Business, Economics, Space, Technology and Society

Reusing Rockets

33 Comments

I love the opening of this piece from James Dean:

In more than 65 years of launches from these shores, a rocket landing anywhere near its launch pads meant something had gone terribly wrong.

Not anymore.

Nope.

[Update a few minutes later]

He quotes Elon as saying that the rocket “costs” sixty million, but isn’t that the price? It has to cost less than that for them to make a profit. And I don’t think they’d want to reveal the actual cost, for business reasons, but it would be nice to know just how much margin they have, and how much they can reduce the price if it’s fully reusable.

[Update a few more minutes later]

And here we have an article from Mike Wall, where he quotes Elon as saying it cost $16M to build (if true, that gives them a huge profit margin and room to drop prices in the face of any competition). I saw others reporting that on Twitter on Monday, but no one really clarified if he said “sixty” or “sixteen.” It would be nice to get the actual number.

[Update a while later]

Here’s Miriam Kramer’s take.

[Update a couple minutes later]

And Jessica Orwig’s.

[Update a couple minutes later]

Rounding the Cape of Bojador.

[Afternoon update]

Here‘s SpaceflightNow’s take on it.

33 thoughts on “Reusing Rockets”

  1. I’m curious about the next couple launches. In January Jason 3 is scheduled to launch from Vandenberg. Will they try to land on land, on a drone ship, or not at all? That rocket is the last of the older, non-full thrust design; if they do recover it, what would it be good for?

    Also in January they’ll be launching SES-9, a 5,300kg comsat, to GTO. Do they have enough performance to get the first stage back to land, or will they try another drone ship landing?

    1. Jim, there are some complicating issues with the Jason-3 launch. It’s unknown whether whatever they did to address the valve stiction issue (that caused the landing failure on CRS-6) has been done to the Jason-3 LV. If not, they may have trouble getting permission. (though they were asking for it for the originally-scheduled launch some months back)

      My wild guess is they’ll try a drone ship landing, because they still need to work on that (and they do have a drone ship on the West Coast).

      What would a recovered older version of the core be good for? Interesting question. Perhaps as an expendable LV? Or for engineering analysis? I really have no idea.

      Regarding SES-9, I’m really curious what’s going on. To deliver that much mass to GTO was beyond the stated ability of the pre-full-thrust F9, and that’s without reserving fuel for recovery. If they can put that much mass into GTO and save enough prop for a Droneship landing (which is less prop than RTLS) either they’ve increased performance more than claimed, or the original max performance was higher than claimed.

  2. In just under a decade:

    1. “No, it’s impossible, it can’t and probably should not be done. And even if it could be, why would you add all that cost, complexity and mass to only a booster? No, just improve the performance and save the rest for payload…”

    2. “You have to convince your customers to fly on a used rocket versus the competitor’s shiny new one,”

    and soon:

    3. “There is no reason not to fly mission critical hardware on a case proven LV. In fact there is *every* reason to avoid flights on untested hardware if at all possible.”

    1. 1. Returning a stage to the launch site for reuse does cost performance. But if it cuts cost enough it’s worth it.

      2/3. I note that what failed on the previous launch was going to fail on the first launch or not at all. There are, of course, other failure modes related to structural fatigue that need to be characterized, which is only possible once you recover a stage intact.

      1. When FH flies, the performance hit for returning the side cores will be even lower, as those will stage at lower velocity than the F9 core (I believe).

          1. I read recently (probably at NSF.com) that they’ve put crossfeed on the back burner. That makes sense as it probably adds complexity. Better to concentrate on just flying Falcon Heavy for now, then add crossfeed later as an upgrade.

          2. Thinking about it, skipping crossfeed means both the boosters and stage one can have very roughly the same return velocities, etc.

            Discard the two boosters, wait ten seconds (or whatever) and then first stage separation.

            That makes -three- that should have a chance at a return with no major changes. Three hopefully reusable parts (without figuring out separate higher-velocity recovery) even better than two.

          3. There can still be a considerable difference between central and external core velocities if they throttle down the center core shortly after liftoff. IIRC, this is done on the Delta IV Heavy. With all of the improvements they made for the Full Thrust version of the Falcon 9, crossfeed may not be immediately necessary to get good performance. That’s something they can work on for a later upgrade to Falcon Heavy.

  3. You have to look at the overhead. Since SpaceX builds rockets internally, labor is more significant than input hardware costs. So it’s more informative to make estimates on their employee base.

    4000 employees – let’s estimate that each employee costs on average $100k fully loaded. That means around $400m per year in labor costs. Add to that facility and operations expenses, production input expenses and various miscellaneous expenses and you can easily get to $600m-$700m per year in overall costs.

    Now some of that are externally funded projects – notably the Commercial Crew project where R&D is funded by NASA. Perhaps that’s as much as $200m – $300m of that annual total (for now – will change as Commercial Crew goes from R&D to operations). That still means you have to generate revenues on $400m of costs. Dragon operations help, since they make additional revenues on that, but otherwise it seems like they need to launch 6 or 7 times a year to make money. This goes for whether the rockets are reusable or not, since you are still carrying a lot of the same overhead costs.

    These are all estimates and they can be wildly wrong in both directions but if anyone thinks I am off by a lot or haven’t considered certain things, let me know.

    1. “let’s estimate that each employee costs on average $100k fully loaded.”

      Ian,
      $100k is way low. A average fully burdened engineer at an “old aerospace” company is around $400k. Granted that includes some unnecessary overhead. I think a better guess at average cost per employee is around $350k maybe a bit lower since not every single employee at SpaceX is an engineer.

      Anyway I know for sure $100k is way low because I have a friend (mid-level engineer) at SpaceX and his salary alone is well over $100k.

      1. When I was working at SAIC in 1997, the rule of thumb was that the average fully-burdened degreed professional would cost ~ $200k/yr. Using a CPI inflation calculator, which probably comes in low, that’s $295k/yr now. Don’t know how that translates into skilled shop workers, etc.

      2. That point is well taken. I started with $100k because:
        1) SpaceX has been known to pay below-market salaries
        2) The 4000 employee number includes a lot of facilities management, security, production and variety of non-engineer ‘blue-collar’ jobs. Which is not even to say that those jobs are low paying – in many cases far from it. .
        3) Fully burdened in my mind means adding payroll taxes, health insurance and various other benefits and possibly other insurance coverages depending on the nature of the work. Stock or option grants complicate this calculation and I’m not sure if I would include since it doesn’t involve an immediate cash cost. But anyways, if I’m egregiously missing something in terms of overhead, let me know.

        But in any case, it shows the economics are worse than I even calculate, which poses real questions about how cheap their launches really will become in the next few years.

        1. Probably not too much wiggle room on price without a drastic increase in flight rate but depends on how much the cost of the first stage makes of the total cost and how much it will cost to refurbish. I’ve been playing around with a spread sheet but I am 100% sure my assumptions are wrong.

          So peanut gallery, how much do you think it would cost to refurbish the stage and if the rocket costs $16m, how much of that is the first stage?

          2. –Rand Simberg
            December 24, 2015 at 7:27 AM

            I think that Elon has said the first stage is about 70% of the cost.–
            Or about 12 million for first stage with each of 9 engines being a million or less.And inter-stage, second stage and fairing being about 4 million,
            Though putting them all together and fueled and on the pad is another couple million- and I am not including cost of having and using a launch pad and launch range, which could be about 5 million per launch. Nor counting the integration with whatever the rocket payload is.
            Or compared to just making the rocket the other stuff is expensive. Also in terms of other rockets, the first stage though the most massive is generally cheaper than other stages. The Falcon-9 first stage is expensive because it has 9 engines, though making a lot of engines allow a low cost per engine and about 9 million for any kind of first liquid fuel rocket engine is very cheap. And with similar engine in second stage- the second stage engine is also very cheap.

  4. –And here we have an article from Mike Wall, where he quotes Elon as saying it cost $16M to build (if true, that gives them a huge profit margin and room to drop prices in the face of any competition). I saw others reporting that on Twitter on Monday, but no one really clarified if he said “sixty” or “sixteen.” It would be nice to get the actual number.–

    16 million for first and second stage seems about right to me.
    But cost is largely related to how many are made per year [day/month/whatever]. One can’t sell it for 16 million as the business has other costs, though if one made say 20 per year, one might be able to sell it for around 16 million, and if say made 30 per year maybe sell for less than 16 million and make a profit.

    Another way to look at it, is the landed first stage is worth about
    5 million or were SpaceX to have a surplus of them, it **might be** worth selling them at 5 to 10 million [assuming anyone wanted them] but it seems to me that first landed first stage is worth to SpaceX more than 10 million. Because they learn how to improve the Falcon-9 from it. And using it again and recovering it again, would increase it’s value to SpaceX [though less to someone else who wanted to use a used rocket]. Or in simple terms a landed first stage has scrap value of about
    5 million. Though again, scrap value assumes SpaceX wants/needs to sell it- which would be crazy at this point in time.

    Then there is the cost refurnishing the first stage. So 5 million as scrap, is one lands it, and someone buys it as is, not the price after refurnishing it and it’s sitting again at pad ready to fly again.
    And finding out how much is costs to getting the stage so it can fly again is quite important. I would imagine they would want to do several of these used stages at same time to get idea of these costs. But this first stage which is now landed will be examined as a preliminary step so one can gear up to refurnish many used stages at one time.

    1. But it seems to me the whole point of all of this, is so one can fly the Falcon Heavy. Or if this was merely limited to Falcon-9, it’s uncertain to me that it’s worth it.

  5. Question for the peanut gallery: which technological development was Monday’s landing analogous to? Doesn’t have to be an aerospace analogy. I always enjoy Edward Wright’s history, I hope he chimes in. My vote is X-1’s first supersonic flight.

    2. As Rand mentioned before; the Wright’s successful second landing of the 1903 Flyer. The first goes to Blue Origin.

      1. Um, the flight of New Shepard is analogous to Shepard’s flight whereas Monday’s is like Glenn’s?

        I was seriously impressed by the video of New Shepard’s landing. The new ground that they broke since the Delta Clipper is that they thing came wiffling in at quite the speed before they did a rocket burn just short of the ground so as to achieve a soft landing.

        The new ground covered by SpaceX is a first for a liquid-fuel fly-back-to-launch-site booster that was part of a launch that not only achieved orbit but delivered a useful payload. Shuttle never achieved that.

        To get that, however, I guess that they resorted to some “tricks” such as the super-chilled propellants to cram more into the booster. But having successfully flown, and having recovered the stage, of course, they can do the structural analysis and re-crunch their numbers and so on.

        I guess the next milestone is recovery of both the booster and orbital stage? If you achieve that, you have achieved the whole “halfway to anywhere” thing because even if you can only orbit small payloads, you could do things like tank aerobraking transfer stages for the trip between low Earth orbit and “anywhere.”

        1. “Um, the flight of New Shepard is analogous to Shepard’s flight whereas Monday’s is like Glenn’s?”

          I like that one

        2. Was this also the first launch of a vehicle with subcooled LOX? They cooled the LOX down to below 67K, increasing its density (from the 1 bar BP) by about 9.5%. The RP-1 was also chilled, but the density increase there was not as great.

    3. This is a serious game changer. I’d compare it to the introduction of the home computer. Prices to orbit are going to drop dramatically, so much so that even small businesses and high schools will be able to put stuff in orbit.

  6. $16m makes more sense. It would be nice to know how much of that is the first stage. Will be interesting to see how long it takes before they reuse one of these stages and then how long before they are a regular part of operations. But the most exciting, will be seeing just how much demand there is at lower price points.

    Also, when NASA will want a better deal?

    I hope Bigelow is ready to roll.

  7. But the most exciting, will be seeing just how much demand there is at lower price points.

    Yes, this goes back to the sci.space.* discussions twenty years ago about CATS (cheap access to space) and price elasticity. There was one thread that I remember that asked about not CATS but FATS (free access to space). That is, if by means of magic, you could put 20 tonnes per month into a LEO orbit (say 500 km altitude at any inclination) for zero cost (*) to yourself and you could charge whatever price you could wring out of the customer, who would your customers be? Consider the revenue flow over both the short term as the business started up and then the long term.

    (*) If zero cost trips your breakers, say it’s $1.00/kg.

    1. I certainly don’t know but it is fun to use the old imagination. The cost of access to space is definitely a major hurdle but one of the same things that made SpaceX’s feat possible, the confluence of maturing technologies, is also having a large impact in other industries as well. Other barriers are also being reduced.

      Can’t wait to see what products and services people come up with and what new markets are discovered.

  8. Yeah – I used to work for Bellcore back in the late ’80s. We were about to roll out ISDN/dsl and everyone was having a big argument: is there really a market for cheap data going over your phone line? Who is going to use it exactly? We didn’t know, but we hoped that uses would spring up.
    Then came the Internet and it all just exploded.
    We’ll see.

    1. We all like to use our imaginations but no one can predict the future. That new space elevator documentary had some good quotes about the future and technology. I’m not sure a space elevator will ever be built but the movie was pretty inspiring about the future in general.

Comments are closed.