Why do cylindrical rockets roll?

Here’s a fun question that not only have I myself asked, but I get asked fairly often. “Why do we hear a call out like ‘Roger roll’ or ‘Roll program complete’?” At which point we can see the rocket rotate or roll on its X axis. The best example of this was the Space Shuttle, which had a very obvious and dramatic roll program. As soon as it cleared the tower, you can see it making a very impressive and sometimes scary looking roll.

Now, a maneuver like this makes sense when a vehicle is asymmetrical like the Space Shuttle, but why do cylindrical rockets like the Saturn V, Titan, Atlas, Delta IV, etc. even bother doing a roll?  Can’t rockets just tip over in whatever direction they need to go? Do a little pitch here, a little yaw there. Just as long as the pointy end is going the direction it’s intended to go, who cares which side of the rocket is facing the Earth and which side is facing space… right?

So today, we’ll first define the pitch, yaw, roll and their corresponding axis on a rocket. Then we’re going to dive into why a rocket rolls in the first place. Take a look at launch azimuths and their relationships to trajectories. Then we’ll look at some unique solutions to orientations, including some rockets that don’t roll on ascent to align with their trajectory.

Let’s get started!

This is one of those topics I love. Where at first the reason feels perplexing, then you hear one explanation and you’re like “Oh, I guess that makes sense”. Then you think of other reasons and learn of all these edge cases and come to find out, there’s actually a lot to unpack here!

And just to clarify things, we’re specifically talking about the roll program of rockets and not their gravity turn, these are two different things. We’re focusing on this…

not this….

So let’s start off with a quick overview of pitch, yaw and roll and how they correspond on a rocket. You may have heard the terms pitch, yaw, and roll, especially when talking about airplanes. On an airplane, pitch is the nose pulling up or diving down, yaw is the nose going left or right, and roll you can think of the wingtips going up or down while the nose stays in the same place. With airplanes it’s really easy to define pitch, yaw, and roll because airplanes have such obvious characteristics, like wings, landing gear, cockpit and a vertical stabilizer. And you might think, how do you define these dimensions with a cylindrical rocket?

Although a rocket is pretty symmetrical, it’s still vital to define these dimensions, otherwise your rocket might go north instead of east or something. So let’s take a jetliner and just remove the wings and tail stabilizer. Oh look, the fuselage kind of looks like a rocket! Perfect! So now we still have our pitch, yaw and roll. Just stand this baby up on its tail and let er rip. And this held true when cockpits were put on missiles, which is essentially what the Vostok, Mercury, Voskhod and Gemini programs were.

So now with a rocket on the launch pad, we can look to the cockpit for that same pitch, yaw and roll. When sitting in the cockpit, your pitch, or nose up and down is rotating on the Y axis, yawing left or right is rotating on the Z axis and rolling left or right is on the X axis.

Unlike an airplane, the pitch, yaw and roll of a rocket generally isn’t controlled by wings or fins, but is actually controlled by the engine itself via a gimbal and perhaps some auxiliary thrusters to help control roll. However, wings and fins are sometimes used for passive stability in the atmosphere. A single engine on the bottom of a rocket can only provide 2 axis of control, pitch and yaw. This is because the engine goes through the center of the rocket, it can only apply torque on two axis.

So in order for MOST single engined rockets to have roll control, you usually see axillary thrusters stuck on the sides or outer perimeter of the rocket.

These auxiliary thrusters are called vernier thrusters and I think they’re most obvious on the side of the original Atlas SM-65A rocket, and there’s several vernier thrusters on the bottom of the Soyuz rockets too. But, some single engine rockets get clever and control their roll via the gas generator exhaust, like the RS-68 on the Delta IV and Delta IV Heavy. You can see that the engineers cleverly point the dual gas generator exhausts on each side of the engine.

If you need a brush up on gas generators and the open cycle, I recently did a really in depth rundown on a few common engine cycles in my “Is SpaceX’s Raptor engine the king of rocket engines” article.

But with rockets that have at least two engines, or at least two combustion chambers like the RD-180 on the Atlas V, you can point the engines in opposite directions which will induce your X axis roll.

OK, so now we know HOW a rocket can control its roll, now we can get onto WHY a rocket needs to control its roll. Well, to begin, a rocket needs to remain stable throughout flight so it doesn’t spin so fast that it tears itself apart. OK, sure that’s the most basic reason why the rocket needs to control its roll, but we still get to the question of WHY do they intentionally roll once they get off the launch pad? So I’m gonna tell you the reason here and then we’re going to want to define a few things. The rocket rolls to align itself to its flight azimuth so its flight path becomes a simple pitch program.

Ooooh baby, we have a lot to unpack in just that one sentence huh? So first let’s talk about the azimuth. Depending on the destination of the payload, rockets need to head to a very specific orbit. And a fun reminder here, I like to say, “To go to space you go up, but to stay in space you need to go sideways really, really fast”, which is really what orbit is. To get to your desired orbit, you want to make sure that sideways part is pointing in a very specific and accurate direction.

If you launch a vehicle on the equator straight east, not only would you take full advantage of the Earth’s rotation which gives the rocket a nice little boost, but you’d also place your vehicle on a 0 degree inclination. Like a nice little belt right around the Earth’s equator.

Another example is the International Space Station which is on a 51.6 degree inclination. It’s on this exact inclination so the Russians can participate and launch directly to it ,without dropping boosters on neighboring China and without doing a costly dogleg maneuver.

And just as a reference, if you launched straight east out of Kennedy Space Center, you’d be on a 28.6 degree inclination, which you may notice, is the exact latitude of the space center.

So here’s where we get to what your azimuth is. The azimuth is basically if you were holding a compass on the launch pad, which direction do you want the rocket to go to get to your desired orbit?

But we should pause for a second here and clear up one thing… because this definitely confused me a bit. Let’s be sure and note the difference between the azimuth and the inclination. The azimuth is what’s on the navigation ball inside a cockpit. North on the nav ball is 0 degrees, while east is 90, South is 180 and west is 270.

This does not line up with inclinations. A zero degree inclination is due east on the equator, while a polar orbit is inclined 90 degrees. But again, minimum inclination depends on your latitude, flying due east will only correspond to a zero degree inclination if you’re on the equator. And another side note, all prograde orbits (or orbits following the rotation of the Earth) are between 0 and 90 degrees inclination. If the rocket is flying south from the equator, it’s still between 0 and 90 degrees, because inclination is really just a measure in degrees how far off angle the orbit is from the equator.

And of course it’s not quite as simple as this… If you want to go 51.6 degrees to rendezvous with the ISS, you don’t actually point at 51.6 degrees, you actually point at about 45 degrees.. but now we’re getting into some fun math that takes into account the rotation of the Earth and spherical trigonometry, which might be getting a little too far into the weeds for this video.

So now that we know that rockets don’t all follow the same path to get to space and to their destinations, we’re starting to get some of the puzzle pieces as to why they might intentionally roll… For our next clue we need to look no further than the launch pads themselves! And since we’ve mentioned rockets like the Space Shuttle and Saturn V, let’s look at one of the most famous launch pads in the world…  A launch pad that saw lots of action from both those vehicles and now SpaceX’s Falcon 9 and Falcon heavy, of course I’m talking about Launch Complex 39A at Kennedy Space Center.

LC-39A is a great example because it is perfectly aligned North / South / East / West. Take a look here, we can see the flame trench and crawler way runs perfectly North / South.

Let’s start off with the Saturn V which first launched from 39A on its inaugural test flight on November 9th, 1967 and last launched Skylab on May 14th, 1973. When the vehicle crawled out to the pad, you’ll see the launch umbilical tower on the north side of the vehicle and its crew access arm swings around and connects to the east side of the vehicle. This is where the astronauts get in and once they’re in, they’re facing with the top of their heads due east and their feet west.

So of course, along with the command module, the rest of the vehicle had certain features. Such as the fuel and electrical umbilicals that connected the rocket to the launch umbilical tower, some external raceways which had some wiring and stuff, but most importantly when talking about the alignment of the rocket was a thing called the IMU. The IMU, or instrument unit, sat on top of the Saturn V’s third stage and housed the rocket’s guidance systems. This included a digital computer, ooh la la, an analog flight control computer, accelerometers and some gyroscopes.

Now in the case of a Saturn V heading to the moon, the launch azimuth was 72 degrees which is 18 degrees north of due east. So while on the launch pad, the flight path and the belly of the rocket were 18 degrees off from each other.

And here’s where we get the first reason for the roll program. Instead of moving the entire launch pad to just face the belly of the rocket at that 18 degree angle, the rocket could simply perform a roll to align its belly with its flight path and simply pitch over. This means if the vehicle performs a simple roll to basically zero out the difference between the flight path and the body’s physical coordinates, it would have a value that’s a nice easy zero. Now all the rocket has to do is pitch over! This made it so the computer really only had to calculate one set of numbers instead of two, making the math and the calculations much much easier. Less variables = a good thing. Keep it simple.

Another physical consideration is gimbal lock. Gimbals can freely rotate on all three dimensions and align themselves to a fixed position in space which can tell the guidance computers where the vehicle is pointing. By zeroing out one of the numbers, you’re keeping the gimbal as far away from a potential gimbal lock as possible. And gimbal lock can be a very, very bad thing.

So in order to demonstrate why zeroing out a vehicle’s roll is a good thing, let’s build a quick rocket in Kerbal Space Program. By default when you build a rocket, it’s aligned perfectly north / south / east / west, with pitch aligned north and south. So to head out on an equatorial, 90 degree inclination orbit, you need to only press a single key the right amount. In this example, that key is the D key which will yaw you over due east. One finger flying, nice and easy!

Now let’s rotate the rocket 20 degrees or so away from aligned and still try and follow a perfect 90 degree inclination due east. Now this can still be easily done, using just two keys this time, but it is noticeably harder to do. Why not keep it simple?

Or here’s another example. This is a map of downtown Waterloo, Iowa. Notice that the streets run NE to SW and NW to SE and are aligned to the river and not true north. When walking around, it’s unlikely that you wouldn’t redefine your own coordinates and start thinking of anything on this side of the river as north and this side as south. It makes navigating a lot easier than thinking about NE and SW.  So if the rocket and the launch pad are always in a fixed position, which spoiler alert, they pretty much always are, well, kind of, we’ll talk about that more in a second, the next easiest thing to do is program the rocket to do a quick roll to align itself with its azimuth.

We take the navigation from a 3 dimension equation to a 2 dimension equation.

Now of course whether the vehicle pitches or yaws is a bit pedantic, because isn’t it really about the same? Well there’s still some important distinctions. Sticking with Apollo, with the astronauts heads were pointing due east, they were actually on the “belly” of the Saturn V. The command module and the rocket actually had 180 degree opposite Y axis orientations, for reasons I don’t actually know – Ha Ha.

But this meant that when the rocket pitched over, the commander could look out the small port window in the blast protective cover and get visual references of their orientation. By zeroing out the roll, the horizon would always appear across the window, which made it easy to use as a reference. This also made it so that if the commander saw the ground suddenly coming up or the horizon spinning, they may have considered aborting, or at least had a good visual reference on if that would be necessary.

Another reason there’s usually a defined “belly” of a rocket is to place the radio antennas and receivers in the optimal place to have best contact with the ground during ascent. This is especially true with the space shuttle which if it had ascended with the orbiter on top of the external fuel tank, it would’ve had much worse line of sight.

While we’re talking about the Space Shuttle, its roll program was even more necessary due to its unique shape. Not only was it structurally the best option for the wings and the struts holding the external fuel tank, but by flying with the orbiter in the wake of the external fuel tank, there was actually a 20% increase in payload capacity!

Although most rockets look relatively symmetrical, they almost always have some kind of protruding feature. Take a look at the Saturn V. It had very large bumps and bulges on the outside that definitely aren’t insignificant when factoring in the ascent profile.  You’ll see these areas where additional piping or wiring is house inside sections called raceways. You’ll notice there are two different raceways on each side of the Falcon 9 and Falcon Heavy cores. As a matter of fact, you can tell the two outer cores of the Falcon Heavy are 180 degrees opposite of each other because of the two different raceways.

But back to the Shuttle. The shuttle controlled its pitch via gimbaling nozzles on the solid rocket boosters. Yes, the main shuttle engines could gimbal too, and a lot, but they primarily gimbaled to maintain the center of thrust going through the center of mass.

By using the solid rocket boosters to control pitch, the gimbal vectors are in line with each other relative to the center of mass… This probably makes it easier to control. It might not be a huge deal, but just note that if a vehicle like this flew sideways, having an engine on the top and an engine on the bottom, their engines would have a different amount of leverage over the vehicle compared to if they were all in a row.

They also fly these rockets flat to the horizon for stage separation so the boosters have the lowest chance of hitting the center core.

While we’re on the topic of Falcon Heavy and SpaceX, here’s a fun little fact. The Falcon 9 does not perform a roll program to align to its azimuth, neither does the Electron rocket. Both of these rockets just pitch and yaw over however much is necessary and roll for aerodynamic considerations or a few other variables as well. But controlling a rocket in a 3D space like this is actually a lot harder than it sounds. It took a generation of grad students to actually solve the linear algebra AND have access to computers powerful enough on the rockets to do the math in real time for this type of control.

So if the Falcon 9 and Electron don’t need to roll, why do they? Well, for fun!

Ugh, I totally just got trolled here by Elon. So on June 12, 2019, SpaceX launched a trio of satellites for the Canadian Space Agency. Soon after lift off, the Falcon 9 did a pretty substantial roll.

Now again, vehicles aren’t symmetrical and although the Falcon 9 navigates along both axis, it’s likely this particular launch had a roll due to payload considerations. The customers might have certain constraints and with this particular launch having an offset payload, perhaps they need to fly it a certain way for the payload to handle the g-forces the best way.

The Falcon 9 is also perhaps a little unique and that it for sure wants to be oriented correctly at stage separation so the first stage has both of its nitrogen thrusters able to do the flip maneuver. Since the Falcon 9 has only two packs of cold gas thrusters that are 180 degrees from each other, it means if the vehicle were rotated 90 degrees, only one set of thrusters could help with the flip instead of 2.

And here’s a fun story while we’re talking about the Falcon 9, have you ever seen the very first Falcon 9 launch? It unintentionally rolled almost 45 degrees immediately after take off. This was due to the gas generator exhaust that has a slight angle to it, just like how the Delta IV’s RS-68 uses its gas generators to roll, the 9 Merlin engines had so much extra torque from the gas generator exhaust, it took a second for the engine gimbals to cancel that roll out.

And one more reason why rockets would roll is for fairing separation. Now, I don’t exactly know what considerations go into choosing whether the fairing should split on its y axis or z axis, it should be noted that this is certainly taken into consideration. And from what I can tell, SpaceX usually ditches its fairings on its Y axis or up and down, while ULA tends to ditch its fairings off to the sides on the Z axis. Why exactly each launch provider chooses to ditch them in this manner, I’m not sure, but it’s fun to note.

So a few 21st century rockets finally took the roll to align to the azimuth program out, but perhaps my favorite rockets that didn’t roll align were Soviet era rockets. Remember near the beginning when I said it’d be too hard to turn the rocket and or launch pad to align with the trajectory? Well, that’s actually exactly what the Soviet Union came up with for their R7 family of rockets like the Soyuz. That’s right! The ENTIRE launch pad actually rotates to align the rocket up with its trajectory.

Now some down sides to this is your azimuth might change ever so slightly throughout your launch window, so by aligning the pad to your azimuth, you might lose some flexibility in the launch window and flight path. This is something the new Soyuz 2 can do away with now that it has a digital flight computer and can align itself to the correct azimuth. Although crewed missions still use a Soyuz-FG which utilizes the rotating table.

But lastly, there was still perhaps the most advanced, ahead of its time rocket, the Soviet Union’s N-1 rocket, which was meant to *cough cough* follow it’s flight path using both pitch and yaw. It had roll control thrusters that were undersized for the first 3 launches and upgraded for the fourth launch, but they weren’t used to align to the launch azimuth, they were just used for stability. Man, I still really wish the N-1 had worked out. It’s such an awesome rocket!

So to summarize… Rockets roll for a few reasons, and like all rocket science and engineering, there’s actually some good reasons. But as for why, it’s generally easier to roll to align the vehicle to its azimuth than it is to move the launch pad, it makes for easier calculations for the guidance computer, rockets roll for aerodynamic and structural considerations, they roll for the astronauts vantage point and visual references, they roll for fairing deployment orientation, they roll to align auxiliary or control thrusters, and they roll for the best line of site for the communications and downlinks.

Well does this help answer that question? It’s another one of those things where you know there’s probably a good reason for it, but it can be pretty hard to come up with all the fun little reasons. Hopefully, this helps us appreciate just how many of these little, but important, decisions engineers and scientists need to come up with. There’s always a reason for all the strange little quirks. Let me know what other questions you have about roll programs, rockets or rocket science in the comments below. I have a crazy long list I’m trying to chew away at here, so stay tuned, there’s a million things yet to learn!

If you want to help me continue to do what I do, please consider supporting by becoming a Patreon member! Patreon members gain access to our exclusive sub-Reddit, an exclusive Discord channel, they get bonus material AND exclusive livestreams!