long burn is the way to go.

KSP rockets restart without any penalty so that goes away, and designing a rocket that just has enough thrust to join circularising with ascent involves more maths than most players are willing to undertake. Most space programs don't have that problem... However KSP rockets are missing a lot of the limitations real world ones are subject to. Mostly cost and restart-ability and mass relative to the other parts

$\endgroup$
7
$\begingroup$

Yes, in very broad terms this is how real spacecraft enter orbit. The maneouver is called a gravity turn, albeit KSP encourages (very) inefficient versions of it, just because it's faster (in terms of ground-to-orbit-time) and more fun (in terms of waiting-for-something-to-happen-time) to add more boosters and raise your apoapsis faster.

An efficient gravity turn in stock KSP looks a bit more like:

  • Turn on the engines and move vertically for a bit.
  • Start a pitch program to turn 10-20 degrees
  • Once the rocket is pointing towards prograde, use SAS to keep it towards prograde
  • Once the rocket has passed maximum dynamic pressure, start throttling down.
    • This is done in order to minimize atmospheric drag and vertical climb dV loss.
    • An easy trick is to keep "time to apoapsis" more-or-less constant at 50secs.
  • Keep going until the Apoapsis of the trajectory reaches the desired height.
    • An efficient gravity turn means that this can take a long time, in excess of 15 minutes.
  • Turn off the engine and wait until the rocket (almost) reaches Apoapsis.
    • Or not, because KSP engines can throttle down to 0.001% and so they can keep a theoretical efficient prograde burn instead.
  • Burn prograde to increase periapsis until orbit is achieved.
    • For an efficient launch, this is less than 20 m/s dV. Can be done with RCS even.

You might want to experiment with the GravityTurn Continued mod for KSP. Note that it's optimized for stock KSP.

I also suggest you try RealismOverhaul for KSP - this changes the physics of the game so it behaves more like launching from the real earth. Let me quote one of its wiki pages, which summarizes differences between stock KSP and Earth-like launches:

You don't want to go up to 10km and then turn over; that wastes a lot of fuel. You also don't want to throttle down, since that ends up wasting fuel as well (due to increasing gravity losses). Instead, go full throttle the whole way, starting your gravity turn when you're going ~50 m/s (for an initial TWR of ~1.7) or ~100 m/s (for an initial TWR of ~1.3) and aim surface prograde until you're out of the atmosphere. If you've done it correctly, you'll be flying down 45 degrees at ~1 km/s.

You're not going to coast to apoapsis and then burn to circularize; you don't have the time for that kind of a burn. Instead you're going to have to keep burning the engines until you're in orbit. Once you're out of the lower atmosphere (~60 km) aim the rocket slightly above prograde to delay reaching apoapsis. For lower orbits you might have to circularize after apoapsis (burn up a lot to compensate). If you think you're not going to be able to circularize before you fall back into the atmosphere, you're not going to be able to; angle the rocket up to delay reaching apoapsis, and do that as soon as you can to reduce steering losses.

If you try out RealismOverhaul, try as well the MechJeb2 ascent guidance, which provides several ascent path profiles. Let me quote from the MechJeb wiki:

Powered Explicit Guidance (RSS/RO) PEG is actual gravity turn algorithms from the Surveyor missions that properly integrates the trajectory.

And if you want an even more hardcore experience on real-world ascent algorithms, have a look at PEGAS. Let me quote its KSP forum page:

I am proud to present PEGAS: Powered Explicit Guidance Ascent System - a kOS script for automated launches using a real-world guidance algorithm taken straight from the Space Shuttle documentation: the Unified Powered Flight Guidance!

Those KSP mods can help you see how the real-world PEG ascent algorithms works.

$\endgroup$
  • $\begingroup$ If you find yourself turning too much in a gravity turn to escape the atmosphere, try a more powerful engine. Conversely, you can throttle down your engines in flight if you're not turning enough. $\endgroup$ – Fax Jun 4 at 10:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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