Orbital Access Methodologies Part I: Air Launched SSTO

As I mentioned last month, I would like to briefly discuss in a series of blog posts some of the more promising potential approaches for reusable orbital transportation. There is often a tendency among engineers to completely dismiss any idea other than ones own preferred approach as being unrealistic, naive, flawed, impossible, inefficient, etc. However, the more I’ve studied the problem, the more I’ve come to the conclusion that there are probably several technical approaches that can be made to work for providing reliable, low-cost access to orbit. Each of them has its own set of strengths, challenges, unresolved questions, and operating characteristics. By their nature, this means that different approaches may lend themselves better to different potential market niches and different development paths.

The first such approach I would like to introduce for discussion is epitomized by a proposed design (illustrated below, credit: Teledyne Brown) that was brought to my attention about a year ago. This proposed design, termed “Spaceplane” was developed at Teledyne Brown by Dan DeLong (who later became one of the founders XCOR Aerospace and is currently their Vice President and Chief Engineer, and who also currently owns all the rights to the Spaceplane design). Dan’s proposed concept was a winged, “assisted” single-stage to orbit (SSTO) design that was launched off of the back of a converted 747. The LOX/LH2 stage, powered by 1x SSME and 6xRL-10s would theoretically be capable of delivering ~14klb of unmanned cargo to a 400km circular orbit. The vehicle would be reusable, using an Inconel-foil over fiberglass insulation concept for its reentry TPS, and using a runway landing for its recovery method.


While the specifics of Dan’s proposed design are now a bit dated (the concept was proposed back in the late 80’s), the general approach still merits investigation.

To Stage or Not to Stage: That Is The Question

Now, before I go into the specifics of this approach, I know at least a few of you are probably already thinking things along the line of “SSTO? He can’t be serious. Everyone knows that SSTOs are totally unrealistic!” While to be honest, I’m mostly a TSTO guy myself (as is Dan DeLong these days), but I think there’s a real danger in how quickly and without contemplation people tend to buy into new conventional wisdoms.

The fundamental reason why anyone would even want to stage a rocket vehicle has to do with the physics of the rocket-powered flight. The rocket equation, says that the change in velocity due to a rocket in flight is linearly proportional to the specific impulse of the propulsion system and proportional to the natural logarithm of the vehicle’s mass ratio (the ratio of the mass at ignition to the mass at shutdown of the engines).

DV = Isp * g * ln (MR)

Another way of looking at this equation is that the required mass ratio of a vehicle is exponentially proportional to the required velocity change divided by the vehicle’s specific impulse:

MR = e^(DV/(Isp * g))

The inverse of the mass ratio is the dry fraction of the vehicle, ie. the percentage of the vehicle’s gross takeoff weight that can be allocated to structures, propulsion, payload, recovery systems, controls, power, life-support, etc, etc. The rest is fuel. Rewriting it in terms of dry fraction (df), we get:

df = e^-(DV/(Isp * g))

Now this is a fairly simplistic way of viewing things (ie. the Isp actually varies quite a bit with time based on the altitude at a given time, the engine throttle level, if you’re using thrust augmentation, etc, etc.), but shows the crux of the problem. The total delta-V needed to attain a low earth orbit can range anywhere from ~8-10+ km/s, while you’d be lucky to get a mission-averaged Isp much higher than ~400-440s even using the highest Isp propellants in service, LOX and LH2. Now there are all sorts of subtle nuances that we could go into. Things like how dense propellants typically require lower overall delta-V because they end up having less gravity and drag losses, or that depending on what latitude you’re launching from you can get a small “boost” due to the earth’s rotation. But the crux of the matter is that for a single-stage system, you’re dealing with a dry fraction of less than 10% (and typically quite a bit less than 10%).

That 10% has to cover all those categories mentioned above while still providing a high enough payload fraction that your system doesn’t have to get too gargantuan to deliver a sufficiently sized payload. And it has to be robust enough to be reused many times. And your system needs to be maintainable. And it needs to have graceful failure modes, and safe abort modes throughout the flight path. And it needs to be buildable on a realistic budget and timeframe.

All of those issues make the concept of staging very desireable. By staging you get to drop off some of your dry mass along the way, instead of having to lug it all up to orbit. This tends to relax the required mass ratios substantially, which makes it a lot easier to do all those things that make a reusable vehicle truly reusable (as opposed to recoverable, refurbishable, or scavengeable).

But that staging comes at a price. Staging creates a lot of complexity, and introduces some potential failure modes that can be hard to actually check-out on the ground. Staging is one of the single highest risks of failure for existing launch vehicles. Additionally, with a TSTO, now you’re really designing three vehicles, not just one. A first stage, an upper stage, and a combined entity. You now have to come up with abort modes for all the different configurations.

Probably one of the biggest headaches for TSTOs is how to recover and reuse the first stage. Getting to orbit is only a little bit about going up, and mostly about hurtling yourself sideways fast enough to “throw yourself at the ground and continually miss”. Doing so entails gathering quite a bit of horizontal velocity with a first stage, which means that the first stage gets quite a bit of horizontal distance between it and the launch site by the time it releases the upper stage. Most of the TSTO approaches I’ll discuss later revolve around how to get that first stage back. This is a real challenge for TSTO vehicles, though as Dan put it about SSTOs, they have their own challenges with getting the stage back (mostly due to trying to pack a robust heat shield and a robust structure into such a limited available mass budget).

So, in spite of the real challenges of developing SSTOs, there is a reason why some sane and rational people still look at them from time to time. There are real drawbacks to all approaches, and if an SSTO can be technically feasible, it might actually be desirable economically.

With that in mind, I’d like to get back to the topic of this post: air-launched “assisted” SSTOs.

The Benefits of Air Launching

One of the lessons I’ve learned as an engineer is that many times the best way to solve a really nasty and intractable-looking problem is to find a way to not actually solve that problem, but to replace it with an easier problem, and solve that one instead. In the case of an SSTO, trying to make a ground launched, horizontal takeoff and landing SSTO is a horrible challenge. You have very little dry mass to start with, and ground launching requires landing gear rated for the fully loaded weight of your vehicle, wings that have to be able to produce sufficient lift at very low speeds for takeoff, engines that can operate near sea level while still being efficient in vacuum (which entails either really high pressure designs, altitude compensations, or carrying around different engines with some optimized for high thrust at low altitudes, and some optimized for high efficiency in vacuum), and several other challenges. According to Dr Livingston, a Boeing engineer several years ago suggested that such a system was just not technologically feasible with modern materials and propulsion systems. While there have been some improvements on both fronts since he made that comment back in the mid-90s, I wouldn’t be surprised if a ground takeoff HTHL SSTO is still unrealistic.

So the real engineer finds a way to cheat.

And a good way to relax all of those constraints is to not try taking off from the ground, but to start at a reasonable altitude, by using a subsonic airbreathing carrier aircraft. Starting, as SpaceShipOne did, at a reasonable altitude gives several distinct advantages over ground launch (the following list comes from Dan DeLong, with some thoughts from me [in brackets]):

  1. The airplane carrier contributes to the overall altitude and velocity. These advantages are small. [Total savings are probably on the order of 100-200m/s. While this is a small fraction of the overall delta-V, the exponential nature of the problem means that even a small decrease in required delta-V makes a big difference.]

  2. Meteorological uncertainties are mostly below launch altitude. Propellant reserves can thus be less. [Or this means that you can fly on a more dependable schedule, and that you can have more robust propellant reserves without paying as much of a penalty for such.]

  3. Total integrated aerodynamic drag losses are less, as the launch is above much of the atmosphere. [This provides a bigger benefit to low density propellant combinations such as LOX/Methane or LOX/LH2, but overall could be worth several hundred m/s of delta-V, particularly for smaller vehicles]

  4. Max Q is less, which reduces structural mass, and may allow lower density thermal insulation. [You may also be able to "split the difference" on the structural mass somewhat--allowing for a higher FOS on the structure, which allows much less maintenance/inspection, while still pocketing at least some of the mass savings.]

  5. Engine average Isp is increased because the atmospheric back-pressure effect affects a smaller fraction of the trajectory. [This means that your mission averaged Isp is going to be much closer to your vacuum Isp than is typical for a booster engine.]

  6. Engine expansion ratio (non-variable geometry assumed) can be greater because overexpansion is less problematical. [For instance, IIRC, you can light an RL-10 at 30,000ft without risk of unsteady flow-separation caused by overexpansion. This can make a huge difference, as it means you can use an engine with a much higher vacuum Isp. Possibly a benefit of as much as 5-10%, with greater improvements seen by lower pressure systems that often have higher reliability than the ultra-high pressure staged combustion engines preferred for booster applications these days. When combined with benefit #5 above, this can have a large impact on the required propellant fraction due to the exponential nature of the rocket equation.]

  7. Wing area can be smaller because the wings do not need to lift the gross weight at low subsonic speed. Air launch Q is greater than runway rotation Q.

  8. Wing airfoil shape need not be designed to work well at high gross weight and low subsonic speeds.

  9. Wing bending structure need not be designed for gross weight takeoffs or gust loads. Wings can reasonably be stressed for 0.7 g working plus margin. This is a large weight advantage made possible by the carrier aircraft flying a lofted trajectory and releasing the orbiter at an initial angle of at least 15 degrees. (25 degrees is much better but not crucial, more than 60 degrees has no value) This initial angle decays in the first 10 seconds of flight but picks up again as propellant is burned and the constant wing stress trajectory yields a better lift/weight ratio. The thing to keep in mind is that the wings are sized and stressed for landing, and that insofar as they exist, are used to augment launch performance. [A comment I heard from a professor of mine back at BYU was that many people try to use composites as "black aluminum", i.e. they don't try to understand the nuances of the material, and thus miss out on most of the benefits. I think that that may often be the case with wings on rocket vehicles--if you design a vehicle to take the maximum advantage of your wings, you can negate some or all of the supposed "penalty" for carrying them in the first place. And that's coming from a VTVL guy!]

  10. Thrust/weight ratio can be smaller because the low initial trajectory angle does not have large gravity losses. This allows a smaller engine, propellant feed, and thrust structure mass fraction. I found 1.25 at release to be about optimum. This is a bigger advantage in air launching because total integrated aerodynamic drag losses are less and the trajectory need not get the orbiter out of the thick stuff as fast. [Lower gravity losses due to the flight angle reduces the required delta-V somewhat, and is probably a bigger benefit once again for high performance, low-density propellants, which typically suffer from higher gravity losses. Lower required thrust-to-weight is also big because your propulsion system is often a large part of the dry mass of an SSTO, so being able to get away with a lower required T/W ratio for the vehicle can make a large difference.]

  11. The lower mass/(total planform area) yields lower entry temperatures. I assumed inconel foil stretched over fibrous blanket insulation for much of the vehicle undersurface. Titanium over blankets, or no insulation worked on the top surface. Payload bay doors peaked at 185 F. [Having a better ballistic coefficient (the relationship of mass to planform area) means that your vehicle starts decelerating at a higher altitude where the atmospheric density is lower. Basically, drag force is proportional to area, while since F=ma, the acceleration is inversely proportional to mass.

    In other words, "Fluffy" is good for reentry vehicles, which means that by necessity, a fixed geometry SSTO is probably going to have gentler reentry heating loads than a fixed-geometry TSTO. This is increased by the fact that many of the benefits/constraints of air-launching push vehicles towards lower density propellant combinations like LOX/Methane or LOX/LH2. This is a good thing, because an SSTO has a lot less mass to cram that TPS system into. This is also good, because lower temperatures and more robust TPS systems mean lower maintenance, lower costs, and higher "availability".]

  12. Mission flexibility is greater. For example, the carrier airplane can fly uprange before release to allow a wider return-to-launch-site abort window. Good ferry capability, etc. [The other major benefit for missions to specific orbital destinations, like say a Bigelow station, is that the carrier airplane can move the launch point around. By being able to place the launch point at just the right position relative to the station, you can provide for first-orbit rendezvous opportunities even if your launch site isn't directly underneath the given station. The ability to move the launch point also potentially opens up longer launch windows. Lastly, being able to move the launch point allows options like operating out of an airport closer to "civilization" while still launching out of an area with low population density, like say over an ocean or a desert.]

  13. [Update: A commenter noticed that Dan and I both forgot to include an important additional benefit of this approach--landing gear for an air-launched SSTO can be designed based on landing weight instead of takeoff weight. This is a big deal for SSTO designs. Boeing had another proposed design, RAS-V that used a trolley for takeoff, but would probably be pretty dicey for an abort. Dan also mentioned the point I forgot to bring up that the RL10s on his design could be used to establish a subsonic cruise of a respectable distance, so you wouldn't actually dump propellants, you'd burn them off in your smaller engines. All in all this ability helps Mass Ratio substantially since the landing gear for a ground takeoff HTHL SSTO is typically a large chunk of the dry weight of the vehicle.]

As can be seen from this list, by “cheating” a little bit on the boundary conditions, assisted SSTO approaches can avoid many of the typically largest drawbacks of ground-launched SSTOs. What was a probably intractable problem before (ground-launched HTHL SSTO) becomes a lot more feasible by adding the air-launch “assist”. Now, technically you could say that the carrier airplane in an air-launched “assisted” SSTO is really a stage, and therefore the idea isn’t really SSTO–and you would be technically correct. But, I do think there is a fundamental difference between an airbreathing carrier plane and a true first stage, such as: no worries about TPS for the carrier plane, no need for RCS systems, no need for rocket propulsion (probably), no need for high propellant fractions, etc.

So all in all, there’s a fairly compelling case that if you’re interested in developing a SSTO vehicle, and a winged one at that, that air-launching is a big win over ground launching.

The Constraints, Challenges, and Drawbacks of Air-Launching

But as with everything in engineering, air-launching is not without its constraints, challenges, and drawbacks. While I’m sure that someone like Dan DeLong, or Antonio Elias of OSC could probably do better justice to this section than I could, I’ll try to touch on some of the high-points:

  1. There are a limited number of existing aircraft designs that can be used for air launching. What this means is that the design space for gross takeoff weight vs. carrier price is not a smooth continuous function. If you are near the upper limits of a given carrier craft, even a small increase in takeoff weight might end up forcing you to use a much larger carrier craft.
  2. Most existing aircraft aren’t that great for air-launching large vehicles. If you drop the vehicle from beneath most commercial aircraft, you’re very limited on maximum volume beneath the wings or the hull. If you launch off of the back of an aircraft, you now need to have a higher L/D wing (or light the engines before separating) so as to not collide with the carrier after separation. Also, if you use a top-launched configuration, now you have to mount the stage on top of your carrier, which requires a substantial amount of ground handling equipment (compared to a bottom-dropper).
  3. Due to needing to fit on an existing carrier aircraft, air-launched SSTOs are a lot more Gross Take-Off Weight (GTOW) limited than ground-launched SSTOs (which can grow to arbitrarily big sizes).
  4. Related to point #3, there are certain systems on a launch vehicle that don’t scale down very linearly. There are also minimum gage issues. These two realities mean that as an SSTO gets smaller, the maximum achievable mass ratio for the system gets worse and worse. Below some minimum size, it’s no longer possible to reach orbit with any appreciable payload at all. I’m not positive where that exact point is (and it probably depends on a *lot* of details, but it is probably in the ~50klb range.
  5. This is still an SSTO, and even if you cheat by air-launching, you still have a very demanding mass ratio to meet while still making the system robust enough for reuse.
  6. Air-Launching a cryogenic propellant stage requires either very good insulation, or some sort of propellant storage capabilities on the carrier craft, or at least some sort of propellant conditioning equipment (ie something to pull heat out of the propellants and prevent them from boiling off). Or possibly all of the above.
  7. Due to upper limits on the size of available carrier craft, this concept is unlikely to be scalable to payloads much bigger than 20-25klb.

Now, none of these are necessarily deal-killers, but its important to know a design choice’s drawbacks.

Potential Enabling Technologies

There are a couple of recent technologies that could make a vehicle like this a lot more realistic than back when Dan DeLong first developed the concept. Specifically, cryogenic composite tank materials, some advanced cryogenic insulation techniques that are under development, the White Knight series of carrier aircraft, thrust augmented nozzles, and orbital tugs.

First off, cryogenic composite tank materials (such as XCOR’s “NonBurnite” flouropolymer matrix composites) allow for somewhat lighter tank masses, allow for cryogenic “wet wings” if desired, and allow for insulation and the tank to be integrated into the vehicle structure.

The advanced cryogenic insulation technique I mentioned would help a lot with reducing/eliminating boiloff issues for cryogenic propellants (particularly LH2 if you go that way). I can’t really go into the specifics on this approach quite yet. I had written an SBIR proposal for pursuing this technology (along with some teaming partners in industry), but we barely lost out, so it may take a lot longer before the idea is proven out. Suffice it to say that it could cut down on boiloff substantially in gravity, and even moreso in microgravity. Keep your fingers crossed.

The benefit of the White Knight series of carrier aircraft should be obvious. Having a large carrier aircraft with a high undercarriage that is purpose-built for carrying large rocket powered vehicles is immense. I don’t have exact specs for WK2 (I figured it would be really bad form to try and pump my friends on the Scaled Propulsion team for such info), but my guess is that its at least 40klb, and possibly as much as 60-80klb. Depending on the exact numbers it might be just barely big enough for a fully orbital SSTO, though I’m not sure how much payload you could get with a vehicle that small. I really don’t have a great feel for how the scaling performance for the SSTO works. There have also been several rumors (from all sorts of sources) about the possibility of a White Knight 3 down the road. T/Space showed such a vehicle in their original presentations. That would likely be capable of carrying a booster in the 300-500klb range, which is about the weight of Dan’s original “Space Plane” proposal. The benefit of using a White Knight 2 or 3 for your carrier plane (above and beyond being able to buy an airplane that is purpose built for air-launching) is that the SSTO wouldn’t be the only customer for the carrier aircraft. Which means the SSTO would only have to pay a fraction of the amortization costs of the WK2/3 development. More importantly, if you can get away with something like WK2, there may very well be several of these built for Virgin Galactic (and other customers), which means that the unit price of the airplane will be lower, parts will be more available, there will be a larger operational/maintenance experience base for it, and depending on the required flight-rate, it might even be possible to just rent a WK2/3 from a SS2 operator instead of having to own one outright.

Ok, I’m sure I’m starting to sound like I have a bit of a hobbyhorse thing going, but I think that thrust augmented nozzles would be a very good match for an air-launched SSTO. Especially if they were running in a “tripropellant” configuration (ie with the fuel in the thrust augmentation section being a denser fuel like kerosene, methane, or subcooled propane). The first big advantage is that it would allow an engine with a much higher thrust to weight ratio compared to a more traditional engine. This would allow for a much lighter engine to be used, which directly translates into more mass for the rest of the vehicle (and the payload). Another benefit is that depending on the fuel used (and the construction technique for the wings), a “wet-wing” tank could be used for the TAN fuel, which would allow a lot more fuel to be carried at almost no extra dry-weight. Combine this with the fact that the LOX tank would be bigger, and the LH2 tank smaller, and it ends up giving you a much higher achievable Mass Ratio for a given construction technology. Using TAN, you can also get away with a larger expansion ratio on the nozzle, giving better Isp after the TAN propellants burn out. Also, if the TAN injectors are broken up into quadrants with separate valves, they could possibly be used for Liquid Injection Thrust Vector Control. This would eliminate the need for the gimbal, and possibly allow for the now much bigger rocket engine to package better into the rocket vehicle. Lastly, if the thrust augmentation is light enough, it might allow for the possibility of keeping some “go-around” propellant for increased landing reliability. While adding the denser TAN propellant doesn’t give quite the same drag and gravity loss benefits as it would for a vertical ground launched vehicle, it would still likely increase the payload fraction for the vehicle at a slight increase in GTOW. Aerojet was estimating, IIRC, a 3x increase in payload for a less than 50% increase in GTOW.

Lastly, space tugs (possibly based on the Orbital Express design, or possibly based on the Loral/Constellation Services tug designs) could greatly help such a system if it turns out to have lower performance than hoped for. Instead of taking the payloads all the way to their destination, a tug could possibly allow the SSTO to place payloads into a much lower temporary orbit (which would increase payload mass). Having a tug would also reduce the mass and complexity of the SSTO as it would no longer need its own rendezvous and docking hardware. Also, having a tug means that the cargo (or propellant) could be stored in generic containers, which would simplify ground handling and payload installation. A pressurized tug would be necessary if you wanted to fly people on the spaceplane, but that isn’t too unreasonable.

All of these new technologies, most of them which have only come out in the past 5 years or so, make a system like this a lot more feasible today than back in 1986.

Preferred Instantiation

[Update: I also forgot to include this section in the original post]

While I think Dan’s original design provides a lot of useful ideas, I think that my preferred instantiation of an air-launched “assisted” SSTO would be a lot smaller. After Space Plane, Dan also went the direction of a smaller vehicle–one he called “Frequent Flyer”. I don’t recall the exact specs for that design, but they were around 40-50klb GTOW, and required a solid strapon “0th stage” to provide enough thrust. I’d go instead with a wet-wing tripropellant design using kerosene in the wings burned in a single RL10 modified for LOX/Kero thrust augmentation. The gross weight would go up a bit, probably to up around 70-75klb (which is hopefully below the upper limit of what White Knight 2 can carry–I don’t know for sure), but you’d get a better mission averaged Isp, would have a fully reusable system, and would probably increase the payload a bit over the Frequent Flyer. But Dan would be in a much better position to say. My goal with this instantiation would be basically either two people, or 1-2klb worth of cargo to LEO. If a vehicle this size works, and if it can fit on WK2, it would be possible to do a larger follow-on using something like a WK3 down the road.

That’s just my opinion though.

Remaining Unknowns and Some Potential Paths Forward

So the question becomes, where are we at now with regards to this concept? What unknowns are that we currently know about? Where do we go from here?

The key “known unknowns” I can think of include:

  1. TPS design and reentry aerodynamics–is it feasible to make a reusable TPS system that will work for this vehicle that is robust enough, and what moldline/airfoil design will provide the best balance of needed subsonic performance, and workable hypersonic aerodynamics?
  2. Cryogenic propellant tanks and insulation–can tanks be designed that are both light enough and robust enough for the application? Can long-lifetime cryogenic composite tanks be built that work at LH2 tempeatures? Can an insulation technique be found that is adequate enough to prevent boiloff during the ferry to the launch site? Do we need to use some form of subcooling, propellant conditioning, or “top-off tanks” on the carrier plane?
  3. Thrust Augmentation–can thrust augmentation actually deliver enough of an advantage to justify its use in this application? Can an existing engine (such as an RL10 variant) be readily modified for use with thrust augmentation? What is the optimal augmentation level? Does the better payload fraction provided allow you to use a smaller vehicle? What TAN fuel is best? Kerosene? Subcooled Propane? LH2? Can the thrust augmentation be combined with an LITVC system? Does that gain you anything? Can you adequately control the CG shift during flight with the TAN fuel in a “wet wing”? Could an RL10 type engine operating on “vapors” in turbine bypass mode provide enough of a core flow to ignite the thrust augmentation for a go-around burn at landing? Or would you need separate go-around thrusters?
  4. Vehicle Sizing–what’s the smallest vehicle size that can reasonably deliver (with margin) the payload in question? What are the actual carrying capacities of WK2 or 3? Would such a minimal vehicle be small enough to fit under WK2, or would WK3 be necessary?
  5. Mass Ratio–what mass ratio would be required for the vehicle? Based on existing technologies, how feasible is that mass ratio to attain? Is the required mass ratio more doable using denser propellants, and if so, can a denser propellant vehicle still keep a low enough GTOW to fit on potential carrier planes?

To me, the most critical questions that are also the most unknown, are the ones regarding the TPS and reentry aerodynamics. Most of the other questions, while important, are much more straightforward to answer.

As for the path forward, I think there are multiple prongs that can be taken.

First off, for the carrier plane, WK2 is mostly built and will probably be flying this year. More exact information about its maximum carrying capacity can probably be had in the relatively near future. Trying to find a way to make a vehicle that closes using WK2 would be the most preferable option.

Second off, the TPS/Reentry Aerodynamics. Some of this can be worked on the “traditional way” using CFD and special wind tunnels (at places like NASA Ames). However at some point, it would probably be worthwhile to move on to subscale models launched from suborbital vehicles. Basically, a suborbital vehicle with a “nanosat launcher” upper stage could probably put up a small, instrumented reentry model to nearly orbital speeds. A lot of care would be necessary in designing the experiment and analyzing the data to get the actual data you want, because there are all sorts of scaling laws going on a the same time. Things like different reynolds numbers, the fact that the standoff distance of the bow shock is going to be proportional to the linear dimensions of the vehicle, so a subscale model is likely going to see more intense heating, etc. It should be possible to design a series of low-cost experiments though that can at least retire some of the risk in advance before trying to build an operational version.

As for overall vehicle integration and Mass Ratio control issues, an HTHL vehicle like Xerus actually provides a useful starting point for working ones way up to an SSTO. Now, the XCOR people aren’t SSTO fans. And they’re especially not LH2 fueled SSTO fans. But, the best approach for trying this would probably be to hire someone like XCOR to try and build a lower-performance iterative prototype first to test out some of the key functionality, and then work your way up to the performance needed for SSTO. The first prototype might use just a traditional LOX/Methane engine with as high of a mass ratio as possible. Make sure that the handling and basic aerodynamics work out right. Test out the cryo insulation, and air-launch cryo-propellant handling procedures. Make sure that the TPS functions as expected in the suborbital (though relatively high velocity) environment that such a vehicle could provide. Upgrade the engine to a TAN system and get some experience operating that and making sure that the LITVC scheme works. Test out RCS functionality. Test abort modes.

Do a second iteration that has LH2 as well as the TAN propellant. Develop and test out a TAN-modified RL10. Get experience using such an engine. Get in-flight performance data. Make sure the cryo insulation still works. Make sure the tank can handle the cold cycles. See how close you can get to the mass ratio. Instrument the crap out of your vehicle and figure out where you can shave weight, and how robust/reusable the TPS is in an almost orbital situation. Figure out if you need to scale up the vehicle, or what other changes will be needed to reach a sufficient payload target. Start expanding the envelope to orbit.

Now some of these steps might be skippable depending on how previous steps go. Some of them might be doable as part of other programs (for instance figuring out the cryo composite tanks for LH2 or the special insulation system might benefit other projects, developing flightweight propellant conditioning hardware that can fit on a WK2 or 3 might also be useful for other projects). But these are just some thoughts on what has to be done from a technical standpoint to get “there” from “here”.

Conclusions

In spite of the bad reputation that SSTOs have earned during the last decade, there are at least some versions, like the air-launched SSTO that aren’t entirely crazy. They still might not make sense, but if any SSTO RLV design ever makes it, my guess is it would likely be something like this.

This entry was posted in Launch Vehicles, Orbital Access Methodologies, Space Transportation, Technology. Bookmark the permalink.

68 Responses to Orbital Access Methodologies Part I: Air Launched SSTO

  1. tom mattison says:

    Hi

    Sorry I’m late to the party. Here’s a thought: towed glider launch.

    Build an remote-piloted glider that straddles the rocket. Design for easy integration of the rocket. Long simple wings, no engines, strong landing gear. Use a tow plane to get the glider+fueled-rocket up to altitude. Release the glider from the tow plane, get a little separation, then release the rocket from the glider and light it. The glider is remotely piloted back to a landing site. If that isn’t the glider takeoff site, you land a tow plane near it and tow it back to the usual glider takeoff site at your convenience. Since there’s no rocket for this trip, and you don’t need to go high, you could probably use a smaller towplane for this trip if it saved you money.

    At the cost of designing a minimal airframe, you get a vehicle optimized for your airlaunch, and have freedom of choice for the big tow plane, which is basically just providing engines.

    I imagine that the sudden weight change on releasing the rocket from a plane is a major issue. It’s a safety and expense issue if the plane is a manned. There are also issues with the rocket colliding with the plane after ignition. But releasing a glider from a manned plane should be no big deal.

    You could perhaps release the rocket from the glider without releasing the glider from the tow plane, but I wouldn’t do that until it was proven you could keep the glider under control so it wouldn’t pull the tow plane out of control.

    It’s true that the same issue of controlling the glider during the release of the rocket, and avoiding collision afterwards, is an issue for a glider as well. But the glider could be unmanned, and cheaper than a big jet (which would be nice since you may trash a few of them in the process of learning to airlaunch).

  2. Anonymous says:

    I’d like to add my observations.

    My favored design uses a 747-200 as the carrier aircraft, and carries the rocket slung under the belly. This clearly implies takeoff using a ‘Tall Trolley”, with the landing gear up.

    Let’s look at the operations of this concept a little more closely. The empty 747-200 sits on a runway, and is approached by two strange V-shaped trucks, one in front and one behind. The trucks are attached together by ground crew, completely encircling the plane’s landing gear. The trucks hydraulically extend support poles, which connect magnetically with weight-supporting ‘hard-points’ located near the landing gear. The entire plane is then fueled up, then lifted about six meters or more above its normal ground position. The landing gear is then raised. The empty rocket is then lifted into position and attached to the 747 using a web of kevlar ropes. The rocket is then filled with LOX and LH2, after which the plane immediately taxis for takeoff.

    The ‘Trolley trucks’ have powerful gasoline or diesel engines, and are capable of accelerating their own mass as fast as the 747 during a MAX GTOW takeoff. When the 747 develops sufficient lift, the magnetic coupling is disengaged and the trolley begins rapid braking.

    It doesn’t appear (to me) that this design has any major technical difficulties.

    I tried to come up with other uses for a ‘bottom-dropping’ 747 cargo aircraft with over 100 metric ton capacity, and I immediately came up with the military use of an oversize fuel-air bomber. If you can get efficient dispersal of 100 metric tons of LH2, the explosive potential is 3.4 kilotons.

  3. Chiya says:

    Even though I didn’t understand all of it, it just makes sense that SSTO would be simpler and simpler would mean less could go wrong with it.

    Anyway thanks for explaining it in a way that I could understand part of it at least!

  4. Anonymous says:

    White Knight 2 has payload 30,500 lbs according to its Wiki entry and from what I’ve heard this is correct. What impact would this have on what it is able to launch? Could it launch an LEO?

  5. Jon Goff says:

    Anonymous,
    30klb sounds a bit light, especially since that’s lighter than the numbers I’ve heard for SS2. I could be wrong, but there may be a mistake there. If it was only 30klb, there’s probably no way to do an SSTO that light, but you could do a non reusable TSTO…but what would be the point?

    ~Jon

  6. Randy Campbell says:

    Jon;

    I mentioned this in the comments for the second article but thought I’d comment here also:
    The “SwiftLaunch” concept is a “stage-and-a-half” vehicle with a reusable engine/OMS/with cargo or passenger accomidations attached to a set of expendable tanks with LOX/RP1.

    Dropped out of the back of the C5 or AN-225 attached to “Sled” that is extracted from the aircraft cabin and then supported by SRB parachutes in the near vertical launch position. Once the vehicle launches the SRB parachutes lower the sled to a water landing.
    The vehicle expends all the fuel in the external tanks and drops them short of orbital velocity, with the OMS propuslion inserting the vehicle into orbit.

    While I think that the ‘sled’ can be reduced or even eliminated, (the people who proposed the SwiftLaunch system now work for Air Launch) the need for the Air Launch vehicle to be sized to allow it to tip out of the aircraft rather than being pulled as in Swiftlaunch, in my opinion are sevier enough to make such an operation, while greatly simplified, much less adaptible for larger sized payloads such as those the Swiftlaunch proposed.

    What would be your take on such a system?

    Randy

  7. Mark Horning says:

    Jon, wonderfull post, looking forward to the rest of the series.

    Several of the advantages of airlaunch also apply to the air to air propelent transfer (AAPT) proposal Mitchel Burnside-Clapp was pursiung several years ago.

    I had a long conversation with him and the two points I remember distinctly were.

    1) landing gear is 3% of your Gross Weight. With AAPT you only need to accomidate the partially fueled take off weight. You also save some structural weight due to lower wing loading.

    2) HTHL or VTVL didn’t matter. The performance difference was down in the 3rd or 4th digit. The important thing was designing a VTHL (ala Shuttle)was the dumbest thing you could do, because it combines the weight penalties of both systems. The main reason he wanted to go with HTLH was he thought there was a greater pool of engineers to draw from.

  8. Vladislaw says:

    What does air launching do to the costs relating to ground launch facilities? How much are the relative savings in launch costs because you do not have to maintain a ground launch facility?

  9. Jon Goff says:

    Vladislaw,
    What does air launching do to the costs relating to ground launch facilities? How much are the relative savings in launch costs because you do not have to maintain a ground launch facility?

    I don’t have enough info to give specific answers, but yes I agree that intuitively it should be a lot cheaper. Especially if you’re doing airlaunch from an underslung carrier vehicle like a WK2 or WK3. At that point, your ground facilities would be little more than a traditional aircraft maintenance hangar, some fueling trucks, and maybe a small cleanroom depending on what sorts of payloads you were flying. You could probably even operate out of a minimally prepared airstrip if you had a customer with such a need…

    ~Jon

  10. Mike Täht says:

    You seemed to miss the idea of partial in-air fueling in this piece.
    (I guess you love hydrogen). Maybe you did, I only just scanned it again. or maybe kerosene isn’t energetic enough…

    With partial in-flight fueling (loading just the cryogenic oxygen on the rocket before takeoff, and the kerosene in flight) You get a low takeoff weight for the carrier aircraft, a faster climb to altitude, less wing loading until you get to altitude and fueling, a
    better abort scenario for all but the final stages of fueling, and a well
    understood problem (in-air fueling), that uses existing production aircraft.

    A KC-135 has a ceiling of 50,000 feet, and can carry 200,000 lbs of fuel
    - RP-1 is not all that different from the normal loads…

    http://www.aviationexplorer.com/kc-135_facts.htm

    Cryo refueling in-flight would be a bitch. Refueling with hydrogen defeated kelly johnson back in the 60s, and you have major problems with boil-off with cryogenics of any sort… a shortened take-off to altitude procedure would help… and if both your propellants are on board at take-off, your manned aircraft is a bomb on the runway… far safe to have a bomb 50k up.

    Just a thought…

  11. Jon Goff says:

    Mike,
    In-air refueling would definitely add some performance, but in a LOX/RP system, the RP makes up only about 1/3 of the propellant mass. While it will make a difference, it won’t be an earth-shattering one, and may very well add enough complication to sink the deal (except when you really absolutely need that last little bit of performance for a mission).

    ~Jon

  12. Pingback: Selenian Boondocks » Blog Archive » Air-Launch Paper

  13. Scott Holman says:

    There is something kind of eerie about seeing what you have been proposing for months given a very rational, scientific treatment, with a positive bent. I am not going to bog myself down in the semantics of whether a carrier wing constitutes a ‘stage’ or not, because that has nothing to do with making the proposal a reality. My greatest reservation about the post is the use of an ‘existing’ aircraft for a carrier wing.

    In order to achieve a practical payload amount, the space plane, or orbiter, has got to be larger than any existing aircraft could possibly carry to a useful launch altitude. Fully loaded, the orbiter is going to weigh around 750,000 pounds, with a dry weight of about 120,000 pounds. These weights are Wild Ass Guesses, based upon a propellant combination of kerosene and liquid oxygen, which are the only ones that have been proven in flying vehicles, and which can be carried on the orbiter for several hours without boiling off. Another necessity is to minimize the altitude requirements for the orbiter. So it has got to be part of a larger system, which would include what I have referred to as an Orbital Transfer Vehicle, as well as a space station where passengers can be transferred from the OTV’s to space ferries, lunar shuttles, or some other vehicle heading somewhere beyond Low Earth Orbit. The Orbital Transfer Vehicle is a logical extension of the advantages a vehicle based in space has for carrying cargo to higher altitudes over a ground launched vehicle.

    Requiring that some orbital space plane be carried aloft by an existing aircraft is a deal breaker, because there are no such aircraft. Just as no one had seen an aircraft such as the White Knight, the carrier wing needed for launching a true space vehicle is something which must be designed from scratch. One prerequisite is that the wing will carry the orbiter on its back. An orbiter of useful size will simply be too big for a wing to straddle, plus the combined vehicle has got to be supported somehow, if an extremely heavy undercarriage is to avoided. Which brings up another aspect which has no precedent.

    When an extremely heavy aircraft takes off, it requires very long runways, with no obstacles for many miles beyond the end of the runway. The stall speed of a fully loaded aircraft can be significantly higher than when the aircraft is empty. For an aircraft which could easily weigh nearly two million pounds, a runway 5 miles long might not be adequate, unless the vehicle is massively overpowered. To provide a support system for the carrier wing, and to provide a means for that carrier wing to reach take-off speed as rapidly as possible, a catapult is needed. This catapult will only accelerate the carrier wing to a speed of, say, 350 miles per hour, but that would probably be considerably higher than the wing’s stall speed. And the catapult could achieve that velocity in a very short distance, just as an aircraft carrier’s catapult throws an aircraft off the deck at something close to 200 miles per hour.

    Properly designed, the catapult offers an excellent abort option, at least for the take-off of the carrier wing. By extending the catapult well beyond the length needed to launch the wing, the ability to slow the wing down from launch speed is created. Also, the cradle of the catapult could be outfitted with grapples which will not allow the wing to launch until it has sufficient lift to climb at 500 feet per minute.

    The design of the wing has got to be focused on the single goal of lift, with speed being well below sonic range, because speed and lift tend to be incompatible in large quantities. To this end, we may see a biplane design emerge, which might make launching from the back of the wing easier, as the orbiter would be carried above the main wing. Another concern about air launching is that the orbiter can begin climbing immediately after separation. Space Ship One was dropped from about 50,000 feet, but passed through around 35,000 feet before it was able to begin climbing, as the engine was not fired until the vehicle had been dropped. With the proper refractory materials, innovative design, and maneuvers at separation, the orbiter would be able to start its engines while still on the carrier wing, and run them up to full power for a few moments, so that it will be climbing away from the carrier wing, instead of falling towards it. The carrier wing would execute a steep dive immediately upon separation, to minimize the chance of collision.

    Your article made an excellent point regarding the lift available from the orbiters wings during the ascent. Those wings that will provide lift at landing speeds can still provide lift in the upper atmosphere at a couple of miles per second velocity, and such lift would enable the orbiter to climb faster than if it were to rely entirely upon engine power to climb away from the Earth. If the orbiter is only traveling at a few hundred miles per hour at 50,000 feet, aerodynamic lift is usable. Increasing velocity will increase altitude, which will allow more velocity. The value of maximum aerodynamic turbulence will diminish as altitude increases, even when velocity is increasing more rapidly than the altitude is, if one begins the launch at very high altitude.

    Even the most optimistic projections will only allow the orbiter to handle a payload in the 20,000 pound range, but that would be adequate to carry a dozen people into orbit, with life support for several days, if that orbit is only about 120 miles high. However, it may be possible to transport the passengers in a module which could be handed off to the Orbital Transfer Vehicle, which would eliminate the need for bulky airlocks and docking equipment. The Orbital Transfer Vehicle would carry a large module which would allow the passengers to exit the launch module once it had been transferred to the OTV, and this module would provide the life support for periods of up to a week, or even 10 days, to allow for aborted launches, or having to remain on station in low orbit to pick up multiple launch modules.

    This entire concept has one purpose only, and that is to provide transportation for large numbers of people into space. We can lift cargo into space right now, and will be able to do so more and more cheaply as the amount of cargo goes up. But we have no technology for transporting large numbers of people into space and back, without which the entire enterprise will falter. Scientists, technicians, engineers, fabrication specialists, and cooks will all have to go into space if we are to be successful in pushing back the boundaries of our activities.

  14. Vincent says:

    Hi Jon, and thanks a lot for this fantastic article.
    I’m currently studying the feasibility of air-launched SSTO, and I have a question about the thrust/weight ratio of 1.25 that you give here, it seems that it was given by Dan DeLong but I don’t find where.
    “Thrust/weight ratio can be smaller because the low initial trajectory angle does not have large gravity losses.” From what I understand, a small pitch angle, i.e. a quite horizontal flight, which is due to the fact that a subsonic carrier aircraft is considered, means that it will require more thrust to counter the gravity. As Dave Salt said (comment 41), it requires much power in aircraft engines to have a high pitch allowing the rocket to climb right after the rocket release. Moreover, a higher thrust implies a shorter burn time, and thus a lower gravity drag.
    I have done some simulation, and I find that for a subsonic aircraft and a 340s Isp for the rocket, the best thrust/weight ratio is around 2.5, without taking into account the effects of aerodynamic lift (nor drag). See this graph: http://lcas.otaski.org/index.php/Rocket:First_approximations#The_gravity_drag_and_overall_thrust-to-weight_ratio_issue . You can see other graphs showing the effect on one parameter on the final mass in the page too.
    Would you have any link that talks about the 1.25 ratio of Dan DeLong? Am I missing something in the process? Are you still reading this page? Thanks!

  15. Jonathan Goff says:

    Vincent,

    I think the 1.25T/W ratio came from conversations with Dan–I worked just down the road from him at the time I wrote the article, and I think I had lunch with him before writing the article to get some details. My guess is it came from his simulations and analysis. I’ve heard other numbers in the 1.4 range from Kirk Sorensen for the the ALTO air-launched TSTO concept that had a winged first stage. I’m no longer out in Mojave, but it might be able to ping Dan to see if he has more details somewhere.

    ~Jon

  16. Vincent says:

    Thanks for the answer. I guess that having small wings on the rocket is the key to such low T/W ratios. Being able to efficiently climb in the first part of the flight, where air can still help, would allow the pitch angle to be lower and the trajectory smoother. Note that the Pegasus has wings and still a high ratio, around 2.5-3.
    If you chat with Dan someday, that would be nice to ask him indeed.

  17. john hare says:

    FWIW, my understanding is that the engine constitutes much of the mass and cost of each stage. So you accept considerable gravity losses early on to reduce engine caused mission cost. At altitude, the 15% or so greater thrust available from a given engine allows considerably more GLOW per engine cost. The altitude also allows near horizontal acceleration almost immediately. A 1.25 T/W ratio would allow a 7m/s horizontal acceleration with some slight vertical acceleration. The altitude launched vehicle can wait until some fuel has been burned off to increase the vertical component. A ground launch under the same conditions would restrict the vehicle to 0.25m/s vertical acceleration of a vehicle that is 15% or so lighter to begin with.

  18. Vincent says:

    Thanks John, I understand better now. In my mind it was a bad idea to gain horizontal speed too low in the atmosphere because of the aerodynamic drag, so I was looking how climbing could start right after release. I understand the cost consideration too, I have not looked into that either.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>