81 thoughts on “The Dawn Of Commercial Spaceflight”

  1. Rand, you’re talking like no one has ever reused a launch vehicle before. The first time that happened was in 1981 with STS-2. This is the first time the booster for an unmanned flight was reused. The refurbishment for the Falcon 9 is obviously much simpler, much less expensive than for a shuttle. That’s called building a better system based on lessons learned. Although, obviously, it’s going to be a long time before we build an actual spaceship again.

    1. Michael, while each of your statements about Shuttle was correct, you were not replying to Rand’s statement. His statement was about making spaceflight cheaper through reuse in an entire market system interested in making spaceflight cheaper. The politicians in Congress making NASA’s decisions, in 1971, in 1981, or 2011, or now, had and have no use for making spaceflight cheaper, so their designs didn’t, either, even though that was the original concept for the Shuttle.

      What happened yesterday was the workings of far more than just a rocket, but of an entire market beginning to free itself of the feudal domination of Congress, to fly to the stars. *That* was what was historical about it.

      It need not be long before we build another spaceship again, since another set of engines for the next one just came off the factory floor to be tested for powering the next, called New Glenn.

      1. Space is getting back to how it was supposed to have happened, hadn’t WWII, and the Cold War, gotten in the way.

    2. Are you REALLY sure it will it be a long time? SpaceX may do it in less than a decade with the ITS spacecraft. If it turns out anywhere close to what we have been told, it will be far superior to the Shuttle orbiter in capability.

      1. I wouldn’t expect that to happen so soon. The ITS design is kinda bonkers and it uses shittons of composites, something which SpaceX and everyone else have little experience with.

        1. Composites for structures are certainly no longer anything resembling unknown and exotic. Carbon fiber composites have been used in aircraft, boats, cars and other transport applications for decades. SpaceX vehicles have had large carbon fiber composite parts in them since the company’s inception.

          Use of such materials in cryogenic propellant tankage is, admittedly, a bit more recent, but there are several firms that seem to have figured this trick out. SpaceX is now, probably, foremost among them.

          And I would call the ITS more bold, than bonkers.

          1. My problem with the ITS is that flip-around maneuver they want to do in the Mars atmosphere. That requires the center of mass to shift from fore to aft of the center of lift in the middle of the maneuver to maintain aerodynamic stability. Tricky. Not impossible, but tricky, and if they screw that up even a little the ITS will land nose-first like a lawn dart.

    3. Yeah I’ve heard this conversation before. But if it was so easy, so obvious, why didn’t anyone do it? I mean I was reading about the German DLR Hopper proposal two decades or so ago, a proposal to have a reusable first stage and an expendable upper stage, with landing in the ocean (on an island but still), however in their mind it had to have wings, and it had to use LOX/LH2, etc. They made it a mess and complicated everything. DC-X is arguably what came closest to this vision but it was basically killed in the crib. I remember all sorts of stupid arguments in the 1990s on why not to use VTVL, from that it was too easy to do (and therefore not worth it), to that it was too hard to do (and therefore not worth it). Totally bonkers.

      The comment that it has “been done before” is kind of besides the point. It reminds me of quotes from the time of the computer revolution. I have a couple Ken Olsen quotes about that here:
      “There is no reason for any individual to have a computer in his home.”
      “On almost anything someone does in the computer business, you can go back in the literature and prove someone had done it earlier.”

      Now Ken wasn’t a dumb person. Not at all. He and DEC were largely responsible for the mini-computer revolution but he still couldn’t see nor adapt to the new paradigm properly.

      The fact is most of the cost in a liquid rocket like the Falcon 9 is in the first stage. It’s that thing about staging, each lower stage is roughly an order of magnitude larger than the next stage. The design is smart enough that they basically reuse the same engine design and electronics for the upper stages. Which is basically a feature similar to the Soyuz rocket which is arguably (for now at least) the most successful rocket in launch history.

      SpaceX has demonstrated first-stage barge landings (not done before), first-stage return to launch site (not done before), and now first-stage reusability. Eventually they’ll get the costs of reuse down, because that’s how they operate, they’ll iterate on the design, they won’t sit on it.

      The Falcon 9 is simpler than Shuttle on some regards and more complex at others. You should remember that Shuttle couldn’t land automatically on solid ground (it had to be piloted); let alone in a floating barge on the middle of the ocean.

      A successful reusable design will require many of these little advances to be made until the large qualitative leap occurs. This is something SpaceX is willing to do and is part of their business culture, and it’s what the conventional space sector mostly forgot how to do.

      1. They only reason the Shuttle couldn’t land on a barge in the middle of the ocean is that they never asked Neil Armstrong to do it.


        1. Heh. There’s no denying Neil’s skill with the controls. I remember trying to play Lunar Lander and it was anything but easy. Which I think makes the achievement of automating that procedure all the more outstanding.

      2. “But if it was so easy, so obvious, why didn’t anyone do it?”

        Sometimes you just have to be at the right time & place where events, knowledge, materials, resources, etc. come together.

        Just like the reason the model T ford didn’t have shatter proof windscreen, airbags, seatbelts, etc.

        1. I’ll note that SpaceX wouldn’t have been viable any time before 1984 due to the Space Shuttle monopoly. And afterward, there was extensive market stratification (which I believe to have been enforced by both NASA and Congress) which mean that the Falcon 1 and 9 would be entering niche monopoly markets with possibly a hostile NASA to deny them government contracts. It wasn’t till the early 2000s that the EELV competition broke up this stratification and allowed SpaceX the opportunity to enter.

          1. The interesting what-if is if the shuttle had not been pushed by Nixon to get California votes in 1972. In retrospect he didn’t need those votes (McGovern crashed pretty spectacularly), US expendables would have advanced faster and would arguably have been able to compete better against Arianespace.

          2. Karl, there was also the vicious hostility of NASA turf warriors, toward *anyone* who might undo their monopoly on spaceflight in the US, to take into consideration. In late 1979, Max Hunter was approached at a conference by Robert Meuller, of NASA HQ, to join the whispering campaign against Space Services Inc. He refused.

            In spite of his long history leading aerospace projects, from Thor onward, Max never led a successful bid with NASA again. In 1983 there was massive pushback, inside NASA, against Reagan allowing contractors to use the launchers developed under government contract, which NASA had long since gotten their development costs back from, to launch satellites privately. This pattern was repeated in 1985, with Pan American Spaceways, and continued right through the Challenger disaster until Columbia also bought 7 farms for astronauts.

            There was a 5 years hiatus, in which SpaceX got established with a bureaucracy in their favor inside NASA with the 2008 COTS contract. That’s the only reason the knife work hasn’t been sharper and deeper.

        2. “But if it was so easy, so obvious, why didn’t anyone do it?”

          Because there were congressional and military roadblocks in place to insure that it DIDN’T happen?

      3. Well, Shuttle *could* have been landed on autopilot. Especially later on in its history. NASA chose not to do so.

    4. The Shuttle wasn’t the first stage. The first stage of STS was solid rocket boosters. They weren’t really re-used, because by the time they spashed down (not “landed”) they were only empty metal tubes. The engine part of an SRB is the propellant mix, and that had to be completely re-cast every single time. And, because they splashed down in corrosive sea water, they had to be gone over with a fine tooth comb, the O-rings between the segments replaced, the parachutes replaced… in short, the SRBs had to be completely re-built.

      Before the SES-10 launch, there was only ever one attempt at re-using a first stage, a splashed-down Titan III, and that launch didn’t make orbit.

      1. SRB’s are like the brass cartridge case in a reloaded cartridge.

        They were reloadable more than reusable.

      2. When you quibble about SRBs being the first stage, you are missing the broader point … the SSME fired from the ground all the way to orbit. Both they and the Shuttle structure demonstrated rocket/booster/engine reusability as a fact.

        One can argue that the SSME’s underwent massive teardowns and rebuilds after each flight (true), but the point remains that reusable, orbital rockets were an engineering reality demonstrated by STS.

        SpaceX has now built a system that seems to be on the way to demonstrating quick, economic booster reusability.

        1. Godzilla helpfully provided the name of the paper a few days ago, but I don’t know if it’s online:
          Dunn, J., “Titan IIIB Recovery Experiment,” SAE Technical Paper 670399, 1967, doi:10.4271/670399.

          1. My original comment was just “Thank you!” but WordPress said that was too short, so here’s some additional text. Thank you!

    5. I think of STS as a partially reusable launch vehicle. Obviously the External Tank (the physically largest component) was thrown away; but even the reused components – the SRB’s and orbiters – had to be so extensively refurbished that it’s a close call as to just how “reusable” they really were. The Shuttle orbiters were probably the closest comp to what is being done with Falcon 9, since the complete engines were reused for multiple flights.

      Now, some will say Falcon 9 is not completely reusable, since the second stage is thrown away. But as Musk says, the first stage is 70%+ of the flight cost, it’s not much of an argument – and it will be even less of one if he can actually succeed at his new ambition of reusing the second stages, too.

      Either way, what matters is whether it succeeds in significantly dropping launch cost and raising launch frequency. Shuttle failed miserably at both. If March 30, 2017 is a historic day, it will be because SpaceX will succeed where the Shuttle failed.

      1. Exactly right!

        I read an interview of Musk, where he described the extreme care SpaceX took with the reused Falcon 9 1st Stage, where almost the only remaining components were the airframe and engines. Understandable considering how experimental the flight was.

        In that sense this 1st reuse of the Falcon stage wasn’t really different in kind from the ordinary flight operations of the dead STS. But this mission was just one step towards the goal of cheaper flight, not the endpoint.

        The STS was a dead end towards cheaper flight operations, foredoomed by politics and engineering compromises. The Falcon 9 today is most of the way towards 70% cheaper operations. By the end of 2017, I imagine SpaceX will get there.

    6. Given that the Shuttle threw away two SRBs and a yuge fuel tank with every launch, I would hesitate to call it “re-using” the vehicle for another launch.

      I would further hesitate doing so considering the Shuttle’s rocket engines needed to be more or less rebuilt after every launch, not to mention the maintenance & refurb done after every mission done to the Shuttle itself.

      It was only re-useable in the broadest sense of the term.

      1. “I would further hesitate doing so considering the Shuttle’s rocket engines needed to be more or less rebuilt after every launch, not to mention the maintenance & refurb done after every mission done to the Shuttle itself.”

        That’s actually not correct. It might have been true at the start of the Shuttle program but with progressive engine upgrades they became more reliable and the mean time between overhauls improved. I quote this document:

        “Block I and Block II Engines
        The next major step in engine advancement was replacement of the high-pressure turbopumps in order to meet NASA’s goal of increasing the period of time between overhauls by flying ten times without removing the turbopumps. Pratt & Whitney was selected to provide redesigned alternate turbopumps. The primary objective for the turbopump redesign was to eliminate failure modes and vulnerabilities in the heritage design. Some of the turbopump parts were originally built by welding together forged segments. These welds were expensive and time-consuming,
        and caused a lot of problems. Accordingly, elimination of the welds was a key specification in the Block I and II SSME design.

        Certification testing on the new Block I configuration SSME was completed at SSC in March 1995 (Figure No. C-17). The new turbopump was designed for a life of sixty missions, and certified for ten flights without inspection, overhaul or maintenance.”

        Had all the future planned SSME upgrades been done after that the engines would probably have lasted like a hundred missions with more time between overhauls.

        I think it’s important to stress this so that people don’t take the wrong conclusions about why the Shuttle was uneconomic per flight. The SSME wasn’t the issue.

        I remember seeing the costs at one time and I think the ET (External Tank) and the OCS/RCS (hypergolic) refueling were substantial parts of the per flight cost. The solids weren’t that cheap either. The ET became especially expensive as a portion of total costs with later upgrades to make the tank lighter to reach the ISS (namely they switched from Al to Al-Li tank construction). The tank was discarded on every single flight. Unlike the SSME and the TPS costs, which kept decreasing with upgrades, the ET and OCS/RCS kept getting more expensive. The ET because of the upgrades, and the OCS/RCS fueling because of the increased regulation costs for handling toxic hypergolic fuel.

        The TPS was also an issue to a large degree because the maintenance wasn’t something that could be easily automated and it added a lot of parts to the system:

        Although TPS costs also improved as time went by (sorry that I can’t find a better link, I read this info years ago) with the further use of blankets in low temperature regions, plus the incremental improvements made to the tiles. Further improvements would have required replacing the orbiter design. e.g. to use metallic TPS the Shuttle couldn’t have remained with the same aluminum structure.

  2. “But on Thursday night it was hard to doubt the immigrant from South Africa who had taken on the bluebloods of the global aerospace industry with a vision of low-cost reusable rockets—and had just validated that entire plan before the world.”

    African-Americans can be proud!

  3. That flight was the beginning of the true RLV era. Now SpaceX can begin the process all of us RLV people dreamed of beginning: seeing what it takes to turn a vehicle between flights. That is the crucial learning that can’t be predicted, only experienced. It marks the point at which one has a real reusable launch vehicle.

    You’re right, Rand. Putting this moment in historical context should be done carefully, because it deserves such treatment. I, for one, think it is monumental.

  4. Whether this moment will be considered monumental will depend upon it being a founding one. That we won’t know that for sure until reuse flights 2, 3, 4, … until it literally becomes a gas and go operation. To mix a metaphor, as Tom Leher once pointed out in his song New Math, as with New Space; it’s the process that counts not the final answer… And once we achieve that process, the moniker ‘New’ will of course be long, long gone.

    Many of us are counting on that….

  5. I’m still of the opinion that a horizontally oriented (rocket on its side) vertical launch and landing will ultimately be a major improvement on tail sitters. The extra engines don’t add much weight. You get massive redundancy (two separate sets of engines), and a return-to-pad abort capability. The rocket can be moved around by a high-school graduate with a tow-truck, and upon return the stage is already horizontal so no extra equipment is needed to mate a new second-stage and payload. All the systems that need to be maintained are easy accessible, much like an airliner instead of a 15 to 30 story grain silo. Regarding launch, the elaborate tower and hold-down clamps disappear, replaced by a concrete pad with engine exhaust spread out along the length of the rocket, not just at the base.

    We’d probably do this already if the military had ever had a requirement for a large, instant-launch road-mobile solid-fueled ICBM that uses small solids to toss it into the air, followed by a 90-degree pitch up just prior to main engine ignition, instead of having the missile aimed upward prior to launch with a slow hydraulic erector.

    1. The reason for pitching up is interesting since it then has to pitch down to achieve orbit. The reason is to get out of the atmosphere as quick as possible, but what if they went through the atmosphere slow using a reusable oxygen breathing stage?

      The rest of the launch industry, he said, is shaking in its boots.

      Elon’s forcing function. The rest of the industry would probably have continued to stagnate if not for SpaceX competition. DC-X should have been a turning point and they killed it ASAP.

    2. The weight penalty of making the vehicle strong enough to withstand thrust in two axis rather than just one, plus the weight penalty of additional engines make this unworkable. At least, unworkable with rockets. Some science fiction drives might make the weight penalty irrelevant, but as it stands now, there’s little enough payload fraction at liftoff already.

      1. Actually the airframe loads might be lower. Consider that you don’t have to use a few big rocket engines, as you do at the base due to space limitations. You can spread much smaller engines out along the rocket’s length to distribute the stress, just like a bridge over a swamp with closely spaced supports.

        The added weight of the engines, based on using Merlins (180:1 T/W ratio), is going to be less than 1% of the liftoff weight, and that weight is almost all on the first stage. In return, you get to lighten the landing gear since you don’t need really long struts that can stabilize an empty tail sitter in pitching seas and moderate winds.

        And of course another obvious bonus is that it’s far, far easier to hover a long rocket horizontally than balancing it on its tail. Heck, Harrier pilots fly that way, by hand, routinely.

        And the idea should really pay off if you put all the horizontal launch engines on two side boosters (similar to a horizontal Falcon 9 Heavy), with the two boosters doing all the initial lift, which means your core stage isn’t even any different except for strengthening some attachment points. The two side boosters could also be given chines like the SR-71 to better cradle the core, while allowing much better atmospheric fly back performance.

        This would all come into play as the rockets get bigger and taller, because construction and maintenance costs go up dramatically with height. A 550-foot tall rocket would be a nightmare. A 550-foot long ship is a common destroyer. The difference is the orientation.

        1. The exhaust impinging on the sides of the rocket in this configuration seems to me to be a bit concerning.

        2. I don’t think it’s that easy. More engines means more plumbing, which also increases weight. Fuel sloshing might become more of an issue than with a wider tank. Then there’s drag. Launch pad construction also become more expensive because you need a larger fireproofed footprint.

          I agree that it might make sense to make the rockets wider than they are now. Especially the Falcon 9 FT, it seems ridiculously long to me, perhaps you could use spherical tanks and something more akin to the VTVL designs of the 1960s. Especially with the LOX/Kerosene rockets, you could have a spherical LOX tank with a toroidal Kerosene tank below it. But I guess its easier to manufacture the cylindrical tanks, and they’re also easier to change dimension (make them shorter or longer) as required to change volume unlike with a spherical tank.

          1. Those are good arguments, length vs. the width. However, I remember reading something to the effect that Elon and his team picked the diameter of the Falcon 9 to be as wide as possible and still enable easy transportation by truck everywhere it needs to go, California, Texas, Florida.

            They’ve gained a lot in terms of low costs, low prices, partly due to this easy transportation (and many other factors).

            Anybody else have ideas on this tradeoff?

  6. Could we have done this 6 decades ago?

    The landing algorithm makes use of an interior point method for solving a convex optimization problem. These algorithms weren’t around 60 years ago; they got their start in 1984 with Karmakar’s algorithm for linear programming. I suspect the computing power needed, even if the algorithm were known, would have been beyond what was practical in 1957.

    The welding technique (friction stir welding) for making the Falcon tanks and structure wasn’t around then. It was invented in the early 1990s.

    1. Friction Stir Welding has been around for longer than that in another form. They used to put missiles together by spinning an entire launch stage while holding a warhead still, and bringing them slowly together. The whole seam got welded at once when they touched.

        1. And also, friction welding != friction stir welding. The former involves welding two workpieces that are rubbed against each other; in the latter the workpieces are clamped into fixed positions.

  7. Paul D.:

    I know that “reference, please” is a hackneyed expression on the Web, but do you have a link or a citation to a paper on the landing algorithm? I have access to online journal subscriptions through my employer.

    You remark about computing power was something I was about to say. I am curious, however, about the advances in algorithms required to make this happen as I had thought that Optimal Control as a paradigm for solving trajectory optimization problems had been around since the Dawn of the Space Age in the late 1950s.

        1. Just as an odd thought, could you solve the problem backwards by assuming that the rocket takes off empty, gaining instead of losing fuel as it flies on a course that takes it to its current position, orientation, and fuel load, with atmospheric drag operating in reverse? Then the problem would simply be flying the pre-computed course backwards.

          1. The issue is computing the path that minimizes some cost relative to constraints. It’s not just solving a differential equation, since you are allowed to vary control inputs. Moreover, the problem has to be dynamically solved, over and over again, during the descent as random environmental effects perturb the trajectory.

          2. Oh, you can control the rocket that way alright. It is just that the resulting trajectory may use up all of your fuel before you land.

            The art in what Paul D. is referencing is controlling the rocket within the limitations of the max and min rocket thrust, the angles in which you can point the rocket, and not running out of fuel.

        2. One thought occurred to me.

          There are two sets of videos of tail-first rocket landings: One is of the SpaceX effort and the other is Blue Origin’s New Shepard.

          Both have the rocket barreling in as if it were to crash and then at the last possible moment, the rocket thrust slows and straightens the rocket for a soft landing. The SpaceX landing, however, appears steady whereas the New Shepard appears to wobble back and forth a bit before it straightens out for the landing.

          Do you suppose that Blue Origin is using a less sophisticated control algorithm?

          1. SpaceX’s vehicle used to tumble around until they added grid fins. Does New Shepard have those?

    1. The provenance may go back to the beginning of the Space Age, but landing a tail-sitting booster like that requires real-time processing & reaction. There’s no go-around for a second try.

  8. Doing this in the 1960s with a 9-engine version of Saturn I first stage would probably have required an RC pilot. Otherwise probably doable. Or a 7-engine Proton?

    1. I don’t think it could have been done until recently, because the computing power didn’t exist. Also, as I understand it, the droneship maintains position at a precise GPS coordinate, and the stage aims at that coordinate. So it couldn’t have been done before the GPS constellation existed.

      Perhaps it could have been done with an RC pilot, as you suggested, but landing a small piloted vehicle on the airless yet solid surface of the Moon would have been a great deal easier than landing a large rocket stage on Earth, while compensating for wind and waves. Could it be done at all in cloudy weather if the RC pilot didn’t have a clear view of the droneship?

      Also, see this charming and lovely cartoon. Click on the thumbnails to view them at full resolution. I’d link Abby Garrett’s tweet, but I don’t know how to make the pictures large enough to be fully readable.

      1. I think a radio homing beacon would have worked, and would have helped the RC pilot gauge wind gusts, along with radar. Military pilots were landing helicopters on ships at sea when I was a kid in the 1950s. Not easy, but plausible.

    2. If I remember correctly, the winged Saturn V first stage plan would have had a crew on board to fly it back to KSC. So why not put a crew on the recoverable first stage and have them fly the landing?

      1. Mainly weight. The RC pilot and his gear are on the ground. I recall autopiloted glide back and tow back proposals for S-IC as well. But a Saturn I RC propulsive landing would likely have worked, leading to a different future from the one we’re living in. Maybe faster evolution of computer control, too. Kerosene under the bridge, now, a 50 year delay, so I’m 66 instead of 16. Oh, well.

    3. With the high acceleration landing used by the Falcon-9 first stage recovery you need better reflexes and precision than a human pilot can deliver. Add the latency of an RC control link and it gets even harder. Hitting the landing point isn’t the hard problem. Hitting close enough to zero speed at zero altitude is the hard part.

      1. True, but we’re not talking about Falcon 9, we’re talking about a hypothetical 1960s stage. Hoverslam is at least partly required by Merlin characteristics.

  9. Just out of morbid curiosity, was there any potential use for the shuttle external fuel tank back in the day? I seem to recall that in the time there was some possibility of using the tanks as some sort of orbital space station, in addition to ISS. But NASA never pursued it, or so I’ve heard…

    1. There were a bunch of proposals (you’ll find some on NTRS, but I don’t have the document IDs handy), but obviously none that actually happened. They varied from putting an airlock on the bottom of the ET so it could become a ‘wet workshop’ to melting them down in orbit and reusing the metal to build new hardware.

    2. http://www.spaceislandgroup.com

      Their problem was the spray-on foam insulation. NASA was rightly afraid of outgassing from the external tank. Problem was, SIG couldn’t demonstrate a fix that would allow them to reuse the tanks in orbit unless NASA let them do it, and NASA wouldn’t let them try until that solution was demonstrated. Chicken/egg.

      1. That wasn’t the only problem. The Super Light Weight external tank weighed 58,500 pounds. That exceeds the heaviest payload the Shuttle ever carried (50,162 pounds for the Chandra/IUS) by a lot, and doesn’t count residual propellants. Even if one assumed that Atlantis could have been stripped down some, and had been able to achieve the 65,000 pound payload originally advertised for Shuttle, they would have to have been below 6,500 pounds of residuals, or 0.4%, in order to get the ET into orbit. The loading uncertainty is greater than that.

        1. The ET was taken almost all the way to orbit anyway, even when the Shuttle was carrying payload. It’s not like putting 1 lb. of ET into orbit would have cost 1 lb. of payload.

  10. STS, as has been said, was “partly reusable.” But the nature of the system meant that it was absurdly expensive, because it required a large standing army for refurbishment. The big thing for F9 is that it apparently will not require extensive refurbishment, and should lead to lower prices. The key difference, to me, is that STS did not significantly reduce costs, and a reused F9 apparently will reduce them by about 30% over a new rocket. If SpaceX can recover and reuse the payload fairings (there was a successful demonstration of guided reentry and parachute deployment on at least half the fairing on this recemt flight) that should allow a further reduction in price. Let’s be even more dramatic: I’d suggest STS was a failed experiment. Clearly it didn’t result in learning that lowered prices, look at the immense development cost of CxP and SLS; and the orbiter, the reusable part, did not beget any offspring. F9, by comparison, is paving the way for F9 Heavy and ITS.

    1. As noted above, the ET weighed 58,500 pounds. The SRB nozzles weighed 23,965 pounds each, and were replaced each flight. So were the case liners, which weighed in the neighborhood of 40,000 pounds each. That’s a total expended weight of 186,430 pounds. With the exception of the Titan IV and Saturn V, that’s more expended mass than any other expendable launch vehicle the US ever flew.

    2. “and a reused F9 apparently will reduce them by about 30% over a new rocket.” Nah – that was a pure guess. Depends on how many times a booster can re-fly, and how much refurbishment. Neither is known yet, tho SpaceX knows more than the rest of us.
      We’ll see how get they can get at it.

  11. It’s worth noting the per flight recurring costs of STS were fairly low, less than contemporary expendables. You get the giant costs by dividing the entire manned spaceflight budget by the annual number of flights. STS failed for non technical reasons.

    1. William, if the cost of a contemporary expendable was in the range of $80 to $180 million, I’m not sure there is anyway to slice and dice the STS numbers to say that the recurring costs of STS were fairly low in comparison. I suspect the STS numbers were purposely buried in the entire operating budget of the manned spaceflight operations because they were so appallingly bad. If the technology chosen for STS required a hugely expensive standing army to keep the fleet operating, that is indeed a failure of technology.

      1. The way you slice is with the STS-83 reflight costs, which were well under $100mln. That’s the stack and fly cost. NASA’s $450mln average published in 2011 includes mission prep, but not the cost of maintaining KSC. SpaceX’s three pads are on maintained bases and I don’t know how much rent they pay, but it’s unlikely to be a pro rata share. Boca Chica will tell that tale.

        1. The marginal costs were admittedly not great. But to be fair, you have to really divide the program costs by the number of flights, which will then amortize the development costs, include the costs of the standing army (i.e., the people who had to get paid whether or not any shuttle flights occurred), etc. According to a space.com article from a while back (http://www.space.com/12166-space-shuttle-program-cost-promises-209-billion.html) the per-flight cost of STS was actually about $1.6 billion in 2010 dollars. Well-meaning people can argue, but that’s probably pretty close.

          The kicker is this: given the great expense of CxP/Orion/SLS to date, plus the built-in standing army cost (let’s be fair, this is why SLS is called the Senate Launch System) I fully expect each SLS flight to cost integer multiples (n>1) of $1.6B. That’s enough to fly 17 F9 Heavies, given the pricing on SpaceX’s website. When SpaceX gets reusability really worked out, the price will come down even more.

          1. Much of the cost assigned to STS by detractors can arguably be said to be the social cost of having a space program. Highway taxes alone don’t pay for the Interstate, either. As far as I know, Boca Chica is the only large private launch complex ever built. Does SpaceX have to pay back a part of the 39A ground structure? How many pilings are that under the pad? In a way, you could look at it as a subsidy. NASA’s $450mln figure is probably a good compromise. But I agree, the SLS development cost is crazy.

          2. William, I have trouble with the argument you are making there. The problem is it could be used to justify just about any outcome of the shuttle program. It proves too much.

            With the shuttle, the human space program became the shuttle program (at least, until the station was up, and it was there to justify the shuttle as well). So just saying “the shuttle is justified by having a human space program” is just circular. “The shuttle is justified by having a shuttle.”

            What would have been lost if the human space program had instead been shut down for a decade or two with the end of Apollo? Very little. Well, unless you were one of those feeding at that particular trough.

Comments are closed.