The Secrets of Rocket Design Revealed
by Tory Bruno
I am going to share with you some of the little-known techniques and implications of rocket architectural design. Why big rockets sometimes do less. Why little rockets sometimes cost more. And why every rocket has its very own, perfect mission.
There is no single, best rocket. Different rockets do different things. As it turns out, the design of a rocket flows directly from the mission the rocket is intended to do, and there are many different missions. Any given rocket is optimal for a specific orbit and payload. Its efficiency falls off as we move away from that perfect case.
Space Destinations: So Little Time, So Many Orbits
So, let’s start with understanding what the typical missions are that most space launch vehicles are designed to accomplish. This graph below captures the orbits that spacecraft are typically delivered to.
You will see right away that there are eight common orbits. These are from the rocket’s point of view, not the satellite’s. For example, the geosynchronous transfer orbit (GTO) is never the satellite’s final destination. Instead, this is the orbit the rocket delivers it to and considers its own job finished. In this case, the satellite will, after separation, and on its own, proceed to raise itself to its ultimate destination in geostationary orbit (GEO) over the span of weeks or months. Another crazy sounding example is the interplanetary mission. The spacecraft might be going to Mars, Jupiter, or even Pluto, but the rocket thinks of this as a “low earth orbit (LEO) mission” because that’s where its job is finished.
Each set of bars on the above graph (Figure 1) indicate the major, driving tasks that the rocket must accomplish in order to perform its mission of delivering the satellite to its orbit. These include things like the number of burns that the upper stage must do, how long the stage must coast in between burns, the total energy delivered, etc. I have normalized each bar, which is to say that I have made these a ratio of what it takes to reach LEO. For example, the longest duration coast for LEO is shown as 1.0, while the duration for a typical geosynchronous orbit (GSO) insertion is shown as nine because it’s at least nine times longer.
It should be readily clear that, as we move from left to right (from LEO to GSO), that the rocket’s task gets harder. The missions on the right require more energy and significantly more complexity. In fact, all of the missions seem to fall into two broad categories: those on the left we’ll call “Low Energy” and those on the right that we’ll call “High Energy.”
This difference is the first big driver for the architecture of the rocket. The fundamental differences, from the rocket’s point of view, are how long the journey takes and the “division of labor” between the first stage and the upper stage. A Low Energy mission is usually a quick, 15 minute ride to space, with just a couple of upper stage burns. A really long LEO mission might still take only 30 minutes. Conversely, a High Energy mission will take from six to eight hours or more (10, 20, even 40 times longer!), involving many burns, long coasts, and complex maneuvers, which are executed at precise times and locations.
Low vs. High Energy Architectures: Why Your Rocket Looks Different
The fundamental tasks performed by each stage are very different because of this. The high energy rocket’s upper stage is primarily charged with transiting from LEO to a higher energy destination orbit, where it will insert the spacecraft. Consequently, we want the booster to carry it nearly to LEO, preserving the propellant (which is the same as to say, the “energy”) in the upper stage for its long, complex journey.
This means that the High Energy booster’s basic job is to climb out of the Earth’s gravity well and deliver the upper stage as close to LEO as possible. Obviously, that booster will fly very fast, and very far away from the pad before finally separating.
The first and second stages of the Low Energy optimized rocket have tasks that are very different from the division of labor we just outlined for the High Energy rocket stages. Remembering that this is only a 15 minute ride to LEO, we can see that the upper stage typically only needs to operate in orbit for five or 10 minutes, not eight hours. This means that we are free to design a different distribution of labor between the first and second stages that is more optimal for the LEO destination. We no longer need to preserve the upper stage fuel for transit from LEO. Instead, the upper stage can share in the burden of climbing out of Earth’s gravity well to LEO.
Consequently, the Low Energy rocket can stage earlier in flight, at a lower altitude, velocity and down range distance. Figure 2 below sums up these architectural differences.
Reuse Strategies: Paper or Plastic, Propellant or Parachute?
If you look closely, you’ll notice a remark about reuse strategy under each architecture. Let’s examine that. Since the low energy rocket stages at a lower altitude, velocity, and distance, it’s much less challenging to propulsively fly the first stage home to be refurbished and flown again. Because of its early staging and shared function with the upper stage, it’s also less impactful to its mass to LEO when we save a large portion of the first stage’s propellant for the return flight. Consequently, it makes sense for a Low Energy (LEO) optimized architecture to choose propulsive fly-back reuse. Of course, when that type of rocket is pushed beyond LEO, it rapidly finds itself without sufficient performance to enable reuse and must often fly as an expendable for the most challenging missions.
The opposite of all of that, is the situation for the High Energy architected rocket. At twice the velocity, altitude and distance down range, flying home is very challenging. And the impact on its payload mass, especially for the highest energy orbits, which is what it was designed to do, is much more severely impacted by reserving the necessary propellant to fly its booster back. That’s because it would either be forced to deliver the upper stage short of its optimal, near LEO altitude and velocity, or significantly reduce the payload mass. Thus, component recovery is a more appropriate choice for reuse in that design.
The high energy architecture’s reuse strategy of component recovery is to expend all of the booster’s propellant for the benefit of delivering Delta-V to the payload, and then separate off something expensive (most obviously the engines) to reenter and be recovered for reuse.
The Cool Toys of High Energy Tech
You’ll also notice by glancing at Figure 2, that there seems to be a long list of unique technologies attached to the High Energy architecture. These are driven by the extreme performance needs of High Energy orbits, but also by the extraordinary duration of these missions. For example, the longer you expose electronics to the much more intense radiation environment present above LEO, the more likely they are to be harmed. Consequently, radiation hardened components and designs are required.
The high demands of energy performance also drive the choice of high specific impulse (Isp, efficient) cryogenic propellants, but the hours and hours of flight time required, threaten significant losses due to boil off. So, very special technologies are employed in the thermal management and efficient use of these propellants, not to mention the endurance of electrical power and other systems.
Last, but not least, the level of complexity of the necessary maneuvers means advanced avionics and an engine that can be started and stopped many times and left idle for hours at a time, in space. All of these, and more, are the unique and specialized technologies required for High Energy missions and, by the way, are not required for LEO missions. Nor are they yet needed in the commercial space marketplace. These are unique government mission needs.
What’s Staging All About, and Is It Worth the Trouble?
I’ve talked a lot about staging and how that is different between the two architectures. But some of you might be wondering, “So what? Why is staging, let alone when we stage, such a big deal?” Well, I’m glad you asked.
Here’s a big secret about rockets. Almost all of what they lift off the pad is just themselves, their own structures and propellants. The actual payload, which is the entire point of the rocket, is typically only 3–5% of the mass that the rocket must lift off the ground! It’s like The Rock (Dwyane Johnson) climbing a giant vertical cliff, expending all the calories required to move that big physique, just to get a paper clip to the top.
The propellant is the heaviest thing the rocket lifts off the pad, but at least that will eventually get burned and deliver some energy to the payload (along with having delivered a lot of Delta-V to the propellant not yet burned at any moment in time). But everything else, the inert, or “dry mass,” is nothing more than “parasitic payload” that subtracts from the useful mass we can get to any given orbit.
Likewise, from the narrow viewpoint of the first stage, everything on top of it, including the upper stage, is mass it must lift. Which is why you can’t always fix a flawed architecture by making the upper stage really big and why the thrust of a first stage, even more than its Isp, is so important for a booster. For illustration, just imagine the extreme case of a booster with such a gigantic upper stage that, together, they weigh more than the booster’s engine thrust. Even though you’ve put a massive amount of total energy in your rocket design, it would just sit on the pad and vibrate… (i.e,. zero mass to orbit).
That’s why we work so hard to minimize inert weight, and, why we really hate having any inert weight that is not absolutely necessary. This leads us to very strong, lightweight structures, built from high performance materials, as well as removing of anything we don’t need. For the High Energy rocket, this demand is even more acute and leads to advanced designs and approaches like the ultra-thin, flimsy, balloon tanks on a Centaur.
You might suppose that a single, big stage that goes from the ground all the way to space must be the most efficient, given its simplicity and clean design, but you would be forgetting that once the rocket lifts off, it starts burning propellant and emptying that big tank rapidly. For example, halfway through flight, the tank is half empty and is, therefore, twice as big and twice as heavy as it needs to be at that moment, and so on as we continue. So, we end up dragging a lot of structure that we needed on the pad, but can’t use anymore, all the way to space. If only we had a way to get rid of the bits we stop needing along the way. Well, we can do that. It’s called staging.
There will come a point in flight when the weight of the extra things on another stage like additional domes (top and bottom) of a second stage’s propellant tank, along with its own engines, and an interstage to connect it to the first stage, weighs less than the empty part of a big single booster. So, by adding this second stage, we can separate from the empty first stage and shed that (now) useless weight.
Staging also gives us the benefit of having a first stage engine optimized for flying in the atmosphere (big thrust, smaller nozzle), and then switching to an upper stage engine better suited for flying in vacuum (high ISP, big nozzle). [For the Venture Star fans reading this, yes, there are very specialized applications where a single stage to orbit (SSTO) solution makes sense, but they are presently rare, and I’ll save that discussion for another time.]
The ultimate choice of exactly when in the flight path we stage, and how big each stage is (the staging ratio), depends on the physics I just described and the destination orbit that you are optimizing for. As I explained earlier, optimizing for Low Energy missions, drives us to stage earlier in flight than if we choose to optimize for High Energy orbits. Any given space launch rocket can fly to orbits other than their perfect, optimal case, they just don’t do so as well.
No Regrets! — The Implications of Architecture Choices
Once you have made your choice of target mission, you’re locked in. You can sometimes make minor adjustments like offloading propellant, but that’s about all. (Unless you have a “dial-a-rocket” architecture like Vulcan, but more on that later). Figure 4 below gives three examples of rocket architectures optimized for Low Energy vs. High Energy: a typical three core LEO optimized rocket, a high energy rocket, and an extreme LEO optimized “super heavy LEO lifter.”
As you can see, the High Energy example has very good performance all the way out to GSO, the typical Low Energy rocket falls off faster as we move right, and the extreme case of the highly LEO optimized lifter delivers a truly massive payload to LEO, but is literally stuck there, being unable, for any payload mass, to lift even its own inert weight beyond LEO.
Definitely Regrets… How Do I Reach Past My Architecture Choices?
Moving away from your optimized target mission space can be difficult without abandoning the architecture, but not impossible. Let’s look at the extreme LEO lifter example for illustrative purposes. We have three basic options to get this beast up and out of LEO: we can figure out how to remove a significant portion of its inert mass, we can refuel its empty second stage propellant tanks in LEO, or we can add a third stage.
Removing a significant portion of inert weight so that there is mass left over for payload may or may not be possible, depending on its configuration. If one can do so, then it’s likely that it was not very well optimized to begin with. Which seems unlikely, but not impossible. If successful, this would give it some capability beyond LEO, but obviously far away from its optimal point.
Refueling in space to compensate for an architecture over-optimized for LEO is intriguing. However, the practicality of this strategy is highly dependent upon where the propellant originates from. If this were done tomorrow, it would obviously be brought up to LEO from Earth. This presents several issues.
To start, we must remember that a rocket’s payload, in this case propellant, is a small portion of its lift-off mass. If we are going to do something useful beyond LEO, that is worthy of using our otherwise highly optimized, massive LEO lifter, we’ll need to significantly fill those empty tanks so they can lift all that inert mass, and a useful payload, to a higher energy orbit. For most cases, that will take several “gas runs.” We will launch one rocket with the payload, then two to six more rockets with fuel, transferring that fuel in LEO, and then proceeding on to the destination.
For example, think of running out of gas on the side of the road while you still have a long way to go on your journey. You would need to be “refueled.” You give a friend a five-gallon gas can to go fill up at a gas station. However, when your friend gets back, she has only brought you one gallon of gas! She shares that the container was full when she left the gas station, but she had to use 80 percent of it to get back to you. While refueling is helpful, it can take many refueling trips to get you where you need to go.
Given that this could be as many as seven launches “for the price of one”, the economics of this architecture will be difficult to close for many missions. The massive LEO lifter is a very large rocket that will have a substantial initial cost to build (with a floor cost, obviously, not less than its raw materials). Reusing it will help, but there will be a limit to the number of times it flies each year because 100mT to 200mT per lift will quickly take a big bite out of the available LEO market, which does not contain an infinite number of launches. Clearly, the vehicle will have a limit to its life as well. So, under these circumstances, it is possible that this can be made to work, but likely for only very specific missions (remembering that all this complexity could be avoided by simply using a rocket that was specifically designed for those high energy missions).
Another consideration is the impact on range infrastructure of this inefficient strategy. Even an occasional use of this scenario would put significant stress on the national launch sites. Turning “one launch into seven”, on a frequent basis, would worsen an already increasingly challenged situation on ranges that are becoming crowded.
Finally, it would be fair to pause and consider sustainability as well. At present, space launch has a massively net positive impact on the environment, with GPS alone, saving a half billion gallons of fossil fuel each day while also feeding 800 million people. However, if a market were to materialize that allowed a sufficiently high launch rate of a super heavy massive LEO Lifter to make its launch cost cheap enough to close on the economics of a “seven for one” refueling strategy, let’s say 100–200 launches per year, that would constitute a massive increase in the total fossil fuel consumption of space launch. Not to mention, another infrastructure challenge at the launch site just to keep up with the demand for propellants. The implications of both would need to be understood.
On the other hand, if the propellant is sourced from space, then the availability and cost of refueling in orbit would be substantially more enticing. The in-space source will be the moon, where massive quantities of water reside, just waiting to be converted into propellant, but that is another topic. Suffice it to say, the sooner we have In-Situ Resource Utilization (ISRU), the better, and the more likely that scenarios like the in-orbit refueling of rockets will make commercial sense.
Another approach to getting our highly LEO optimized big rocket to a higher energy orbit would be to add a third stage. We would subtract a portion of the payload mass and repurpose it to enable the mass of a third stage. The first and second stages would see the same mass above them as before and proceed to do their job of getting the third stage, and the now diminished payload, to LEO before running out of gas. The third stage would then separate with the payload and take it to a higher energy orbit.
Obviously, this would be an “add on” solution to an existing rocket that was originally designed for another purpose. So, it would be less efficient than if it had been purpose built for High Energy from the beginning, but depending on the details of its original design and the market dynamics present, it is possible that it could be made to work.
What About All Those Tiny Rockets?
Now let’s flip to the other end of the spectrum. What about “Micro Launchers”? These are rockets designed to lift 0.5mT to as much as around 1.5mT (i.e., very small payloads). These are, in turn, very small rockets and can be built much more cheaply and launched with less sophisticated infrastructure than a heavy class launch vehicle.
However, as the rocket gets smaller, its ratio of inert mass to propellant degrades. As we scale down the overall size of the rocket, the propellant tanks can be shrunk faster than engines and structural components. Consequently, while the individual cost of the micro rocket comes down with its size, its $/kg to orbit rises. In fact, a small launcher, with an individual launch cost of only 1/10th the cost of a single Heavy class launch, will have a cost per kilogram of payload that is around 10x to 15x higher than a heavy lifter. The physics of volume, mass, and density for the components a rocket dictate this and cannot practically be overcome.
This means that micro space launch vehicles are ideally suited for individual, small, and experimental or demo missions (a modest, but real market), but not for large scale constellations. Nor would it be practical to launch the thousands of spacecraft needed to reach a revenue generating network, in a timely fashion, by launching them one at a time.
Unfortunately, when these were first proposed to investors and appeared in the marketplace, they were touted as being suitable for individual launches of the very small spacecraft that would constitute the so called commercial mega-constellations going into LEO. At one point, over 100 companies existed to develop these vehicles and pursue this market. I was somewhat unpopular when I suggested that the driving economic factor for these constellations would be the cost per spacecraft on orbit, analogous to the $/kg on orbit discussed above. Therefore, these missions would be dominated by the much more efficient heavy class launch vehicles. Time has validated both the physics and economics behind my assessment.
Commitment: Picking Your Perfect Mission
By this point, I’m hoping you have grasped the principle that different rockets are generally optimal for different missions. As a business, one targets a market, and then picks its most common mission, to design a rocket that will be both well suited and competitive across that market. Once you do this, you’re locked in. Payloads can get bigger or smaller. Destinations can be lower or higher energy, but your rocket is, more or less, the same.
You can off-load propellant for easier missions and make other minor adjustments to improve efficiency for the off-optimal orbits that are adjacent to your original target, but your basic architecture is now fixed and your flexibility is limited. You have one perfect mission. Other missions are less efficient for you. Some can’t be done at all. Thus, different rockets do different things, as I have said earlier.
At present, the largest single market is the commercial mission to LEO. Specifically, the commercial mega constellations that are bringing global, high-performance (low latency) internet to everyone, from space. Consequently, most of the current and new launch vehicles are designed to be optimal and most efficient for that mission, at the expense of their high energy capabilities.
But, contrary to everything I have just told you, it doesn’t always have to be that way…
Rebel Rockets: Defying the Rules
I will share with you how we have cheated the architectural rule. The Vulcan design is, in fact, a life hack for the tyranny of having to choose your one perfect target mission. Unlike conventional architectures, Vulcan employs a “dial-a-rocket architecture.” It is, quite literally, eight different rockets within a single architecture. The common core Vulcan rocket can be configured with zero, two, four, or six very large solid rocket motors, at the launch site. This allows the maximum lift-off thrust to be scaled across a factor of 4x. Effectively giving Vulcan four different first stages. Vulcan also has two different versions of its Centaur V upper stage and two different length payload fairings. One Centaur V version is optimized for High Energy and a second version is optimized, within the overall Vulcan architecture, for Low Energy LEO missions.
Rather than being an after the fact bolt-on fix, like adding a third stage to the LEO optimized Lifter I discussed earlier, this strategy is baked into the fundamental architectural design of Vulcan. It has allowed it to compete effectively for the most difficult High Energy missions, being judged as the most efficient, cost effective, and best suited in an open competition with participants from across the industry, while also being able to later compete effectively in the commercial Low Energy LEO market, winning the largest commercial space contract ever awarded. Sometimes, it’s OK the break the rules.
Whew! You Finished The Paper!
If you have made it to the end of this somewhat lengthy treatise, thank you for your interest and attention. Your reward is that you are now privy to some of the subtleties of rocket architectural design that few aerospace professionals and even fewer laymen have understood.
