The huge costs associated with the long-serving "semi-reusable" Shuttle are prohibitive for commercial interests operating in space. So also are the huge costs associated with existing or planned expendable launchers. It does not look to me like the new Constellation design will be either timely or a cost improvement.
The idea of a reusable space plane of some sort was supposed to reduce these costs, but never actually did in the Shuttle's case. Too much of that vehicle is not reused, there is way too much rework between flights, and the complexity of the system demands a supporting logistical tail the size of a major American city. That can never be inexpensive.
The X-30 "Orient Express" project during Reagan's administration was supposed to fix this by using high-efficiency airbreathing propulsion in a single-stage spaceplane. It failed primarily because the necessary propulsion technology, supersonic-combustion ramjet (scramjet), did not exist in a technologically-ready form. Even today, it is still very experimental. A secondary reason is that the design trajectory was essentially re-entry flown in reverse at very high drag. This eats up a lot of the benefit of the higher-efficiency airbreathing propulsion.
The X-33 project of the 1990's was supposed to address launch costs in a reusable, chemical-fueled rocket, single stage vehicle. The cost reduction was to come from reusability. The only way to accomplish such a thing is to drastically cut the weight of the structure, and the payload weight. That's just the plain facts of physics with the rocket equation. Structures that light (under 8% structural fraction) are inherently very fragile, while the spaceflight environment is very punishing. As it turned out, every time they set an X-30 fuel tank on-end for a component test, its own weight crushed it. The entire vehicle had to be built the very same lightweight way. So the project died.
The old lifting-body work of the 1960's was a part of what went into the Shuttle. Another part was the idea of carrier-airplane launch. The original Shuttle design was a two-stage airplane, with a gigantic carrier craft reaching low hypersonic speeds in the lower stratosphere. There it was to release a rocket spaceplane orbiter. Obviously, with government budget-shortfalls and all that political-football nonsense, this plan did not happen. Yet, as the original ASAT missile launched from an F-15 fighter proved, such a thing is possible.
Most folks do not remember (or ever knew) that we actually tested nuclear-thermal rocket engines in the Nevada desert from about 1959 to 1972. These offered vastly-improved efficiency performance over chemical rockets, but with two important problems. One, the engine thrust/weight ratio was too heavy for use as a lower stage or for direct surface launch. The other was the severe radiation hazard of the exhaust stream, and also of the post-fired hardware. That stuff they tested long ago in Nevada is still out there "glowing blue" at night. That second radiation-hazard problem is probably the most insurmountable.
The nuclear rocket work, promising as it was, died when President Nixon killed all manned spaceflight beyond Earth orbit in 1972, right in the middle of the Apollo landings on the moon. It was supposed to be the prime propulsion for the planned Mars landings of the 1980's. When it lost the rest of Apollo and the Mars landing, NASA killed its nuclear rocket program as "no longer needed". In hindsight, I think everyone might agree with me that Nixon and NASA made very bad decisions.
All of this got me to wondering if there really is a way to have less expensive access to Earth orbit, and if so, which approaches might "work best". There seem to be three basic ideas to pursue: staged rocket, nuclear upper stage, and staged airplane. Any of these will need something new added to the mix, or we are right back where we started before. And yet, this "new" item had better be an existing technology. Otherwise, it will be stuck "forever" in development, just like the scramjet-powered X-30.
One of the most important things I learned in development engineering work was to question all assumptions. Another was that you cannot do something new if you do not change something, somewhere, somehow. After all, repeating the same actions endlessly, and expecting the outcome to change, is a very good definition of insanity. So, I looked closely at a few of the launchers we have operating today: Atlas-5, Titan-4 and Delta. Each is an update of an old 1950's missile design, with upper stages, or strap-on boosters, or both. Atlas and Titan were our first liquid-propellant ICBM's. Delta's core is the old Thor IRBM. All of these are mid-50's in origin.
You must remember what it was like in the 1950's to be designing rockets that might reach orbit. The only technologies ready-to-use were chemical rockets that could barely do the job, provided that structures were one-shot, throwaway light, and that multiple stages were used. (Ramjet assist could theoretically have helped, except that the ramjet hardware was dead weight until the vehicle reached a feasible ramjet speed.) Even with all the improvements made since then, that basic design approach is still constrained by those facts. All modern launchers are fundamentally based on that same idea.
Airbreathing technologies are inherently far more fuel efficient. But, in the 1950's, ramjet was barely considered ready-to-field with stage-off boosters (SA-4, Talos, and Bomarc), and 1950's turbine simply was not adequate for the job. Lots of things have changed since then: ramjets have been updated with integral boosters (SA-6, ALVRJ, ASALM-PTV, and SS-N-22 "Sunburn"). Turbines have been improved, with air bypass to the afterburner, to cover speeds up to about Mach 3.5 at relatively high thrust (the J-58's that pushed the SR-71 "Black Bird").
Yet none of that was ever fed back into any of the launcher designs. Lots of studies, sure, but no hardware programs. I think this was largely because of the same budget-shortfall and political-football nonsense that screwed up the original Shuttle design. That ought to be a criminal offense. Are you listening, Congress? You did it: you ruined NASA with that crap!
Looking at the basic fast-ascent stage rocket, the lightweight throwaway structures seem doomed to one-shot use, precisely because rocket efficiency does not allow the higher weight allowance necessary to provide protective systems and robust structures capable of absorbing the inevitable punishment of repeated flights into space. That begs a very interesting question: could adding some higher airbreathing efficiency afford us the weight allowance neccessary for effective recovery and re-use?
Another good question to ask comes to mind when examining the complexity and logistical tail behind the Shuttle, and to only a slightly lesser extent, the expendable launchers. If we could vastly-simplify these designs, could that reduce the size of the logistical tail to something substantially less expensive?
If those two questions have positive answers, and I think they do, could we put those answers together into one design and achieve a breakthrough in spaceflight launch costs? I think that might indeed be possible, but no one has ever tried it with hardware!
I organized my activities around 3 basic ideas: the staged fast ascent rocket, a nuclear upper stage used in such a way that its exhaust stream never hits the atmosphere, and a staged airplane.
For the fast-ascent staged-rocket idea, I added simple integral-booster ramjets (not scramjets!) to the lower rocket stages in parallel-burn format. I also added simplified ablative-protected pressure fed engines, letting the tankage provide a strong airframe as well as containing high pressures.
This turned out to be so promising that I reworked the same basic ramjet-assisted launcher into a single stage form for a lob-up trajectory for the nuclear upper stage. This is a non-optimal trajectory from an energy standpoint, but the nuclear upper stage has the "ooph" to handle it at a large payload fraction anyway.
For the staged airplane, I reworked the "air turbo-ramjet" design of the J-58's in the SR-71, into a 100% air bypass configuration for ramjet-only operation beyond Mach 3 all the way up to about Mach 5 or 6, at about 100,000 feet altitude. With this more-energetic carrier airplane technologically feasible, I settled on a 2-stage rocket "payload" for the hypersonic airplane to carry.
None of these studies are "complete" in the sense that they give "the" final answer, but they do provide some very interesting guidance.
For example, in the assisted stage rocket study, I rough-sized a 3-stage vehicle roughly 200 feet tall and a million pounds, capable of throwing 12,000 lb to a 1000-mile circular orbit, with structural weight fractions near 40% in all stages. The second stage had wings for flyback, with integral-booster ramjet nacelles on them.
During the first stage rocket burn, I fired the ramjet integral boosters at launch, then transitioned to ramburning around Mach 1. Staging 1-to-2 occurred at about Mach 2, with ramburning continuing in parallel with the second stage core rocket, to burnout at Mach 6, 80,000 feet. All of this was essentially vertical flight averaging 6.5 gees. The third stage was an all-rocket unit, since most of its burn occurred outside the sensible, usable atmosphere. With no moving engine parts to be damaged, I assumed ditching in the ocean near the launch site for second stage recovery.
In this study, I assumed simple pressure-fed kerosene-LOX engines with missile-type ablative liners in the first and second stages. The same engine using LH2-LOX was required of the third stage to keep up a high payload. I think that if I were to do it again, I would shut down the second-stage ramjets at about Mach 6, 100,000 feet, but continue to burn the core second stage rocket to about 10,000 feet/second, so that I could revert to a common kerosene rocket engine in all three stages.
The first and third stages used rear-mounted parachute deceleration into a front-first ocean splashdown for recovery. Unlike the Shuttle SRB's, these closed-tank units have inherent floatation, and being a pressure-fed system, these tanks are strong enough to be reliable for this kind of recovery. The third stage had a heat shield and featured N2O4-UDMH maneuvering engines in addition to the main engine. The storables were for circularizing and for de-orbit, and would be bladder-expulsion technology. The main engines in all the stages would be fed from simple free-surface tank technology.
In the second study, for the "lob-up nuke", I was able to hit the same Mach 6 / 80,000 feet point with a single-stage ramjet-assisted rocket, flying vertically upward at about 6.5 gees average, and carrying a single nuclear upper stage. This vehicle was around 1/3-million pounds at liftoff, and delivered the same 12,000 lb to 1000-mile circular, with 40% structural fractions in the lower stage, and 10% in the nuke. The lower stage was the same basic winged flyback design with ramjet nacelles on the wings as the second stage in the first study. As a single ramjet-assisted stage, this design was successful enough on paper that I would try that approach in the all-chemical stage-rocket mission, if I were to do it again.
From lower stage burnout, the lower stage uses its wings to bend the trajectory over for flyback, while the nuke just coasts straight up to an inertial apogee about 100 miles up. It fires horizontally for the full 26,000 feet per second orbital delta-vee at that point. This has the advantage that the radioactive exhaust stream never enters the Earth's atmosphere, a real safety item for using nukes to launch things. The nuke stage never returns to Earth, either, although it could be refueled and re-used on-orbit.
The staged-airplane study was different. It did not take long to realize that the airplane's "payload" should be a plain rocket, since most of its effort would be outside the sensible, usable atmosphere. I used pressure-fed LH2-LOX engines similar to those in the other studies, plus an N2O4-UDMH maneuvering system on the third stage for circularization and de-orbit. Both rocket stages featured the same 40% structural fractions and parachute-retarded ocean recovery as the first study. Recovery of the second stage would be very problematic, as it would splashdown essentially halfway around the world.
I had originally hoped to transition from air turboramjet power to rocket power in the carrier plane, so as to reach Mach 10 around 150,000 feet in a 45-degree sudden pull-up maneuver for payload release. There just was not enough room in a practical aircraft to do that. I looked at several supersonic bombers for guidance as to structural fractions. These included the RA-5C, the B-58, the XB-70, the B-1, and the SR-71. Excluding the Navy carrier bird, combat-capable mass fractions appeared to be 50-55%. A launch aircraft does not have to be combat-capable, so I used 40%. To hold a decent payload fraction, I had to "give up" at Mach 5, 100,000 feet, with the 100%-bypass air turboramets in ramjet mode from Mach 2.5 to Mach 5.
Even so, I could only insert 8000 lb into the same 1000-mile circular orbit as the other two studies. The airplane that "did this" looked like a gigantic, 4-engine SR-71, and grossed near 1.6 million pounds at liftoff. This was a horizontal takeoff design.
The same thing as a vertical takeoff design looked pointless to me, for that is right back to the stage rocket scenario. However, I did learn two important things from this study: (1) practical designs with reasonable-size aircraft will handle only small payloads (like the old ASAT), and (2) subsonic-combustion ramjet will still produce usable thrust at 100,000 feet M5-6 conditions, not just the 80,000 feet I had used in the other two studies.
Since I ran these studies, I have run across a Russian folding-wing flyback booster proposal that has appeared at airshows in mockup form, named "Baikal". That idea combined with my integral-booster ramjet nacelle could allow a path for easier land recovery of stages as separate smaller subassemblies, especially at long ranges. If I had it all to do over, I'd try that idea against the basic winged all-in-one stage that I looked at.
A key thing to glean is this: while basic subsonic-combustion ramjets were ready in the 50's, integral boosters were not, which turns out to be major reason why ramjet assist was not tried back then. It is the integral booster that makes sure the added weight of the ramjet engine is doing useful work, right from launch ignition. Without that integral booster, my results would never have looked this favorable.