Wednesday, 18 October 2017

SpaceX SFR: Small Falcon Rocket

The Small Falcon Rocket is a scaled down alternative to SpaceX's Big Falcon Spaceship that fits on top of existing Falcon 9 boosters. 
We will discuss the advantages and disadvantages of such a design.

SpaceX's Big Rockets

The BFR, or Big Falcon Rocket, is comprised of the Big Falcon Spaceship and the Big Falcon Rocket booster. It is a scaled down and simplified design based on the ITS, or Interplanetary Transport System.
The BFR is a BIG rocket.
The ITS was revealed in June 2016, although work on the design has begun in 2013 under the name 'Mars Colonial Transporter'. The ITS promised to deliver 300 tons of cargo to Low Earth Orbit, or up to 550 tons if reusability was ignored. It would have massed 10500 tons on the launchpad. The vehicle had a diameter of 12 meters and a height of 122 meters, making it one of the largest rockets ever plausibly considered. 
And the ITS was positively massive.
The upper stage, called the Interplanetary Spaceship, was supposed to hold 1950 tons of propellant with a dry mass of 150 tons. Without a payload, the mass ratio was 14. 

The BFR replaced the ITS in September 2017. It is a smaller, more sensible design that SpaceX believes it can actually deliver in the next few years. The diameter is reduced to 9 meters and it will mass 4400 tons on the launchpad. Payload capacity is reduced to 150 tons.
The upper stage BFS should have a dry mass of 75 tons, but Elon Musk states that this might rise to 85 tons due to development bloat and overruns. It holds 1100 tons of propellant, giving it a mass ratio of 13.9.

It is important to note that despite being up to 78% smaller than the previous ITS design, the BFS stage maintains the same mass ratio. Why? Because we are now going to scale down the BFS again.

Why go smaller?

Going big is the best way to reduce the cost per kilogram for sending payloads into orbit. SpaceX jumped from the Falcon 1 to the Falcon 9 because the larger rocket can deliver payloads much more cheaply into space. When first considering options on how to make travel to Mars affordable to the general population, SpaceX immediately came up with a gargantuan tower of rocket fuel over three and a half times larger than the Saturn V! 

A big rocket is also easier to develop. It is more forgiving of development bloat that increases mass over time as the designs are perfected. It has larger safety margins and room for many backups, such as multiple engines. 

However, bigger is not always better.
The total development costs will be higher, as large components need large factories. It is much more difficult to test the components too, and a full testing regime of the completed rocket will require launching and even destroying a full-scale model many times. Remember the failed Falcon 9 booster landing attempts, and imagine them replaced with a vehicle eight times bigger.

There is also the fact that the second sure-fire way to reducing launch costs is to have rapid turnover. This involves loading up rockets, sending payloads into space, recovering the rocket and refurbishing it for another launch in a very small time frame, measured in days or even hours. Rapid turnover and minimal refurbishment would allow the space launch industry to more closely resemble existing airline business models. The main benefit of this approach is that a small number of launch vehicles can handle a large volume of missions, critically reducing the initial cost of the vehicles and reducing the amortization rate.

Even if SpaceX manages to develop rockets that liftoff and land several times without needing to go to a workshop, they'd still need to solve the issue that there just aren't enough payloads on the market that need to be lifted into space to fill the BFR, let alone the ITS.

For example, even the BFR's 150 ton payload capacity can cover all of last year's payloads in about two or three launches. Three launches is far from sufficient. Elon Musk is betting that the space industry will be able to fill the BFR's cargo bays with new satellites and LEO payloads once the lowered cost per kg is offered to them... but there will be a long delay between the launch costs being reduced and the industry contracts appearing en masse.

Cost per kg in orbit is only part of the picture.
Waiting for more contracts to appear and bundling them together to use the most of a BFR's cargo capacity is not a good solution. It will force SpaceX to delay launches until the mass delivered to orbit reaches a profitable amount - launching BFRs nearly empty with the usual 2 to 5 ton satellite is surely wasteful and a loss for the company. 


The SFR, or Small Falcon Rocket, is a possible solution to the development costs, under-utilization and low expected launch rate of the BFR, or Big Falcon Rocket.

The SFR is a scaled down Big Falcon Spaceship sitting on top of an existing Falcon 9 booster. It will carry a smaller payload to orbit, but will have a capacity SpaceX is sure to fill up. Existing Falcon 9 boosters can be mated to a fully reusable upper stage, drastically cutting down on development costs. 

We will now look at the details of the SFR's two stages.

The upper stage is the only new part. It is a BFS scaled down to 3.7 meters diameter, using the same Raptor engines rated at 1900kN of thrust at 375 seconds of Isp. We will call it the SFS, or Small Falcon Spaceship.

The Raptor engine.
The SFS will be (9/3.7)^2: 5.9 times smaller than the BFS. The dry mass is expected to be only 85/5.9: 14.4 tons. It will be 19.7 meters long. 

Based on the mass ratios calculated above, the SFS will be able to hold 187.2 tons of propellant. An SFS with no cargo and full propellant tanks will therefore mass 201.6 tons and have a deltaV of ln(14)*375*9.81: 9708m/s. The Vacuum-optimized Raptor engine is quite large, with a nozzle opening 2.4 meters wide.  It is unlikely that more than one such engine can be fitted under the SFS. It will provide enough thrust for an initial Thrust-to-Weight ratio of 0.96, which must be compared to the current second-stage initial TWRs of 0.8-0.9. For retro-propulsive landing, we will not be able to fit, or even need, the sea-level version of the Raptors. Instead, we will use two of the existing Merlin-1D engines with 420kN of sea-level thrust, but possibly with a lower pressure rating as the thrust generated makes them too powerful for landing. The alternative is the SuperDraco engines with 67kN of thrust and 235s sea-level Isp.

Rocket engines in the Raptor + 2x Merlin configuration would represent 13.2% of the overall dry mass, or 8.1% if the Raptor + 4x SuperDraco configuration is used instead. The Raptor engines are assumed to have a TWR of over 200, so their mass should be lower than 969kg. There are no numbers on the SuperDraco's mass, but it should be at most 50kg. These ratios seem not too outrageous when compared to the 7% engine-mass-to-dry-mass ratio in the BFR's original design.
Merlin-1D engines.
The SFS's mass is based on the 85 ton figure for the BFR's dry mass, but this is a cautious estimate with room given for development bloat and mass budget overruns. The BFR's design on paper gives a dry mass of 75 tons instead. Using the on-paper mass, the SFS could have a dry mass as little as 12.7 tons. 

The SFR's booster is the Falcon 9 Block 4. The booster will mass 22.2 tons when empty, and can hold 410.9 tons of propellant. This gives it a mass ratio of 19.5. The nine Merlin 1D engines have a sea-level Isp of 282s and an vacuum Isp of 311s. Because the booster stage does not spend a long time at sea level and performs most of the burn at high altitudes with negligible air pressure, we will use 300s as a low-ball estimate of the average Isp. The true average might be a few seconds higher. 

Taken all together, the SFR will mass 634.7 tons on the launchpad without any payload in the SFS's cargo bays. It stands 89.7 meters tall. 

We will now calculate how much cargo it can lift into Low Earth Orbit in expendable or reusable mode, and where else it can go.


To achieve a Low Earth Orbit, we will set the deltaV requirement as 9400m/s. In reality, it could be achieved with as little as 9200m/s, but we want decent safety margins.

Expendable mode is the easy part. It assumes every bit of propellant is consumed and the SFR's stages left dry. Using a multi-stage deltaV calculator and setting the Falcon 9 Block 4's Isp to 300s and the SFS's Isp to 375s, we work out that the booster provides 2877m/s of deltaV and the SFS provides 6513m/s for a total of 9391m/s with a payload of 24 tons. 

Recoverable mode is harder to calculate. The propellants cannot be completely used up: some must be kept in reserve to perform a retro-propulsive landing burn.

BFR landing.
A landing burn by the SFS requires that about 400m/s of deltaV be held in reserve. This represents 2.24 tons of propellant with Merlin-1Ds or 2.72 tons of propellant with the SuperDracos. 

The Falcon 9 booster needs to retain 15% of its propellant reserve to make an ocean landing. This gives it a deltaV of 3910m/s, which is largely enough to cancel most of its forwards velocity and make a very soft landing. However, holding back 61.63 tons of propellant means it boosts the SFS by much less. 

In recoverable mode, the SFR's cargo capacity drops to 12 tons. Using SuperDracos lowers this again to 11.5 tons.

If the SFS follows the paper designs more closely and achieves a dry mass of 12.7 tons, it will have cargo capacities of 25.5 tons in expendable mode and 14 tons in recoverable mode.

The SFS could achieve a deltaV of 2350m/s after launching on top of a recoverable Falcon 9 booster and without any payload. This is not enough to reach the Moon, so the range of missions the SFR can take payloads on is limited to Low Earth Orbit.

Smaller rockets might solve the problem of having to crane down cargo from the top of a tower. 
However, if it is refuelled in orbit, then the entire Solar System is available. It can deliver 50 tons to Low Lunar Orbit (5km/s mission deltaV). It can send 35 tons to the Mars Low Orbit (5.7km/s mission deltaV) or 21 tons to Mars's surface (6.7km/s mission deltaV). Refueling the SFS will take between 15 and 17 tanker launches. 

With 14.4 tons of dry mass and a propellant capacity of 187.2 tons, the SFS has a maximal deltaV of 9.7km/s, enough theoretically to put itself far above Jupiter or even Saturn.


The SFS is a limited vehicle. It is restricted to Low Earth Orbits and can deliver payloads of 12 tons, up to 14 tons, at most. It is far from the multi-purpose machines the BFR or ITS promised to be. 

However, it is enough to dominate the medium lift launch market, as it is fully recoverable. The re-use of existing Falcon 9 boosters and the smaller number of Raptor engines (one per rocket) will drastically slash the development costs compared to something like the BFR. The smaller payloads are easy to fill, meaning every launch is profitable. Multiple launches promises rapid turnover and a maximization of the return on investment on the craft. 

With re-fueling, the SFS in orbit can complete missions that require it to send decent payloads to the Moon and Mars. With minor improvements and operating in fleets of multiple vehicles, it can even match the payload capacity of the BFR to various destinations. 


  1. I'm dubious. SpaceX claims a BFR launch will cost much less (not less-per-Kg, but less!) than a Falcon launch - something like $10M dollars. It may be more attractive to just launch mostly-empty BFRs than try to pay off the R&D on a SFR at less than $10M/launch savings, especially after SpaceX has apparently had to put more effort into Falcon Heavy than expected.

    Another possibility; the BFR 2nd/spaceship stage (BFS) is itself within rounding error of an SSTO with a tiny payload. We might see an embiggened or otherwise enhanced BFS developed prior to its booster as a solution for small payloads; this would fit into the SpaceX roadmap for a larger BFR later.

    1. That might be a deal-killer, I admit. If SpaceX can send up mostly empty BFRs en masse and still make a profit, then there is no need for a smaller vehicle... but that begs the question - why spend millions upon million on developing, building and servicing a vehicle you know for sure that you will not use to full capacity?

      The BFR as SSTO has many problems. First and foremost, it needs to lift off on 330s Isp engines, and nine of them. If the Raptor engines mass about a ton, adding seven of them to the existing engines will remove any SSTO capability.

      There is a window though. If SpaceX manages to stick to the on-paper 75 ton figure, then it has about about 3-4 tons of cargo capacity left over and still achieve orbit as an SSTO... but you'd need about 9 tons of propellant in reserve to land again. I seriously doubt Elon Musk will go for one-way SSTOs with only 3-4 tons of payload capacity.

    2. I don't expect the BFR to be used as a SSTO, or at least not until SpaceX develops aerospike engines. Arca is hoping to make a small SSTO powered by a linear aerospike with 12 combustion chambers burning RP-1/hydrogen peroxide (these are picked for their high density, reducing tank mass).

    3. The ultra-high-pressure Raptor engines that SpaceX was developing (until it downrated them to 170 Bars instead of 300 Bars) are impeded very little by atmospheric pressure.

      A sea level Isp of 330s is impressive, but most importantly, it is over 92% of the vacuum Isp of 356s. Other high pressure engines achieve 80% of their vacuum Isp at sea level or less. Even the older aerospike designs, such as the XRS-2200, had a sea level isp of 77% of the vacuum Isp.

      All-in-all, I think high pressure engines are the best development path right now instead of attempting to see the benefits of an aerospike.

      Arca is not a great example. The propellant choice is fascinating as it allows for extreme mass ratios using simple construction techniques. However, the aerospike is absolutely necessary - it just barely reaches orbit without any payload if it used conventional rocket technology, as the sea level Isp of a
      HTP/kerosene engine is expected to be only 268s.

    4. I expect SpaceX to jump directly from combustion engines to continuous detonation engines. Those are basically an advanced version of the aerospike, and by the time SpaceX has a need for those, the technology should be industrialized already.

      Given that so much of the cost of a rocket is on R&D, there may be a business case for launching near-empty and often if demand fails to materialise.

      It seems that Arianespace isn't expecting such increase in demand. They made an interesting argument against reuse in that case: they could only build a handful of engines and then shut down the engine plant, so not only each engine would be significantly more expensive, but they could not retain the skills and potentially the facilities beyond that.
      SpaceX is in a more favorable position as it has access to the massive US military market which, against something like ULA, is more or less a guaranteed captive market.

      Arianespace does have reusability plans on the backburner, though, so if the market explosion occurs, they could probably make the switch in a few years.
      I'm very curious to see how both strategies will do.

    5. That's a massive leap! A new, untested engine technology is not something SpaceX wants to do. It is successful because it is innovating in how to use old tools, not because it is spending millions over years to devise new solutions.

      A continuous detonation engine would allow for a big leap forward in propulsion technology...but the same effect can be achieved by just loading up very big rockets with a whole lot of propellant and re-using them.

      Arianespace is not led by a visionary risk-taker and better like Elon Musk. It has to answer to shareholders, often the same governments and industries that feed it its contracts. Every decision has to be weighed somberly against returns on investment, market projections, development life cycles and even how many jobs are being created. Elon Musk's SpaceX is hindered by none of these things, so it push for a supply-first scheme where SpaceX builds the capacity and then sits on its assets while it waits for the market to catch up.

    6. @Matter Beam: You're quite right that just building a reusable 2nd stage for Falcon (or Falcon Heavy?) would be the sane move. But Elon really, really wants his Mars colonization ships, and the new BFR is a compromise that can do that and still take over the earth-orbit payload market. Unless the market hugely expands, once you cut production prices 80%+ with reusability your costs are dominated by R&D - you really want only one rocket with only one engine (type) if you can make the numbers work at all.

    7. In that case, cutting R&D costs and maximizing the number of flights would be a priority, right?

      You increase the number of flights so that the apportioned price you add to each launch to recoup the initial investment gets divided. If you develop a $10 billion design and one make one vehicle, then you'd need to make $10 billion with that vehicle. If you make a hundred vehicles, each performing dozens of missions per year, then your $10 billion would be adding less than $1 million to the cost of each launch!

    8. Matter Beam: "why spend millions upon million on developing, building and servicing a vehicle you know for sure that you will not use to full capacity?"

      Stating the obvious but, to go to Mars.

    9. So the BFR lives or dies on whether or not the US government unlocks billions of dollars for a manned Mars mission...

    10. SFR only makes sense (maybe) at an intermediate flight rate. If it's flying a thousand missions a year, it would have to compete with BFR launching every week or two, and that turns into a cost-per-Kg contest that BFR probably wins.

    11. (assuming SpaceX can and does build BFR, which is kind of insane but I'd hate to bet against it at this point.)

    12. One of the SFR's big advantages is that its cargo capacity will match any expected market demand for the next decade or more, so there is no waste, and even if its development cost is not a lot lower than that of a full BFR, it can divide that cost for recovery over hundreds to thousands of flight instead of over a handful of flights as with the BFR.

      The BFR presentation suggested that the development cost would be divided over the BFR's 1000 flight lifetime... but that is not a reasonable apportionment if it takes several decades for those 1000 flights to be done. It is much more likely that SpaceX will have to recover the development cost of a number of years, maybe ten, before the BFR design becomes obsolete.

      SpaceX is gambling on the Mars mission making the 1000 flights actually achievable in a short time. Charging forwards and building the BFR before it is needed would actually make the Mars mission more likely to happen, which would give the BFR a bigger chance to succeed in a rather circular manner.

    13. Wouter Debois writes:
      """ I don't expect the BFR to be used as a SSTO, or at least not until SpaceX develops aerospike engines """

      Not necessarily. Using ablative nozzle inserts to reconfigure the vacuum nozzles for the boost phase, and then discarding them at altitude, provides a means for a BFR to go SSTO with more than just a tiny payload.

    14. Full-pressure Raptor engines have such a small difference between sea-level and vacuum Isp that no sort of nozzle extension or aerospike is needed.

      The problem really is the amount of propellant that needs to be held back in reserve for landing. An SSTO-variant of the BFR would probably involve the use of tankers that stay in orbit to refill the landing retro-propulsion tanks. This allows the SSTOs to launch with empty landing or header tanks, which increases their payload capacity significantly.

  2. Absolutely wonderful read. Thank you.

    Follow-up question: What would it take to make the BFS serve the same function (i.e. a fully reusable SSTO medium-lift launcher).

    I'm asking because in his Reddit MMA, Elon mentioned the BFS will be testing reentry under its own force (before the BFB is developed), and that it's capable of reaching "orbital speeds" (whatever that means).

    1. You are quite welcome!
      It is likely that the BFS during testing will use a full set of sea-level optimized Raptors so that it can lift off on its own. This will give it a deltaV of 9201m/s, which is enough to reach space and fall down through the atmosphere at re-entry like velocities (~8km/s).

      To actually be useful as an SSTO... it will need a drastic reduction in the dry mass and a set of engines with even higher thrust to weight ratio!

    2. How would the numbers change if SpaceX went to "Raptor Full Thrust", a.k.a. the IAC 2016 proposed numbers (3050KN instead of 1700KN at sea-level, etc)

    3. Then the stick-to-the-plan 75 ton version of the BFR would be able to go SSTO and land again with a payload of about 2-3 tons.

    4. well that's disappointing.

      Side question: I want to be able to do these calculations myself. Where do I find the equations cheat-sheet?

    5. I simply used the deltaV equation and a multiple stage deltaV calculator, and incrementally increased the payload mass until I reached the deltaV limits...

    6. Shai: One of the better tutorial/references I've found is actually on the wiki-guide to Kerbal Space Program (!), here:

      A couple of things (specific deltaV requirements and the optimal thrust/weight ratios for takeoff) are specific to the Kerbals' very peculiar solar system, but the basic equations are all there.

  3. You suggest that the craft could get 21 tonnes to the Mars surface but a whole 35 to Mars orbit. I'm not sure that this is correct. While the trip from LEO to LMO might have less delta-v needs than the trip from LEO to Mars surface the latter can use aerobraking quite a lot when landing on Mars, while the former will find it less easy, though not impossible, to use aerocapture to enter LMO then aerobraking to reduce the apoapsis. Given ths I would susect that the craft could get either the same mass to LMO or the surface, or if it can't do the aerocapture but can do aerobraking for a direct entry (this latter appears an easier task as probes do it routinely while aerocapture has only ever been done by returning soviet zond lunar craft) then get more mass to the surface than to LMO. Unless the 21 versus 35 tonnes is accounting for extra mass needed for specialised aerobraking equipment and heat shields which can survive deeper into Mars' thin atmosphere.

    P.S. The F in BFR does not stand for "Falcon", Elon Musk doesn't find giving his craft some rather direct and blunt names.

    1. I added 1km/s to the deltaV budget for landing. Elon Musks stated that 99% of the kinetic energy upon intercepting Mars will be dealt with through aerobraking. That's 90% of the relative velocity.

      A Mars intercept will probably happen at velocities of between 5 and 7km/s, so the landing budget must include between 500 and 700m/s; 1km/s is overkill.

      The aerobraking thermal requirements for a Mars-bound BFR are about 8/6^2 * 1/0.006: 296 times lower than for re-entry on Earth.

      I know about the F in BFR but the SFR is truly a 'Falcon' rocket as it sits on top of an F9 stage :)

    2. SFR = SFS + Falcon 9 booster.

  4. I think the biggest drawback are still the engines themselves. Chemical thrusters are very weak. Elon should think about using NERVA nuclear powered rockets. The problem is going to be the political dimension- plenty of public outcry.

    1. Politically, that's a non-starter. Elon shouldn't spend energy and good-will tilting at windmills.

    2. @John Triptych: But chemical rockets are what we have for the foreseeable future. SpaceX is big and successful, but has nowhere near the resources to take on the development of a new propulsion technology to modern standards on its own...

      @Shai Machnes: Definitely. Elon Musk would have to spend SpaceX's entire budget on lobbying and ad campaigns just to be given permission to start testing nuclear rockets, let alone flying them commercially.

    3. @John Triptych: On the contrary, chemical thrusters are very strong; what they are is inefficient. The massive weight of a NERVA-style solid-core nuclear engine means it's at best a marginal improvement for takeoff from the earth's surface; it might be a win for a second stage. For interplanetary maneuvers I suspect various types of ion engines are superior, but I suppose the higher thrust of nuclear-thermal might be superior because Oberth effect etc. depending on details.

    4. I think it will be quite some time before the government permits a private carrier to operate nuclear reactors.

    5. Nuclear rockets are going to be off the table, regardless of their technical merits. The "Timberwind" nuclear rocket designed under the Strategic Defense Initiative had calculated Thrust/weight rations of 30 and ISP's as high as 1000. These had to be designed in secret, and once the Cold war ended, they were quietly discarded, despite being by far the most advanced nuclear rocket engines ever designed.

      Politics will keep nuclear rockets grounded for the foreseeable future, and even a nuclear engine for the BFR or SFR second stage is not going to be approved, regardless of the giant increase in performance a Timberwind engined spaceship would provide.

  5. Have you considered a version of this proposal for Blue Origin's New Glenn re-usable 1st stage?

    My impression is that Elon and SpaceX have been clear about their hesitancy to do a re-usable 2nd stage on the Falcon Heavy. They think it would be a distraction and Elon said they've already started to order tooling for the BFR.

    However, I think BRF/BFS might start to put pressure on Blue Origin. B/c if SpaceX can get a per-launch marginal cost down below $10m (i.e. below Falcon 1) then New Glenn with a non-reusable 2nd stage will face a lot of cost pressure.

    (I get that Elon's wonderfully ambitious. So it make take a while and a few BF-explosions before his exceptionally aggressive BFR marginal cost number is realized.)

    1. I don't have any data on which options Blue Origin has considered for making the second stage re-usable, if they have at all. It would entirely be speculation at that point.

      My personal proposal is something similar to the SFS, but without any of the fuss. A vacuum rated engine at the bottom, a single fuel tank and a large cargo bay with hinged fairings. The nozzle of the engine is retractable, allowing the second stage to fit snugly inside a thermal protection system on the lower half. The actual recovery and landing mechanism is parachutes and a short retro-burn much like the SuperDracos for the Dragon pod.

      The SFR proposal can guarantee that SpaceX remains competitive in the short to medium term, and gives it a way to monetize its investments in BFR-like designs if the Mars mission doesn't get funded and the BFR never gets built.

  6. I believe there is an error in your scaling down of bfs. You scaled down with 5.9, the ratio of the cross sectional areas, not accounting for the greatly reduced length. So your numbers might be good for a 48m x 3.7m sfs, but such a thin design could have some issues related to controllability or structural efficiency, especially when stacked on a 42m tall Falcon 9 first stage.

    1. The length difference between the 42m and 19.7m SFS would be mostly propellant tanks. They mass very little per meter height, when compared to the sectional density (kg/m^2) of the spaceship.

      If anything, the dry mass of the SFS should be lower, but only by a few percent.

  7. You proposed mixing one Raptor with two Merlin engines, but these engines use different propellents. Merlin burns RP1 (aka refined kerosene) while Raptor uses liquified methane. That would mean having separate kerosene tanks, piping, etc, just for the landing engines, which would incur a substantial weight penalty.

    Also, your scaling was based on the square of the diameters, but that implies the length would be unchanged. You should instead base your calculations on the ratio of the volumes. Interesting thought experiment, though!

    1. The mixed propellant proposal is only a stop-gap measure until SpaceX develops small-scale or low-pressure Raptors that can throttle down to 31kN of thrust. If the Raptors are available immediately or at cheaper cost than converting the SFS to multiple propellants and back, then they will be used instead of other options.

      The weight penalty is already included in the header tanks that the SFR would have to have anyway to keep a load of reserve propellant separate form the main tanks and ready for landing.

      The area scaling does make the SFS very long and thin, but I assumed that the majority of the length would be taken up by propellant tanks, and their dry mass by length is very very small (just a section of the tanks) compared to the dry mass by area (all the systems and engines). So if anything, yes, the SFS's dry mass should be even lower... but only by a handful of percent.

      Thanks! Following the intense discussion on reddit on this blog post, it seems reasonable to try doing a more detailed set of calculations for a 5m diameter SFS.

  8. 12 tons to LEO - that's not what SpaceX want and ridiculously small payload to GTO.
    Mix two types of fuel on one stage is an irrational idea.

    Every BFR launch is going to be profitable because customers are paying for payload delivery, not for BFR launch itself. It's impossible to launch all year satellites on one BFR due to different orbits, inclinations and readiness.

    1. 12 tons to LEO is enough to service the entire small and medium launch market. For satellites that want to reach GEO, they can be launched empty. The SFR's rapid recovery and re-launch rate means a second launch can refill the satellite with 12 tons of propellant at a time.

      The only thing the SFR can't do is place massive satellites in low orbit (>12 tons dry mass) or serve the Mars mission.

      However, the number of massive satellites is very, very tiny... and a regular expendable second stage on the Falcon 9 booster can handle them up to 22.8 tons.

      If the BFR has to launch several times to handle the different inclinations and orbits for the same tiny expected market, then it will be even less profitable!

  9. I've been thinking about this post for a while, but really could not word the issue that was bothering me. I now see this as a case of making an unwarranted assumption about scaling. A smaller spacecraft capable of being launched from a Falcon 9 or Falcon 9 heavy isn't going to resemble the BFR's spacecraft component in much the same way a Ford F-150 pickup truck does not resemble a Mack Truck with a cargo box.

    Yes, there are similarities, but components do not scale at the same rate, and the fractions needed for things like the engine, transmission and frame will be quite different between the larger truck and the smaller pickup. You can do the same sort of comparison for ships, airplanes or other devices and come to similar conclusions. I would have to say in general, a much large fraction of the small ship will be devoted to structure and systems, so the performance will be far lower, or the payload far smaller, or both.

    So the SFR will scale differently, and its performance will not be comparable to the BFR. This makes most of the assumptions rather sketchy (probably close enough for a handwave, but not good enough for any accurate analysis). Since I don't actually know how the SFR wold scale, I can't get into much more detail. Please don't think of this as an attempt to trash the idea, but rather to look for ways to refine it.

    1. No, you make a good point. Scaling won't be so straightforward as I describe above. But, what I have read so far on the topic of a smaller BFR actually suggests that the dry mass of my supposed SFS stage would actually be lower. This is based on the claims of Robert Clark on Polymath and Gary Johnson from 'An Ex Rocket Man's Take On It'.

  10. I might suppose the SFR will be shorter and have a wider diameter to utilize its volume more efficiently, and provide a degree of stiffness to the structure to allow it to function in the various flight regimes.

    That being said, it seems a bit unfathomable (unless I am either reading SpaceX literature wrong or missed something) that they are not thinking in terms similar to the proposed "Shuttle C" (an orbiter like vehicle which would fly unmanned, allowing for more cargo or better orbital performance since it didn't have the mass of the flight deck, crew accommodations and life support systems aboard) or the "Shuttle Z", which was simply an orbiter sized payload fairing and a detachable engine pod to lift even larger and heavier payloads to orbit.

    Something along these lines would then allow for the SFR for manned missions, an "SFR-C" for heavier payloads or long duration space missions without crew, and an "SFR-Z" for maximum payload to orbit. While there wold be a larger upfront cost, the additional flexibility provides customers a larger degree of flexibility when designing their payloads or negotiation a flight contract.

    1. I think a deeper analysis of the SFR's mass is worthwhile. Shuttle-Z is very interesting!

  11. My name is Brendan Hennessy and I volunteer with a group in Chicago called Our group hosts webinars on interesting space topics and this blog post certainly qualifies! Would you be interested in being a guest speaker on one of our upcoming webinars? My thought is that you could cover in more detail the subject matter from this post. If so, can you reply here and we can figure out logistics offline? Thanks - and again great thoughts here in this post!

    1. Hi!
      Thank you for the offer, but I am not able to free myself from other responsibilities to speak in webinars.

      I can offer my help though, by answering any questions you or any of the webinar participants may have, and by allowing you to use the material in this blog freely, so long as you reference the origin.

  12. It seems to me that returnable Mars missions are going to require smaller shuttles to ferry astronauts to and from the surface, despite, or perhaps, because of Musk's overly ambitious ideas about fuel production on Mars. So, unmanned BFRs would populate the planet with supplies and habitats, while other BFRs would never land, and just shuttle people to SFRs which would land and takeoff. Some BFRs could be designed primarilyfor weightlessness interplanetary travel, while others would be drones designed to land in close groupings to start the habitats. The BFRs should be designed to have the remaining fuel removed to service SFRs or the occasional BFR liftoff until fuel production is a reality, decades from now (maybe). Once the fuel is removed, the BFR should be designed to have its lower section easily converted to quarters or storage until permanent structures can be built.

    In any case, I think this is a mandatory backup plan because I think fuel production on Mars is unrealistic for decades and something needs to be available for liftoff. The other option is sending fuel in a tanker version of a BFR, but that is a lot of waste. Better to leave the BFRs on the ground permanently and shuttle back and forth as needed to space for return journeys. No need to put the stress on interplanetary BFRs at all. They could be lighter and better shielded and carry the fuel for return trips, versus 1-way landing craft, or smaller 2-way landing craft.

    1. There is a halfway ISRU solution that Musk hasn't considered: bringing hydrogen to Mars and using to produce methane and oxygen directly from the atmosphere. Ars Technica has a good article on the challenges:

      SFRs designed solely for use on Mars would have to be redesigned slightly: their drag to weight ratio is lower (cube square law works against them) so they would need large wings. However, their deltaV and thrust requirements drop drastically, so fewer engines and propellant is needed. Even better: since the Mars atmosphere is so thin, you can take off directly using vacuum engines like the Raptors with the original huge nozzle.This gives you 375-380s Isp directly on takeoff.

      If the deltaV to reach Mars low orbit is 3.8km/s, an SFS would only need a mass ratio of 2.77. It could lift up to 91 tons into Low Mars Orbit!

  13. If the SFR would fit on the Falcon9 what capability would it have for a Moon shot with the Falcon Heavy?

    1. Sadly, an SFR as described in this post won't be able to make it to the Moon without in-orbit refuelling.