# What technological breakthroughs were required to land booster stages?

It is only recently that SpaceX developed first stages that can land again, and be reused.

The (until recent) non-existent landing of re-useable rockets is presumably, partially, due to a lack of will (the Apollo program did not care about reusing rockets, just about going to the moon), and partially due to a lack of ability (even if the Apollo program did care about reusing rockets, they would not have been able to do it with 1960s technology - but why not? what was missing?).

I imagine that landing a rocket is a difficult exercise in control theory, and may require significant computing power that was unavailable in earlier decades (let alone the 1960s), but this is speculation on my part.

What technological breakthroughs were required to land rockets?

• I'm assuming you're talking about landing booster stages. As the Lunar Module certainly landed on the moon! (Does that still count as a rocket?) Feb 8 '18 at 12:31
• I think a large part of the technology that enables this today is similar to the technology that made drones mainstream a couple of years ago--that is advanced, compact digital sensors and processing units that can handle thousands of sensor inputs at the same time. I'd say the earliest precursor to automatic reusability was the Russian Buran shuttle which could operate completely autonomously. Feb 8 '18 at 12:38
• There were proposals for recovering Saturn boosters. But the "race" nature of the Apollo program meant that it wasn't a priority. Remember that Saturn booster production was cancelled long before the last lunar landing. Feb 8 '18 at 12:43
• @Edlothiad The Surveyor unmanned missions landed autonomously on the moon. The technology of the day couldn’t guarantee a safe landing on flat ground (modern computer vision algorithms would help), but the basic landing algorithms were there. Apollo LM had a hands-off landing program, but it was never used — Lovell intended to try it on Apollo 13. There’s not much to the Falcon 9 system that couldn’t have worked in the 1970s, albeit more crudely. Feb 8 '18 at 14:56
• @Dragongeek automatic landing is much older than the Buran, having seen production use in airliners in 1968. Feb 10 '18 at 8:09

SpaceX's demonstrated booster-landing ability isn't the result of a breakthrough but rather a bunch of small incremental improvements. The major limitation has been funding and the will to make it happen.

In 1966, unmanned spacecraft landed on the Moon under rocket power in the Surveyor program. It used (IIRC) three fixed-position thrusters, pulsed, to control its attitude and rate of descent. This was a relatively small spacecraft (about 3m tall and massing 300kg), which makes things easier; its landing legs were wide so it could remain stable if it touched down at the wrong attitude. Under the moon's lower gravity, response times are somewhat less critical, so it can be thought of as "easy mode" for autonomous rocket landings. However, it proves that a basic radar altimeter, an inertial platform to determine the spacecraft's attitude, and a simple control loop are sufficient to land with.

In the Apollo program, the LM descent stage was only a little more sophisticated; a throttleable engine on a gimbal mount was used to control direction and rate of descent, and smaller thrusters used to control attitude. The human pilot could, in principle, designate a landing point and then let the computer do the rest of the flying all the way to touchdown, but in practice every Apollo commander took control in a semi-manual mode at around 150m altitude, controlling the ship's attitude and rate of descent. The major limitation at this point was that the autonomous system had no way of knowing if it was coming down on flat ground or on a pile of boulders; the radar altimeter was a single low resolution probe.

Landing a rocket on Earth, under six times higher gravity, requires faster control response: not necessarily a much faster computer (you can manage with only a few thousand operations per second; the equations are not that complex), but things like fast and precise throttle valves. I don't know much about the history there, but I imagine this kind of thing was available in the 1960s as well. On the up side, when landing on Earth, it's much easier to arrange for a large flat area to land on!

As @Dragongeek mentions, the USSR flew the Buran space shuttle in 1988; its only flight was an unmanned launch to orbit and return to Earth. It landed in winged, horizontal flight, so it's not directly comparable to Falcon 9, although the guidance and control problems are similar in complexity.

In 1993, the DC-X project demonstrated rocket-powered, autonomous, vertical landing of a 12-meter tall vehicle from 3km altitude. DC-X had some problems, to be sure, and was eventually canceled for lack of funding, but there weren't any breakthroughs that needed to be made. Like Falcon 9, DC-X used multiple, gimbaled, deep-throttling engines as its primary flight control, augmented with aerodynamic surfaces and attitude-control thrusters, and used GPS in its guidance and control system.

GPS has been a big help to precision landings, of course; the Falcon first stage knows its own location via GPS and guides itself towards a specified point in absolute space -- the center of a landing pad or ASDS. Without GPS, which was first deployed in the late 1980s, it would be possible to place a radio beacon at the landing point and home in on that.

Arguably, the key to SpaceX's success has been that a failed reusable launcher can still be a successful expendable launcher. By making a commercial orbital launcher that is practical to fly in the expendable mode, and experimenting with landings on the payload customer's dime, they've substantially offset the costs of landing R&D.

Modern computer vision algorithms would make it possible to find and target a flat landing site on another planet or moon, which SpaceX may have to do for their first Moon or Mars landings. This technology is a more recent development. It requires fairly modern amounts of computing power to derive a terrain map from camera views in real time, but that isn't required for what SpaceX is doing today.

• Comments are not for extended discussion; this conversation has been moved to chat. Feb 12 '18 at 13:49

One technological improvement that SpaceX does have over Apollo-era NASA is the computer power to do full-up simulations during design.

Read up on the development of the Saturn V F-1 engine - they had to deal with oscillations in the fuel and combustion systems. Back then they actually fired up the engine on a stand, risking blowing it up each time, to test ideas which they calculated with slide-rules. Today they would collect data from one or a few tests and then run computer simulations (maybe on desktop, but NASA and SpaceX do have access to some hefty "mainframe" computers when needed). Eventually you still need a live test, but not nearly so many, and with smart people building the simulations, most of the tests will be "successful". And of course today a few engineers can do all the math on a PC instead of needing teams of engineers with slide-rules. SpaceX has far fewer engineers than NASA employed during Apollo.

NASA was also handicapped by politics and competition with the Air Force's requirements - which influenced some design decisions in the Shuttle program. SpaceX can do pretty much what they want until Elon runs out of money.

Lastly, we stand on the shoulders of the giants who came before us. SpaceX didn't have to invent the turbo-pump, they just had to refine it. The math behind rocket nozzles is now well understood, but design specifics can always be improved. Etc...

• "today a few engineers can do all the math on a PC instead of needing teams of engineers with slide-rules" - or rooms full of women mathematicians. Feb 8 '18 at 22:55
• @FredLarson Yup. My parents have a 1930s dictionary that defines "calculator" as "one who calculates." Feb 10 '18 at 10:53
• Last week I saw an actual F-1 engine at the Boeing Flight Museum, part of the traveling "Apollo Destination Moon" exhibit. For local flavor they added one of the smashed F-1's that Jeff Bezos recovered from the floor of the Atlantic. And standing 2-stories high next to it is an intact F-1 used for ground testing but not flown. *&^%$that sucker is huge! And the Saturn had 5 of them. The smashed engine is in pieces, so we can clearly see the fuel dispenser -- text associated describes the problem of the oscillations and how the fuel "injector" was designed to avoid it. Way cool. May 27 '19 at 17:18 The answer by @RusselBorogove misses the key technical advancement that makes landing boosters possible which is why I am adding this answer. All the 'landing' examples given, amount to solving the inverted pendulum problem + guidance problem assuming no constraints on the control input. The problem is if you want to actually land an orbital class booster, your constraints are difficult and you have to think about optimality. The LM, Surveyor, and DC-X all demonstrated Soft Planetary Landings; they did not demonstrate Optimal Soft Planetary Landings. While these problems have some small passing similarities, they are definitely not comparable. Solving the Soft Planetary landing problem means you have to size your fuel reserves for worst case conditions. Solving the Optimal Soft Planetary Problem means you can size your fuel reserves for the best case. The first gets you down assuming you have a full tank, the second gets you down assuming you have a basically empty tank. Solving the first is trivial; the second is anything but. The key breakthrough SpaceX made is creating an algorithm that solves the Optimal Soft Landing Problem that has guaranteed convergence (it will always generate a solution within a certain amount of processor time). The optimization uses an Interior Point Method (IPM) which means faster processors definitely help. I don't think IPM existed in the 60s and I imagine most optimizers would definitely be way to computationally expensive on microchips from really anything before the 2000s (especially since finding better optimizers that use fewer computational resources is still a very active area of research). You can't land a 1st stage booster just using a soft landing problem algorithm because you will run out of fuel before you reach your destination (even with optimal soft landing boosters still have run out of fuel before landing). You can't just fill up the boosters with more fuel either because then your payload fraction will start dropping to zero. Nitty gritty details of SpaceX's algorithm for solving the Optimal Soft Planetary Problem. Rocket engines have a limit to how far they can be throttled down (many engines can only go down to 40% ish). This is known as a non-convex constraint which makes the whole optimal problem non-convex. As all optimization people know, you can't prove a global minimum on a non-convex problem. SpaceX (although really Lars Blakmore who is the lead of landings at SpaceX) invented something called Lossless Convexification. This algorithm takes a low dimensional non-convex problem and poses it in a higher dimension which makes the problem convex in the higher dimension where an Interior Point Method is used to find the global optimum. The global optimal solution in the high dimension is then reduced and applied to the oringal non-convex problem. Read this paper if you want to understand lossless convexification of the optimal landing problem better. As a side note, SpaceX spent years developing these algorithms with the grass-hopper test vehicle. It took a group of talented people several years to figure out how to solve the booster landing problem. I would also add that during testing there was not enough attitude authority just using nitrogen gas thrusts. It wasn't until SpaceX added Gridfins that landings started working. Gridfins are great because they can provide attitude control without using mass like gas thrusts. TL:DR SpaceX solved the Optimal Soft Planetary Landing Problem which is much harder to solve than just the Soft Planetary Landing Problem which was the actual technical breakthrough + Gridfins. • Well well, my speculation that it was a control theory breakthrough had some truth to it after all! Thanks! Jun 30 '19 at 9:25 • A lot of truth definitely! Some retired engineers forget that software breakthroughs are a thing. Software breakthroughs are often as difficult or more difficult than hardware advances. In their day software often was simple and trivial; that is definitely not the case now. Jun 30 '19 at 19:21 • What about pseudospectral discretization as well? Jul 1 '19 at 5:16 • I think SpaceX uses a direct method for trajectory optimization, although pseudospectral and spectral methods can be more computationally efficient. I think the dimension of the optimal landing problem is still tractable enough that the curse of dimensionalality doesn't hit them too hard. People I know who are better optimization then me, tell the jury is still out on whether direct or indirect methods are ultimately better. Jul 1 '19 at 17:51 • Interesting stuff! Do you have any references describing how much quantitative difference the more advanced optimization algorithms make? I'd argue that the "full tank"/"empty tank" business is a canard; LM and Falcon 9 first stage have very different mission designs, so of course one starts its landing with full tanks while the other starts with near-empty ones. The optimization makes a difference to the size of the F9 first stage, and might even make the difference between economic practicality and not, but it's not the difference between "possible" and "impossible". Aug 23 '19 at 18:36 Adding to Russell Borogove's answer, I would also argue two additional points: • On the technical site, the engines have been designed with powered landings in mind. They can be fitted with the necessary hardware to restart while in flight easily. Also, the Falcon 9 was the first rocket to use 9 small engines, allowing for a powered landing on just one or three engines running at high efficiency (depending on configuration). • Economics plays a large part, too. SpaceX provides launches cheap enough that customers are willing to fly on a rocket that doesn't use all of its fuel for their mission. They in turn can do this because they've reduced cost internally significantly. On the other hand, the Apollo program was plenty expensive as it is, as are other large projects like SLS. Can you imagine what would happen if someone suggested SLS be 30% larger than needed for payload delivery at taxpayer expense for a feature that may or may not pay off? The same goes for private companies like ULA. Spaceflight isn't that large of a division compared to the whole of Boeing and as long as they can sell the same simple design over and over again, they have no need to pour their own development money into it. This is also illustrated by the solid rocket booster fallacy. SRBs are incredibly cheap to develop and use a few times but very expensive long-term - the boosters on the shuttle cost something like \$50 million a launch. An Atlas V has up to 5 boosters at \\$10 million each, costing almost as much as an entirely reusable Falcon 9. However, these designs were chosen as there was little upfront cost and the customer is willing to pay for these with each launch.

• You should edit your answer - "wall of text" is hard to read (split it into bullets or paragraphs) , and making key phrases/words bold improves readability. +1 for effort, good first answer and welcome to SX. Feb 9 '18 at 19:37

What technological breakthroughs were required to land rockets?

Parachutes.

And they existed since the 19th century.

The other answers refer to the powered, vertical landing of booster stages, but landing rockets with parachutes was done since the start of the space race. I feel that parachutes deserve a mention here, so let me quote https://www.nasa.gov/missions/research/f_sounding.html :

Since 1959, NASA-sponsored space and earth science research has used sounding rockets to test instruments [...]

In most cases, after the payload has re-entered the atmosphere, it is brought gently down to Earth by way of a parachute and is then retrieved. By recovering parts of the payload, it can be refurbished and flown again, resulting in tremendous savings.

Recovering booster stages was also done, up to some degree of success. Let me quote from https://www.nasa.gov/sites/default/files/files/5.pdf :

The first Decelerator Subsystem, which included a clustered assembly of the three main parachutes, a drogue and pilot parachute assembly, and load cells and fittings, arrived in November 1978, for installation in the first assembled SRB.

[...]

Following the successful launch of STS-1, three significant issues related to SRB hardware reusability were identified during the post-flight assessment: aft skirt ring structural integrity, aft skirt internal reentry temperatures, and electrical cable salt-water intrusion.

[...]

The first flight scheduled to fly refurbished hardware (other than parachutes) was STS-7 using STS-3 hardware; the parachutes were scheduled for reflight on STS-4.

So the NASA shuttle program was already landing, recovering and refurbishing booster stages by 1983 (maybe even earlier, I haven't checked all the documentation). The technology was already there.

It's worth noting that SpaceX's recovered stages also undergo refurbishment after a landing. But it should be obvious that a parachute landing means more damage (and higher refurbishment cost, and less reusable surviving parts) than a powered, vertical, controlled landing.

I also would like to mention the Ansari X-PRIZE lunar landing challenge and Lunar Lander Challenge. The developments there were instrumental to enable fine control of powered ascents/descents.

• NASA's approach with the SRB was a dead end, however, refurbishment (due to salt water getting everywhere) was almost as expensive as building a new one. Feb 9 '18 at 11:54
• @Hobbes I thought it was mainly because "refueling" SRBs meant casting the solid fuel into disassembled segments which is almost like building a new one? The saltwater sure did not help though. Feb 9 '18 at 12:05
• I'd have to check, but IIRC they had to remove the insulation on the inside of the steel tube and reapply it. And all the hardware had to be stripped to bare metal and rebuilt. Casting could be part of it too (it'll take a while for 150 tons of fuel to solidify). Feb 9 '18 at 12:45

In another answer, Russel Borogove makes a valid point that zero technological breakthrough needed for landing but instead is just incremental improvement.

Much of space technology stayed same through mid 1960 to present but a very small point has been included in a reverse engineering fashion to make rocket land back on surface or ocean.

**at 0:02/0:08 : "Lateral thrust" is imployed, which turns it to right.

at 0:03/0:08 : It is stopped from going further right.

at 0:04/0:08 : Full thrust has been introduced to give it the properspeed for the right side direction.**

Now this technique has been made public in late 2010 and these practices has been in action by Russia since late 2002 or 2003.

Elon musk and SpaceX have re-used "lateral thrusts" from Russians into their landing rockets and after several tests, rocket landed back with just addition of small step in a reverse engineering fashion.

Further Russell Borogove elaborated about computer vision. But really there is zero breakthrough and 100% copy/paste from allied disciplines of rockets and missiles.

Control theory has no deep applications in Elon musk's spacex, but what they do is copy/paste from other countries allied or core disciplines and do experiment to keep success in public eye and whatever they show to public, has been well experimented by other nations a lot long earlier.

• Lateral thrust is not new. All rockets since the R-7 used to launch Sputnik 1 use reaction control rockets to change their attitude. Its application to military missiles is more recent (most use vectored thrust or aerodynamic controls instead). Feb 19 '18 at 15:48
• And lateral thrust is not the primary control method for the Falcon first stage, either. Feb 19 '18 at 17:25
• yes, it not primarily but its a decisive one, to give it a turn while going down and its been employed with all other allied mechanism like speed down etc.. Feb 20 '18 at 7:03

I have read that the first stages of Soyez boosters are so rugged that they sometimes are reusable after crash landing on the ground. If that is correct one very simple way to make reusable booster rockets was and is simply to overbuild them and make them much stronger than needed, making them less efficient and unable to launch as heavy payloads, but able to be reused with little effort.

Another method used for a long time is parachutes. Combining parachutes with ruggedness might possibly have resulted in reusable booster stages at an early period.

The Pegasus launch vehicles are launched from airplanes, with the first launch in 1990, and carrying small satellites into low Earth orbit.

Airplanes, of course, are usually intended to be used for many flights, so these airplanes count as reusable first stages.

A rockoon is a small sounding rocket launched from a weather balloon. Of course the balloons are not recovered and reused but weather ballons are cheap. Rockoons have been launched as early as 1955.

More recently, the JP Aerospace company has developed and used rockoons as part of its space access plans.[6] Additionally, Iowa State University and Purdue University (Purdue Orbital) have started programs to develop rockoons[7][8] and significant work has been recently done by Leo Aerospace based in Los Angeles and a Romanian space company, ARCASPACE. The Spanish company zero2infinity plans to launch a toroid shaped rocket from a balloon called Bloostar in 2018 to carry micro satellites to low earth orbit. UK base company, B2Space is developing the concept to launch small satellites into low earth orbit [9]. Stofiel Aerospace is also aiming to launch Cubesats with their Hermes Rockoon System.

Note that an airship can be steered, though no airship has reached the heights that whether balloons have.

A long track on the ground with a rocket sled could accelerate a rocket plane and thus give a rocket plane a head start in reaching orbital velocity. Thus the track and the rocket sled could be considered a reusable first stage.

Smaller sized rocket sleds have been used to test equipment as early as 1954 in the USA, and there are rumors the Germans used a rocket sled to launch a rocket on 16 March 1945.

Thus it seems possible that routinely reusable booster stages could have been developed decades ago if there had been sufficient will to do so.

• Soyuz boosters are used for scrap metal after hitting the ground. They are not reused as boosters. Feb 9 '18 at 17:53