"Real" Gravity Turns

[davidpsummers]

So to do a "real" gravity turn, you pitch over a bit and let gravity turn your rocket as you ascend. OK. The question I have is why?

Are we doing it to simulate real life (which is a perfectly good reason) or does it give some advantage?

I means, It seems to me that in a lower part of the atmosphere, you want to go as straight up as possible to get out of it, and _then_ start to pitch over when the air is thin.

Edited May 25, 2015 by davidpsummers

[KarateF22]

A flawless gravity turn is the most efficient way to ascend to orbit, as you waste virtually no fuel on rotation, letting gravity do all the positional work for you. It should be noted that you generally are going nearly straight up in the lower atmosphere, however you want to establish at least a 5-10 degree tilt early so that gravity will tilt you in that direction, if you go straight up the direction that gravity will tilt you is less predictable; additionally the earlier you start your gravity turn the less drag you create by not pointing prograde.

[n0xiety]

Perfect gravity turns are always more efficient and believe me it is noticable in big numbers when you have a mod like mechjeb or kerbal engineer where you can monitor dV values. There is one thing about the gravity turn that you shouldn't do tho and that is to pass terminal velocity. It will make you loose alot more to drag at low levels of atmosphere. Having a SLT of 1.2 - 1.4 at launch works well.

[Murph]

When done properly, a gravity turn is the most fuel and aerodynamically efficient path to orbit.

Some highly simplified things to think of:

At all points of the ascent, your real priority is gaining horizontal velocity, not vertical. At sea level, horizontal is very limited before it becomes highly inefficient. At all altitudes, vertical is inefficient, as it has none of the velocity that you actually want, it's just using finite fuel to repeatedly punch gravity (which has infinite stamina). Once a velocity vector is established, it takes time and energy to change that vector efficiently, hence the slow progressive turn, which also should avoid putting the nose too far across the airflow (which can create a growing force pulling you away from the desired vector, as well as far more drag than going directly into the airflow).

It's all about getting horizontal velocity as soon as possible, but without forward velocity becoming so high for the current altitude that you're fighting huge drag from it. As the practical horizontal limit rises (due to lowering air density), you want to be following that limit as closely as practicably possible.

The curve of a real gravity turn is mostly vertical in the early portion of the flight, only slightly horizontal until you're up into much thinner air, but gives you much better than 0 horizontal when you reach the thinner air, with overall velocity vector much closer to the horizontal that you really want/need it to be.

Alternatively, think of a playing field. To go from corner to diagonally opposite corner, you can walk along the edges, but it's quicker and uses less energy to walk diagonally across it. It's also quicker and uses less energy if you cut across the intermediate corner in a long smooth curve (which is closer to a gravity turn).

[lincourtl]

The initial pitch maneuver, which should be done as soon as you are moving fast enough not to fall out of the sky from cosine losses, also makes sure you don't rain flaming death down on the space center from spent boosters or your own RUD. If avoiding flaming death for your ground crew and any looky-loos is a priority for you.

Edited May 24, 2015 by lincourtl

[SAI Peregrinus]

Real launches do gravity turns for the reasons mentioned above. The Wikipedia article has some more details, including the (often overlooked) fact that an inverse gravity turn is the optimal landing trajectory on airless bodies.

[TORCHWOOD]

Off topic, but that's an awesome photo!

[cybersol]

Devil's advocate post incoming. First of all, realistic gravity induced turns are fun and have a definite place in the game. However, they are usually not the most efficient turns when I measure how much dV is used getting to space in 1.0.2. A more efficient profile dV wise typically has more thrust (1.7-2.2 starting) and goes shallower (45* at 12k). The reason has to do with how the relative contribution of the various loss terms that sap your dV.

When you spend thrust off axis from prograde to accelerate the turn, you lose dV to steering loses. As you fight against drag, you suffer drag loses due to aerodynamics. All the while you must fight against gravity, and the thrust you use perpendicular to the ground does not help your orbital speed.

It turns out that almost always gravity losses dominate and drag losses are less than half as important. Meanwhile, steering loses are tiny unless you point in crazy orthogonal directions or spin the rocket around a few times. So minimizing gravity losses will get you the most savings to your dv budget. Notably, the way to minimize gravity losses is to leave Kerbin faster. To leave faster you need the higher TWR. But when you are giving gravity less time to act, you often have to pull the gravity turn over faster to get the more dV optimal shallower path.

[lincourtl]

First of all, realistic gravity induced turns are fun and have a definite place in the game. However, they are usually not the most efficient turns when I measure how much dV is used getting to space in 1.0.2.

Aye, it's the difference between an ideal ascent trajectory and an optimal ascent trajectory, the latter of which is incredibly hard to calculate. It goes beyond my math skills anyway.

In real life rockets rarely allow gravity to guide their ascent, but rather follow the computed optimized ascent trajectory.