Orbital Access Methodologies Part VI: Air-Launched Glideforward TSTO

Back when I first gave the guest-lecture at the University of North Dakota that kicked off this series, I had only introduced four actual technological approaches to making RLVs work. The balance of the time I spent talking about the economics of reusable orbital transportation (and the development process for getting from here to there). When I thought of doing this blog series, I figured I’d just give some quick intros those four approaches, and then I’d go into what I thought at the time was the more interesting part–my thoughts on the business ramifications of all this technical stuff. However, this series has somewhat taken on a life of its own, and I ended up going into a lot more detail on each of these ideas than I did during my lecture. If you would believe it, my notes for the Air Launched SSTO that I discussed in Part I of this series took up only two 3×5 cards front and back!

While fleshing out the technical discussion of these approaches, I also got a lot of useful feedback from commenters and some experts in the industry. In my second post, one of the commenters (a good friend of mine, John Hare) pointed out a fifth RLV approach that I had completely overlooked–airlaunched glideforward TSTO. I plan on making that concept the subject of my last post in the series on RLV approaches–even though there may be other approaches worth discussing, I think the discussion on these five approaches should suffice for now.

The First Stage Recovery Problem: A Review
As we discussed in Part II, one of the key challenges of TSTO RLVs is finding a way to get the first stage back to your launch site without requiring a lot of extra people and infrastructure. It’s also important to get the stage back in a manner that facilitates rapid reprocessing so you can turn the thing around as quickly as possible. While it’s totally possible to do a down-range recovery first stage, and while that may actually be the preferred mode for ELVs trying to transition over into the world of partial reusability/refurbishability, it definitely adds a lot of additional overhead to the operations (more people for running the down-range recovery and return operations), and slows down your possible op-tempo (since recovery and return now take a significant amount of time), as well as placing big restrictions on where you can launch from and to.

In Parts III-V, we discussed the three TSTO options that I had mentioned in my UND lecture. First we talked about pop-up stages, where the first stage intentionally never gains much if any horizontal velocity, so it tends to land back where it came from. Then we talked about techniques that gave horizontal velocity to the upper stage as well as altitude and dealing with gravity losses. We talked about techniques that either used aerodynamic gliding, or rocket propulsion to turn the first stage around and return it to the landing site. The problem is that after a certain amount of downrange velocity (about Mach 3), you end up having to spend a lot of propellant on getting the first stage back. This is propellant that can’t be used for performing the primary mission of putting a payload into LEO. While this mass requirement is worthwhile from the standpoint of easier operations and lower required headcount and facilities, it does come at a respectable cost, which ends up requiring a much bigger vehicle system to place a payload into orbit than a comparable ELV. Ideally, what you would really like to do with an TSTO RLV is to have all of your propulsive efforts going toward putting the payload into orbit, just without the maximum burnout velocity restriction of the glideback TSTO concept we discussed in Part IV. You want the propulsion-free RTLS maneuver without all the restrictions that would normally imply. By launching your TSTO from a carrier aircraft, there is a way to achieve just that.

Air-Launched Glide-Forward TSTO
The concept is actually quite straightforward–John Hare’s suggestion was to launch a two-stage rocket off of a carrier aircraft at a location far enough up-range of the originating spaceport to allow the first stage to reenter after staging, and glide-forward to land back at that same spaceport. This concept manages to meet all three goals discussed earlier: the first stage has RTLS capability, the rocket stack can use performance-optimal staging without the constraints of the glideback concept, and there is no wasted propellant associated with the RTLS maneuver–all of the first stage propellant except for a tiny landing reserve is used for putting the upper stage onto the trajectory it wants.

One of the people who I had read this article before I posted it mentions that launching from the carrier ship at a position uprange of your landing site makes a lot of sense even for air launched SSTO designs. The ability to make it back home in case of an aborted launch is going to be very important for making either SSTO or TSTO air-launched designs work.

Advantages and Drawbacks
While this concept shares some of the advantages and disadvantages of the Air-Launched Assisted SSTO concept, there are a few unique pluses and minuses.


  • Obviously, by going with two rocket powered stages, the required performance of the orbiter stage is relaxed a bit, making it much easier to build.
  • Depending on how much you need to relax the required performance of the upper stage in order to make the design close, the first stage may actually be in the same performance class as a suborbital booster (such as XCOR’s Lynx). Or, if you go with a more aggressive first stage, it makes the orbital stage mass fraction a lot more achievable.
  • Between the performance boost of air launching, and the performance boost of being able to stage along the way, this combination probably ends up with the minimum GTOW (for the rocket stages) of any realistic configuration. While GTOW isn’t everything (dry mass tends to be more important from a cost perspective), it means that you might be able to actually close the case for a manned orbital RLV using a carrier plane as small as WK2. Since the carrying capacity of WK2 is only 35klb (according to a little bird who works on the other side of Sabovich St. in Mojave), as opposed to the 60klb number Wikipedia was providing previously, a manned air-launched SSTO wasn’t going to work.
  • By keeping the stages both relatively small for a useful orbital vehicle, fabrication and testing facilities can also be kept small. For a minimal two-person to orbit RLV, you are probably talking about stages not much bigger than SpaceShipTwo–ie something small enough it could be built in a place like Mojave.
  • By sizing the stage to fit under existing carrier aircraft (like WK2 or “WK3″ if it ever gets built), you can possibly get away with just leasing the carrier airplane during development, flight testing, and early operations–you don’t have to design your own carrier aircraft or run its development program. You also don’t have to buy the plane outright and staff its operations. Keeping development and early operations costs as low as possible is going to be important.
  • Depending on the design of the carrier plane to spaceship interface, it can double as a gantry crane, thus allowing for ground-level maintenance and processing, without needing separate crane systems. This reduces the amount of infrastructure needed to support a launch.


  • Because the first stage has to be able to glide forward to land at the originating spaceport, you are more constrained in where your carrier plane can go for an up-range launch point. This means that you don’t have as much flexibility with regards to orbital phasing for first- or second-orbit rendezvous.
  • For near-term carrier planes like WK2, the GTOW limits are probably going to force you to use a cryogenic fuel on the upper stage to make the design close(Methane or Propane at least, but the performance requirements may actually move you towards LH2) . Dealing with cryogenic fuels such as LH2 on the way up to the launch point is going to be tricky, and is going to take a lot of work to make properly. Even keeping the LOX cold on the way up to the launch point is going to take non-trivial insulation systems.
  • With a LOX/LH2 upper stage, the rocket might end up being a lot larger than SS2, which might make fitting it onto something like WK2 tricky–this would have to be looked into.
  • Your system now has two staging events, one from the carrier airplane and one between the two rocket stages.
  • You still need to develop orbital reentry-capable TPS, and depending on the staging velocity of the first stage, it may also require TPS.
  • If you’re trying to build to work with something like WK2, you may be down in the size where minimum gage issues start biting you. There are lots of systems that don’t scale down very well past a certain point, and especially on the upper stage portion, those can make life challenging. While staging does reduce the mass ratio requirements on the upper stage, the benefit may end up being largely offset by the fact that you’re shoving the upper stage further into the minimum gage range of the design space.
  • Also, even though going two-stage can reduce the performance burdens on both stages by quite a bit, they may still end up being really aggressive mass-wise, particularly on the lower end of the scale.
  • Since there are two stages to deal with, the launch licensing process is going to be more challenging, especially since there are a lot more trajectories that would need to be analyzed for E-sub-c, and more potential failure points along those trajectories. Also, since you are more constrained on your launch point location relative to the originating spaceport, you can’t get as far out of civilization before launching (and you have to keep “civilization” out from under the instantaneous impact point as much as possible).

There are probably other benefits and drawbacks, but unlike Part I, I didn’t have the brains of much smarter people to pick while writing this. So this will have to suffice for now.

Thoughts on Various Instantiations and Trades
As I’ve played around with various concepts within this specific approach, I’ve been mostly focused on first-generation systems–i.e. the stuff that could be built in the near to foreseeable future. As you may have noticed, for instance, I’ve been fairly heavily focused on the concept of using WK2 as a carrier plane, especially now that I have some better numbers on its carrying capacity. While it would be possible to design a larger custom-built aircraft that would therefore free-up more mass and make the design easier to close, that would open up a whole can of worms in other areas. It would add another vehicle that needs to be designed and qualified. It would have to go through FAA’s certification process. And it wouldn’t be used as much as a multi-purpose aircraft like WK2 that at least right now has a very large anchor tenant that plans to buy and operate several of them. Which means that the operational burden of owning the carrier airplane may be required right from the start, thus driving up the cost to get this idea off the ground. So for now, I’ve been assuming that WK2 would be the carrier plane, and that therefore the GTOW has to be kept down to 35klb. While it may be possible over time to increase the GTOW capability of WK2, by up-engining the design, I doubt you’d be able to stretch that capacity to more than 50klb. And for now, such upgrades are far off in the future.

Also, I’ve been assuming a payload of approximately 1000lb, which should be enough for a pilot and a passenger, or some mix of cargo. While 1000lb for two people may not sound like a lot, remember that the upper stage already has its own propulsion, GN&C, TPS, RCS, communications, batteries, and other systems that take up most of the weight of a capsule-type system. Mostly for the two-man configuration, you’d be adding in a pressurized cabin, two seats, some controls, some short/medium duration life support, the people themselves, and probably some sort of Intra Vehicular Activity suit (and possibly a hatch of some sort), oh and two parachutes for emergency bailout at lower altitudes/speeds. When two people aren’t needed, this cockpit might be swappable for a small cargo canister, or for some extra propellants for going to a depot. But any way you slice it, the goal would be about 1000lb.

One other common factor in all the approaches I’ve looked at is going with a cryogenic fueled upper stage and a LOX/HC first stage. While there is a lot to be said for avoiding LH2 for an air launched system, its performance advantage definitely makes it look tasty. Plus, with LOX/LH2, you have the RL-10, which is a reliable, high-performance, and readily available engine that is in just about the right thrust class for the job. While a larger carrier plane or in-air refueling might allow for an all-LOX/HC design, both of those add complexities that I would like to avoid if possible. You might be able to make an all LOX/HC vehicle in the size range we’re talking about work if you had a really high performance engine, but you’d still be running into minimum size issues with a lot of components. While it is harder to achieve a given propulsion system mass ratio with LOX/LH2, you have more mass to work with, and a fully-functional orbital stage requires a lot of other subsystems that don’t scale down linearly. For instance, on the Centaur stage, the tanks and engines only weigh about 40% of the stage mass. The rest of it goes into several other subsystems (batteries, electronics, RCS, structures, etc) that while scalable are not as directly tied to propellant density as tanks and engines are. Using the estimates I got from some LM friends, the actual LOX/LH2 tanks and engines would take up less than 1200lb for most of these concepts, leaving quite a bit of mass for those other subsystems. An all LOX/HC approach *might* work, but you’d only save about 500lb on the propulsion structure, and you would have about half as much dry mass to play with for the rest of the subsystems (and that’s assuming 350s Isp on the upper stage, which is pretty impressive as far as LOX/HC engines go). It isn’t clear to me that LOX/LH2 for the upper stage isn’t actually the easiest way to go.

Now, if Scaled is able to increase the performance of WK2 any (or if it turns out that the carrying weight constraints can be relaxed if necessary), that may change things, but for now, with the limited GLOW available, I think LOX/LH2 for the upper stage is probably easier (if you can solve the thermal issues).

Based on those constraints (35klb gross, 1klb payload, LOX/LH2 single RL-10 upper stage, LOX/Hydrocarbon FS), I’ve looked at a few approaches.

The most straightforward would be a serially staged system, where the first stage fires, burns to near depletion (leaving a little propellant for landing), the two stages separate, then the upper stage lights its engine and continues on to orbit. For this system, you want a first stage with a decent amount of thrust. Previous studies on air launched SSTOs talked about a T/W of ~1.4 being desirable. That would equate to about 50klb. That’s bigger than the C&Space Chase-10 engine, or the Air Launch QuickReach upper stage engine, but smaller than the Merlin-1C. Depending on how aggressive you want to go with the first stage design (and whether you want a pilot on that stage or not), you have a range of options. On the conservative, piloted side, I ran some BOTE calculations that suggested you could possibly have a first stage that would have about the same dry mass (just a bit less due to being a 1-seater, and being a little more aggressive on the materials and engine T/W ratio) and propellant loading as a XCOR Lynx Mk1.

This would result in an upper stage that had about 42% the propellant loading of a Centaur upper stage, while having about 78% of the dry mass, not counting the payload. This gives a propellant mass fraction of just a hair over 80%, which while aggressive for a reusable stage isn’t off in fantasy land. The burnout velocity for the first stage would actually be under Mach 3, which means you could actually launch right over your launch site and glide back, or operate anywhere within about a 50miles diameter. By having your launch point be so close to your orginating spaceport, it might make it easier to get launch licenses for the various trajectories you would want to fly.

OTOH, by going with a more aggressive first stage (say LOX/Methane, no pilot, more aggressive materials, higher tankage fraction, etc) you can reduce the required upper stage delta-V by quite a bit, which means that with the same 3500lb of dry mass to work with, you might be able to get the upper stage propellant load down to about 31% of a Centaur loading (about 14klb), which is a more manageable 75% pmf. This would have a much higher burnout velocity (above Mach 5), which would imply needing to launch several hundred miles uprange. Which way you go depends on the tradeoff between easier development and operations of the first stage, and the challenge of getting the upper stage to design to close.

[Just for kicks and grins, and because I know that most alt.spacers will decry the mere suggestion of using hydrogen, here's what you would need to get a LOX/HC two stage combo to work. You'd need high Isp engines: 340s Isp for the first stage and 350s for the upper stage. And you would need dry masses of about 2000lb for the first stage and 1900lb for the second stage--not counting the 1000lb payload--and propellant masses of about 15000lb each. In other words, the first stage would be very similar to the first stage shown above, but the upper stage would also have to be very aggressive. That's about an 87% pmf. It's not impossible for vehicles this size, but I think it may end up being a lot more challenging than the LOX/LH2 design.]

The other main branch of the tradespace I’ve looked at involves using just a single main propulsion engine (the RL-10) modified with TAN injection and using propellant crossfeed from the first stage. You could either use the cross feed just the TAN propellants, or the TAN propellants and the core oxidizer. Both have benefits and drawbacks. Neither case ends up getting you results drastically different from the serial staging route, but you only have one main propulsion system to deal with. If you cross-feed the RL-10′s core LOX during the burn, you can actually reduce the required size of the upper stage a bit, but at the expense of shifting the stage mixture ratio more towards hydrogen, which will cost you density-wise. The benefits compared to a serially staged system are fairly modest compared to the added complexity, so it might not be worth the hassle.

One more thought on instantiations. There’s nothing per se that requires an air-launched TSTO system to be a winged HTHL configuration. In fact, it might be possible to do it with two VTVL stages. One air launch paper I saw on L2 of NASASpaceflight.com talked about a massive air launcher for a Delta-IV type launch vehicle. They found that if you light the rocket engine with the airplane still attached, that you can get enough thrust to pitch the aircraft up to a high-enough angle before staging that you eliminate the need for having wings on your stage. It costs you a little extra propellant for a few seconds, but if it ends up making it easier to achieve your mass ratio, because you don’t need wings, all the better. Your stages wouldn’t have as much cross range, so your launch locations would be a little more constrained, but this is still probably doable. Also, for VTVL stages you would either need landing engines, or some sort of altitude compensation or flow-separation control in order to allow you to use high-expansion-ratio engines for the main acceleration for landing. Just food for thought.

Related Technologies
Like the other RLV approaches, there are several related technologies, and areas where we need to gain more experience with before such a design can close. Here are what I see as the key related technologies, and a bit about how I would go about filling in the knowledge gaps:

  1. TPS–This shouldn’t come as a surprise to you. Robust, reusable, survivable, and lightweight TPS is pretty much the key to any orbital RLV. As mentioned in my previous articles, there are lots of ideas out there, but precious little data on all but a few of them. Using a platform like WK2 coupled with some sort of kick stage, it should be possible to test out various TPS approaches, gradually expanding the envelope. The process for systematically testing out orbital TPS in a low-cost, rapid-iteration manner is probably a topic with more discussion at a future date.
  2. Cryogenic Propellant Storage–Even if you don’t go with LH2 for the upper stage, you’re still probably stuck with at least one cryogenic propellant in order to get the performance you need. The long trip from the launch site to the uprange launch point is going to expose your vehicle to a lot of convective heat transfer on the outside. Designing proper passive and active cooling systems that are lightweight enough to carry to orbit (or which can be located on the carrier plane with the minimal amount of interconnects) is a tricky problem. Fortunately, in many ways its related to other cryogenic fluid management challenges that need to be solved for on-orbit applications. Low conductivity tank-to-frame connections, internal insulation, and possibly various active cooling techniques will all need development. Fortunately, you don’t really need even an honest-to-goodness rocket stage to test these technologies. It might be possible to make a testbed that’s just a big propellant tank that you could fly using WK2 to gather data before you jump into the full development program. By being able to retire the risk and mature the design with a simplified non-flightweight prototype, you can get answers to your key questions a lot quicker. Even with all these precautions, some sort of onboard tankage and propellant conditioning equipment will probably be needed, but the optimal solution is likely going to be a mix of passive vehicle-side stuff combined with active mothership-side hardware.
  3. Airframe–At least for the upper stage (and also for the lower stage in most of these configurations), it is going to be quite a challenge packing that much hardware into that small of a mass budget. This isn’t as much of a problem for VTVL vehicles, which tend to have much simpler structures that use the tanks themselves for much of their strength. Fortunately projects like XCOR’s Lynx MkI and MkII will help provide some experience on lightweight, high-propellant mass fraction winged vehicles. Sure, they’re only suborbital, but that’s the point. It’s a lot easier to take a reliable, working system, and then start finding ways to make it lighter, than to try to design a super lightweight system from scratch and then make it reliable. A suborbital winged vehicle will require most of the subsystems that a winged orbital stage will require, but it can serve as an incremental step along the way–showing where the high-mass components are, and allowing you to start figuring out what it will take to close the design. More importantly, it may turn out that the data from such an excercise may show that an orbital vehicle this small just ain’t going to happen. That’s useful information too, even if discouraging. Much better to learn if something looks feasible from a smaller, profitable project, than to toss a bunch of money into something you don’t know will work.
  4. Orbital Prox-ops Tugs–For any small RLV, you’re usually fighting pretty hard agains minimum gauge issues and such. Anything that can allow you to remove hardware from your system, and transfer it to something that can be left on orbit, is a huge win. It might be possible to allow the prox-ops tug to carry almost all of the mass associated with rendezvous, docking, and even people transfer. Imagine a tug with multiple arms and a “transfer tunnel”. The tug goes out, rendezvous with RLV, grapples it with one or two handholds, and brings it back to the station. Say its transfer tunnels have extendable portions (kind of like the loading ramps for commercial jetliners), that has an electromagnetic ring with a small ferrofluid reservoir. You have a ferromagnetic matching ring built into your vehicle around the crew ingress/egress hatch. The transfer tunnel seals against that ring providing a vacuum tight seal, using the robot arms to articulate things correctly. The tug then hauls you up to the station, grabs a handhold or two located near an EVA hatch, and repeats the process. The people could then leave their vehicle wearing their IVA suits. They’d enter through the airlock itself. That way if there’s any problem with the sealing, it doesn’t endanger the station, and the vehicle is designed to operate even with a loss of atmosphere…Just sayin.

    Simpler tugs, possibly without the transfer hatch could also be of great use for other types of missions (propellant transfers, delivery of external spares, delivery of small satellite kick stages, etc). One of the keys to propellant depot operations is going to be minimizing the amount of hardware necessary on the tanker vehicle. If each tanker vehicle has to be built as its own Autonomous Rendezvous and Docking robot, with full RCS suite, full GN&C suite, etc, there’s no way you’re going to get costs down or propellant efficiency (ie propellant mass as a percentage of the orbited payload mass) to a reasonable level. What you really want to do is have the propellant tanker literally just be two big tanks with a little plumbing and some grapples. The stage that launches it provides attitude control long enough for the tug to come and snatch the tank delivery and haul it up to the depot. Then, you want the plumbing on the tanker side to be the minimal stuff to store and handle the fluid between launch and delivery to the depot. You want as much of the transfer related hardware as possible on the tug or station side. Ideally, I’d have the plumbing on the tanker side just be a panel with some grapple points, a couple of manual lever actuated ball valves, and some quick disconnect receptacle fittings, and have all of the smarts and all of the robotic bits on the part that already has to have articulating arms and end effectors. Make the whole thing so simple and cheap that the State Department will give you ITAR permission to sell it to anyone in the world to integrate into their own dumb tanker module.

    But the key is that by not requiring all of the specialized hardware normally associated with rendezvous, docking, and fuel transfer to be hauled around on the vehicle for every flight, it makes it a lot more likely that you can close the design case. The only drawback is that you really need someone to develop a good tug at that point, and unless they can find an existing market to justify their tug’s existence before you start your RLV project, you might end up in the unenviable position of needing to develop both the RLV and the tug at the same time–thus greatly increasing your odds of not being able to pull it off.

  5. Thrust Augmented Nozzles–In situations like this where you need good Isp, but you also need good propulsion T/W ratio (to make the mass budget close), having a technology like TAN that helps you on both scores could be very useful. Also, TAN as mentioned earlier, can enable you to do a project using only one main propulsion engine, provided you can cross-feed propellants reliably. Right now, TAN has been proven out on a small, laboratory scale, but there’s still work to do to bring it to primetime.
  6. Composite Cryogenic Tanks–At least for the LOX, the ability to store some of the propellants in the wings allows you to save some weight for the vehicle (since now you’re combining two structures into one), and also probably helps with controling your vehicle’s CG. Cryo composite tanks are also likely going to have integral insulation, and thus be much better at reducing boiloff. I don’t think XCOR’s “Non-burnite” is rated down to LH2 temperatures, but even if only your LOX tank can use the technology, it’s still a big win. And for LOX/CH4 or LOX/subcooled-Propane designs, you can use it for both tanks.

The Path Forward
Like all of the other proposed RLV approaches, the path to orbit lies in suborbital vehicles. Regardless of stupid arguments about differing energy levels required and all that crap, the reality is that suborbital RLVs end up having to develop almost all of the subsystems you need for orbital launch vehicles. And you have to integrate them into a working, reliable package. And you need to do so in a way that can have quick turn-around operations. Sure, orbital operations are going to require lighter airframes that have more propellant in them. And more powerful and efficient engines, but many of the lessons still apply. The interesting thing is that on the scale we’re talking about for this specific approach, both stages end up being not much bigger than the XA-1.0/1.5 concept we’re working on at Masten or the Lynx concepts working on at XCOR, and in fact the first stage for such a system, as I’ve shown earlier, may very well be on the same performance level as either of our or their vehicles. So, a lot of that work is just continuing down the path that we and several other companise are already going.

The other big path forward would be using the WK2 platform as an engineering testbed. Both for testing out cryogenic storage and Airborne Conditioning Equipment for keeping those tanks toped-off till launch, and also for launching small stages carrying TPS experiments. Though honestly, many of the suborbital companies in development are also looking at converting their vehicles into nanosat launchers, which would be just fine for launching TPS test hardware. For airlaunched VTVL style vehicles, work could also be done to demonstrate both the ground handling aspects (tipping the vehicles over so they can be mounted to the mothership), and the launch aspects (lighting the VTVL stage, doing the pitchup maneuver, and separation). Lastly, they could always work on ways to try and increase the cargo capacity of WK2 a bit, to make our lives easier. Maybe using the airlaunched first stage as a JATO bottle?

One last big challenging area of development is going to be working with the AST to get such a system licensed in a manner that allows it to operate as flexibly as it needs to. One suggestion I heard was that if this technology were being developed for the government first, that it could build-up reliability experience to the point that the AST would be more willing to work with it. The other option is that the data built up from WK2/SS2, Lynx, and other suborbital vehicles can also be used to make the regulatory case easier. Right now, to be conservative, the E-sub-c calculations are done assuming that your vehicle will crash and die every single time you fly it. But once you’ve flown something several times, you’ve demonstrated a certain reliability level, that can be factored into E-sub-c calulations, thus allowing you to operate over slightly more populated areas. If you’re launching uprange of your landing site, and gliding forward to landing, odds are your Instantaneous Impact Point at separation is going to be fairly close to your launch site. If you design your trajectory so that staging occurs with an IIP slightly past the originating spaceport, then most of the challenge on the launch licensing is going to involve your first stage operations, as most orbital spaceports will probably be sited such that there isn’t much population nearby to the northeast, east, or southeast. For the first stage operations, since it really isn’t much more than a suborbital rocket, getting lots of safe flight experience with such a combination, over unpopulated areas should be fairly easy to do. Once that’s worked out, it should be much easier to make the E-sub-c numbers look reasonable enough that the AST will allow you a loose enough launch license that you can move your uprange launch point around in the way you’d want to to really take advantage of air-launch.

As I mentioned at the start of this post, this is my last post for now in this series. My main goal was trying to show people that there are multiple realistic ways of solving the reusable earth-to-orbit transportation. I also wanted to introduce people to some of the challenges and tradeoffs inherent in developing such systems, and show try to paint some sort of a vision of the path forward from here. Lastly, I wanted to lay a sort of framework for discussing the business aspects of orbital access. I hope you all enjoyed this series, and I think everyone for their comments. Now its time to roll up our sleeves and get back to work.

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

27 Responses to Orbital Access Methodologies Part VI: Air-Launched Glideforward TSTO

  1. Karl Hallowell says:

    Well, this isn’t a comment about this particular method. Having read through this series, I guess I had to say that I remain skeptical of methods that sacrifice significant delta v to make stages easier to recover to the launch point. I mean here that the default option, reusable stages with little to no effort to make them easier to recover, seems very competitive.

    As I see it, the only part of Earth-based transport that is expensive is the early recovery. Once you get the stage picked up, it’s cheap to ship through the vast Earth-based transportation infrastructure.

    Fast turnaround might be a nice feature, but these stages don’t strike me as being expensive assets in themselves. So just building a bunch more and allowing for slower turnaround seems a viable option here.

  2. Jon Goff says:

    Thanks for the comment. I agree that not having to sacrifice a lot of propellant for RTLS maneuvers is nice. However, I think you’re underestimating the cost of down-range recovery of launch vehicles. One thing is that if it lands on land, there are very real size limits to what can be readily shipped home via terrestrial transportation networks. Both rail and roads are limited to certain diameters beyond which you just can’t ship the thing. There’s a reason for instance, why Roton is stuck in Mojave. It isn’t because the Mojave airport was the highest bidder for keeping it. It’s because it was physically impossible to relocate very far due to its size. While some smaller TSTO RLVs probably could be “roadable”, you’re rapidly going to hit the point where there’s no way to get it back to your launch site….

    But I’m just repeating my point from Part II. There are very real costs to downrange recovery. It requires a lot more people, slows your flight rate down, etc making it much harder to close the economic case.

    But RTLS techniques that don’t require big performance penalties are definitely nice.


  3. Colin Doughan says:


    when weighing the options for Air-Launched Glideforward TSTO, would you see more promise in:

    1. Building two identical stages for simplicity, cost control, shared spare parts, etc. – with the only difference being the payload, or…

    2. Design each stage for its unique roll in the TSTO process -recognizing this, in effect, means each stage is its own unique vehicle design with unique design challenges?

  4. Jon Goff says:

    I’m not a huge fan of the whole Bimese concept. It might work, but you’d probably get better results from having the two stages more optimized for their own job. In the end, making both stages identical would probably cost more than it saves you in development costs.

    Just my opinion of course.


  5. john hare says:

    Just a thought on landing a VTVL stage with large expansion ratios. Some upper stage engines have nozzle extensions that are extended only after staging to keep the whole vehicle compact. These extensions could be copied with a thought to retracting them in between the end of main burn and landing to get proper expansion ration down low?

    Could the extra ratio area be expended before landing? Or puncture the skirt to get the same effect? This is more development that might be useful for avoiding HL requirements.

  6. Jon Goff says:

    There are a few ways to reduce the expansion ratio before landing. Part of why I was so busy last week was that I was writing a grant proposal for investigating my personally favorite way of doing that…

    But one option if none of those work well would be to have both a space rated main engine as well as a few small low-expansion-ration landing engines.


  7. Charles Pooley says:

    An alternative long considered for
    future Microlaunchers is RTLS by tow
    back by an aircraft which takes the
    empty (ground launched) first stage
    back to the launch site.

    A demonstration of this, using model
    planes, is under way.

    This is a fly-back to launch site

  8. Rodney says:

    You don’t have a phasing problem if you’re destination/station is in an equatorial orbit. This will be the most practical destination since it will be passing over the launch postion 14 or 15 times a day.

  9. Jon Goff says:

    The problem is that there’s almost no good places to launch over the equator. Almost all of the land sites tend to be in underdeveloped, politically unstable third-world countries. Launching over water comes with its fair share of issues. I’m not saying you couldn’t see equatorial sites being used. But I am saying that my opinion is that most RLV launches are going to take place from landlocked sites at mid-latitudes. Just because they’re closer to infrastructure.

    But with a good destination in a good resonant orbit at the right inclination, you can often get a good 1-2 “first orbit rendezvous” opportunities per day per launch site. If demand gets high enough that that that flight rate can’t accommodate, you add two or three more locations in the same inclination and altitude, but with an orbit phasing so they pass overhead at a different time of day.

    Most of the industrial world is not on the equator.


  10. Anonymous says:

    In the first scenario, the first stage only reaches mach 2.8.

    Is there any reason why this couldn’t be provided by a jet engine of some sort?

    If the vehicle was allowed to dive to a significant speed I imagine a very simple turbojet with only a few compressor stages and a properly designed intake, or even a ramjet, might be a relatively light-weight alternative to using a rocket for the first stage.

  11. john hare says:

    The problem with jet engines on the first stage is weight and complexity.

    The 50k rocket engine would mass perhaps 500 pounds with the 50k turbojet at about 5,000 pounds. The jet also requires fairly heavy variable geometry intakes if it goes supersonic. The intakes, afterburners, nacelles, and variable geometry nozzles can mass as much as the engine itself. You end up with an installed engine that masses almost the whole mass budget for the first stage even before you add wings, fuel and control systems.

    Supersonic flight with air breathing engines is an entire complex field that would have to be tamed for the speciallized stage. Air breathing engines are extremely sensitive to small variations in angle of attack and intake pressure recovery. Several very expensive planes are said to have been lost to intake unstart at high speeds. By the time someone is ready to build a stage for this job, they will have a reliable rocket engine.

    To see my arguements on the other side of this question, dig up my old posts on usenet. I have personal selfish reasons for wanting airbreathing systems to work.

    The problem with ramjets is they only work in a limited speed range. They barely work at mach 1 and only become effective at mach 2. By that time you are almost ready to stage. At no time do they approach the simplicity or thrust/weight of rocket engines.

    At the end of the day, almost all jet recommendations for spaceflight assume that rocket engines are too expensive and unreliable, even if not explicitly stated or recognized. By the time serious orbital operations start, neither of these assumptions will be true.

  12. Dave Salt says:

    Hi Jon,

    The air-launched glide forward method would be a natural for systems that used atmospheric LOx collection, like Andrews Space’s “Alchemist” Air Collection and Enrichment System (ACES), since the up-range cruise phase would provide an ideal period in which to “farm” the necessary LOx.

    Certainly, the collection and storage of LOx after take-off would lead to significant reductions in take-off mass and reduce the associated contingency risks. However, such an approach obviously locks you into using LH2 though, as you point out, this is not necessarily a bad thing.

    Moreover, if the ACES machinery and storage tanks could be installed in an existing airframe (e.g. a 747), the cost of developing this “high altitude launch pad” may make for a more credible business case… all other things being equal :-)


  13. Mike Puckett says:


    Could the GTOW of the WK2 be worked around by configuring it or the LV for some sort of mid-air refueling?

    Either the LV could be launched dry and refueled or the WK2 could be launched partially fueled like a B-52.

    Or perhaps a combination of both if the WK2 and the first stage would both burn Kerosene. I suppose it would not be a big deal for the WK2 to burn RP-1 instead of JP.

    Let the tanker top off the WK-2 and fill the first stage.

    Proabally be like feeding steak to a hound dog when he would have been just as happy with hamburger ;)

  14. john hare says:

    Would it be possible to simply buy tanker flights from the military? If the recieving gear is adaptable to WK2 the tankers exist. Might fit in with the govt development to retire regulatory risk.

  15. Jon Goff says:

    In the first scenario, the first stage only reaches mach 2.8.

    Is there any reason why this couldn’t be provided by a jet engine of some sort?

    If the vehicle was allowed to dive to a significant speed I imagine a very simple turbojet with only a few compressor stages and a properly designed intake, or even a ramjet, might be a relatively light-weight alternative to using a rocket for the first stage.

    While you might be able to do something like that, there are a few problems:

    1-Designing supersonic jet powered aircraft is challenging and expensive–rockets are easier quite frankly.
    2-jets don’t do well accelerating to high speeds, and more importantly don’t do well past certain altitudes. You’d end up having to stage too low, and your upper stage would now be a lot more difficult.
    3-jets have crappy T/W ratio compared to rockets, which means that your first stage would likely be a lot more complicated and expensive.

    Once you factor everything in, it’s not clear that there’s any benefits from having jet engines on that stage. Use them on the carrier plane to get up to altitude, sure. But for acceleration, and getting up to a good staging point for the upper stage? Just use a rocket.


  16. Jon Goff says:

    should’ve read your comments first before writing that last reply. Thanks!


  17. Jon Goff says:

    I should’ve mentioned something about alchemist. I had in fact been thinking about that. The big question is what is causing the constraint on lifting mass for WK2? Is it engine thrust at liftoff or is it power at altitude, or what? Especially if it’s thrust at liftoff, something like Alchemist (possibly combined with a little liftoff push by the first stage engines acting as a JATO) could make a big difference. That said, Alchemist is going to require a lot of work yet.

    Also, while a 747 would give you a lot more wiggle-room weight-wise (and thus would allow for bigger RLVs)…I’m a fan of starting smaller. Plus, I really like the WK-family approach to design that allows for easy ground integration, and easy separation mechanics. Piggyback definitely has its drawbacks when you can get aircraft with large undercarrying areas.

    Here’s to hoping that WK2 is enough of a success that there’s eventually a WK3…


  18. Jon Goff says:

    Mike, John,
    I’m sure you could add refueling capabilities (and as Dave pointed out reoxidizering capabilities) to help boost payload…but I’d really prefer if such things were just that “boosting the payload” above a certain entry level capability. The more stuff you have to have work right just to get anything to orbit, the less reliable the system is going to be, and the more expensive.

    If you can make a bare-bones system though that can get 1000lb to orbit using just the WK2, then as you add in-air LOX collection, in-air kerosene refueling, JATO assist on WK2, etc you could eventually upgrade the system till its much higher capacity.


  19. john hare says:

    You sir, are a bad influence.
    Causing me to rethink directions.

    Could the WK1 be used to work out issues? Noticed over at Clarks’ a mention of it in use for other customers purposes. It might be more available and affordable than WK2 for quite some time. It would seem that an XA1.5 could be the RLV first for a disposable nanosat second.

  20. Jon Goff says:

    You sir, are a bad influence.
    Causing me to rethink directions.

    Heh. I guess that means I’m doing my job.

    Could the WK1 be used to work out issues? Noticed over at Clarks’ a mention of it in use for other customers purposes. It might be more available and affordable than WK2 for quite some time.

    Actually, that’s not a bad idea. Of course, it’s not like I have money to actually pursue any of this myself alas…but using WK1 to try doing some stuff wouldn’t be a bad starting place. At least it’d give you a nice platform to try out some of the Airborne Conditioning Equipment stuff, and the cryogenic insulation.

    It would seem that an XA1.5 could be the RLV first for a disposable nanosat second.

    While we typically don’t fix the details for a given vehicle till we’re actually building it, our preliminary models are showing we could probably do nanosat launches off of XA-1.5 without needing air launch. Though we could also test it as an air launched first stage as well. Interesting idea.


  21. john hare says:

    Using the WK1 with XA-1.5 might help on logistics, insurance, and regulations more than physically placing the sat. Launching from Mojave would be much cheaper for you than having to support another launch site with water to the east.

    The XA-1.5 would be a pop up stage with no downrange velocity to speak of from Mojave. If you one stage from there to orbit, it would be very technically challenging from that small of a working mass. If you use two stages from pop up, you either drop one stage on land or do polar launches over the pacific. Your IIP for the upper stage(s) will be over land for an uncomfortable amount of time for the regulatory and insurance purposes.

    Staging from WK1 would seem to allow you to work from your home base on eastward launches. A single upper stage would have a rapidly moving IIP over land. Somewhat less percieved danger to the endangered Mojave Jackrabbit that lives downrange. Air launch in various flavors is far more about these issues than performance.

    At some point I could be crossing the line here. Talking about business issues that could affect your company is certainly borderline. I don’t mind being wrong. I do mind doing anything detrimental to the industry. If/when that happens, I would rather be warned than banned.

  22. Jonathan Goff says:

    You make good points. I hadn’t really thought about using air launch for nominally VTVL stages until I started writing this post (thanks, btw for giving me the idea of glideforward air launch in the first place). It definitely has some merit to it, but we’re still a *long* way from the point where we could use that idea. It’s one we’ll keep in mind though.


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

  24. Another option for “glide forward:”

    This would be a pad-launched (vertically-launched) vehicle, and therefore could be larger than the air launched one:

    Make the booster stage a “sea plane”

    Not really a sea plane or amphibian, but really a straightforward cylindrical booster with hydrofoil-type “skis” on it. Instead of parachutes, you’d use a deployed paraglider (or wings), and the skis/skids would allow for a horizontal water landing. The booster would settle into the water on its belly once there’s no more forward momentum. Since it’s a large hermetically sealed tank, it should float just fine.

    The engines I think, would have to a bit off center (so they are as far above the waterline as practical) and the vehicle should have a “boat-tail” under the engines as well. These modifications would help protect the engines form the elements.

    While it could then be towed as is, it would probably be better to either recover it onto a floating drydock, or place and secure it onto a pontoon barge. The barge could be submerged just below the surface and maneuvered under the booster before resurfacing with the booster on top of it. The barge can then be towed to land.

  25. Pingback: Selenian Boondocks » Blog Archive » Random Thoughts/Orbital Access Methodologies VII: Air-Launched Glideforward TSTO with Exo-atmospheric Suborbital Refueling

  26. Pingback: Selenian Boondocks » Blog Archive » The Business of Altius Space Machines

  27. Alex says:

    I’d love to see an addendum/update covering Stratolaunch – http://www.stratolaunch.com
    Also what about the AN-225 as carrier aircraft (this was considered for HOTOL)? External payload 200,000 kg according to wikipedia – http://en.wikipedia.org/wiki/An225

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>