Toward Distant Suns
by T. A. Heppenheimer
Copyright 1979, 2007 by T. A. Heppenheimer, reproduced with permission
Chapter 5: The Sweet Combination
There is a very simple reason why the space shuttle cannot build powersats: It carries only thirty-two tons of cargo. A powersat thus would need three thousand shuttle launches just to haul its basic structure to orbit, to say nothing of the extra weight for its assembly facilities and for the propellant to take it to geosynchronous orbit. At sixty launches per year, the most the shuttle's launch facilities are planned to support, each powersat would take fifty years to build. Powersat enthusiasts thus can be forgiven if they seek something which will do the job more quickly.
Through the years many proposals and not a few projects have sought to achieve such advanced rockets. As early as 1898, the Russian pioneer Tsiolkovsky was writing that rockets for space flight should burn liquid fuel. In 1903 he pointed out the advantages of using hydrogen and oxygen as rocket fuels. In the days when wood-and-fabric biplanes were the very model of modern aircraft, not only Tsiolkovsky but also his American and German counterparts, Robert Goddard and Hermann Oberth, had predicted the performance of many of today's well-proven combinations of rocket fuels.
In the fall of 1945 rocket research began at NASA's Lewis Research Center, near Cleveland, now a leading center for research in advanced rockets. Overnight, the NACA (National Advisory Committee for Aeronautics, the forerunner of NASA) management switched the laboratory emphasis from piston engines to jets and reorganized the staff from top to bottom. A small group was even assigned to rocket research. The reorganization caught the lower-level supervisors and researchers by surprise. One scientist went home deeply engaged in writing a report on spark-plug fouling. The next morning he found his desk in another building, and he was officially engaged in research on rocket motor cooling.
The political climate in Washington was such that the directors of NACA did not want to proclaim that they were supporting research in anything as far-out as rockets. In Pasadena, California, this attitude dictated that a Caltech center for rocket research be given the name it holds to this day: the Jet Propulsion Laboratory. At NACA-Lewis the rocket group was officially called the High Pressure Combustion Section. Not till 1949 could the group emerge from cover and change its name to the Rocket Research Branch.
Even in 1945 rocket research was not new. The German V-2 project had already seen its wartime misuses. Its alcohol and liquid oxygen propellants, the standard rocket fuels of the 1940s, also fueled the Navy's Viking research rocket and the X-1 and X- I A rocket airplanes. The Jet Propulsion Laboratory was proving out such combinations as aniline and nitric acid; other laboratories were deeply involved in work on kerosene as a rocket fuel and in improving the performance of solid propellants. The Lewis people actually were rather late to the field, and to make a contribution they had to work in less plowed areas. They concentrated on liquid propellants of high energy.
Their propellant work was straightforward. They first computed the theoretical performance of candidate combinations, then selected the most promising for experiments. At times their evaluation led into more detailed studies of propellant characteristics, including the starting of rocket motors, their cooling, and problems of combustion. In 1948 Vearl Huff and his associates, who did this theoretical work, made a major contribution. They developed a rapid mathematical procedure for the calculations, which had previously been quite laborious.
Paul Ordin headed the early experimental work. He and his group investigated hydrazine, diborane, and ammonia, to be burned with liquid oxygen, hydrogen peroxide, or chlorine trifluoride. The rocket men were not averse to taking risks. One man was given a sample of hydrazine in another city and had to transport it back to Cleveland. The stability of hydrazine was in question, and it could not readily be shipped. He solved the problem by putting the sample in his pocket and bringing it home on the train.
In those days some important rocket propellants could be gotten only at one place, the Buffalo Electro-Chemical Company, and safety in transport was occasionally low on the list of priorities. When the first Viking rocket was in preparation for launch, in March 1949, the New Mexico launch crew found themselves running low on peroxide. Two of the Viking designers, in Baltimore and itching to be at the launch, solved the problem by driving a company station wagon to Buffalo. They put a fifty-gallon drum of the explosive fluid in the back seat, then drove at full speed despite snowstorms and other delays to reach New Mexico in four days—without benefit of interstate highways. Something like this happened at Lewis, too. Their first diborane came from Buffalo also and was delivered by an engineer in his own car. It came nestled in dry ice on the rear seat, complete with a safety device which soon saw use: a whisk broom. When some diborane leaked past a valve and spontaneously ignited, the engineer neatly whisked the flame away.
In May 1948 the Lewis people held a conference on their latest fuels research. The rocket research had used diborane with both liquid oxygen and hydrogen peroxide and had discovered a problem. Boron tended to form gluey combustion products, which produced deposits on parts of the motors. So it was that the rocket group became intrigued with the ultimate oxidizer, the most powerful available: liquid fluorine.
Other investigators had tested the use of gaseous fluorine, but Paul Ordin wanted to use it in liquid form and he succeeded. Initially he used diborane as a fuel. Huff's calculations had predicted high performance and no deposits of boron fluorides. However, the 9,200°F combustion temperature melted test engines in less than a second. What's more, they saw no reasonable way to cool a diborane-fluorine engine, for diborane is not a good coolant and fluorine is too reactive.
Their first experiments with fluorine, however, only whetted their interest. They worked with fluorine throughout the fifties, using it neat with ammonia-hydrazine mixtures as well as with ammonia and hydrazine. Their biggest effort with fluorine came after 1955 and involved burning it with liquid hydrogen. They also started a program to burn hydrogen with oxygen. The first studies of hydrogen-oxygen rockets dated to Tsiolkovsky in 1903 and Goddard in 1909. By the end of the forties, hydrogen-fueled rockets had been tested at Ohio State University, JPL, and Aerojet Engineering Corp. Aerojet's work was the most advanced, featuring a hydrogen liquefaction plant and three-thousand-pound–thrust motor tested in 1949. The Lewis rocket group wanted to work with engines up to twenty-thousand-pound thrust and with longer burning times.
To carry this program forward, NACA provided Lewis with a new rocket-testing facility. It was capable of not only operating twenty-thousand-pound–thrust hydrogen-fluorine rocket motors, but also of removing the deadly hydrogen fluoride from the exhaust and muffling or silencing the engine noise. The facility began operating in 1956. Over fifty thousand gallons per minute of water completely muffled the noise and absorbed the hydrogen fluoride. The water flowed into a tank and was treated chemically, producing an inert white powder (calcium fluoride), which could be hauled away. As one of the key people in this effort wrote in later years, "We were ahead of the environmentalists."
As this work advanced, it received an extra boost from one of Lewis' directors, Abe Silverstein. Silverstein was very enthusiastic about the potential of hydrogen not only for rockets, but also for aircraft. With his strong backing, the rocket group built and tested lightweight hydrogen-fluorine and hydrogen-oxygen motors, cooled with their own hydrogen fuel, with thrusts of five thousand and twenty thousand pounds.
By 1959 they had achieved full success with a hydrogen-fluorine motor under simulated conditions of space operation; actual performance reached almost 100 percent of theoretical. They measured a key performance parameter, exhaust velocity, at over 15,400 feet per second, the highest ever attained by a chemical rocket up to that time.
But the difficulties of working with liquid fluorine had not been lost on Silverstein. Work on hydrogen-oxygen had proceeded in parallel with the work on fluorine, and when Silverstein saw a hydrogen-oxygen motor run, the sweetness of that fuel combination came through to him loud and clear. Everything worked smoothly and simply, and performance was high.
In 1958 NACA gave way to NASA, and Silverstein was called to Washington to be director of Space Flight Development. His first task involved the Saturn project. Saturn had started in August 1958, when the Advanced Research Projects Agency (ARPA) gave Wernher von Braun the task of developing a booster with 1.5 million pounds of thrust using available engines. With the advent of NASA, President Eisenhower transferred the Saturn project to the new agency. The early emphasis had been on the first stage of Saturn, and Silverstein was named to head a group that would prepare recommendations for its upper stages.
Late in 1959 the committee recommended the hydrogen-oxygen combination for the upper stages of Saturn. The upshot was Saturn's use of the first commercial hydrogen-oxygen motor: the RL-10, built by Pratt & Whitney of Hartford, Connecticut. With fifteen thousand pounds of thrust, the RL-10 was the first hydrogen-fueled rocket to be built for actual use in space. It saw only limited use with the Saturn program, but it continues in use to this day with the high-energy stage known as Centaur. Centaur launched the Pioneer and Voyager spcecraft to Jupiter and Saturn, as well as Mariner 9 and the 1976 Viking missions to Mars.
The advent of the RL-10 meant that leadership in developing advanced rockets was passing into the hands of industry. Pratt & Whitney had grabbed the first plum, but it was not long before there were other key projects. Most of those went to an established industry leader: Rocketdyne, a division of North American Aviation.
North American got started in this area in 1946 with an Air Force contract to develop what today would be called a cruise missile. With the name Navaho, the project called for a ramjet-powered pilotless aircraft to fly 5,000 miles at three times the speed of sound at an altitude of 100,000 feet. Since ramjets develop thrust only when they are at high speed, the Navaho required a rocket booster to carry it to the speed and altitude where the ramjets could take over.
Even today, developing such a craft would be no mean feat; in 1946, the needed technology simply did not exist. The need was for new guidance systems, new electronics; for ramjets, for supersonic aerodynamics, and for rocket engines. To push forward in these areas, North American set up their Aerophysics Laboratory, with a rocket test facility in the Santa Susana mountains north of Los Angeles. In 1950 they began testing Navaho rocket motors there. With a company reorganization in 1955, Rocketdyne, headquartered in Canoga Park, was established as a separate division.
The nation's missile programs in the 1950s brought much new work on rocket propulsion. Rocketdyne won the early contract for the 78,000-pound thrust motor for the Redstone, America's first major missile, which burned alcohol and liquid oxygen. By the mid-1950s, Rocketdyne had contracted to develop a 150,000-pound thrust engine, burning kerosene and liquid oxygen. This engine, or slight modifications of it, was selected for four of the nation's most important rocket programs: Atlas, Thor, Jupiter, and Saturn.
As early as 1956 NACA-Lewis and Rocketdyne had started work on what would later become the F-1: a single rocket motor burning kerosene and oxygen, developing 1.5 million pounds of thrust. The early work concentrated on a forty-inch-wide injector for the propellants of this immense motor. In January 1959 the F-1 was established as a formal project to be built by Rocketdyne. In 1959 Rocketdyne engineers at Santa Susana were already testing an early version, which featured the forty-inch injector and a combustion chamber, though the chamber was built of heavy metal and was not cooled. The early test runs went only one and a half seconds. By June 1961 a complete F-1 was fired successfully. That was less than three weeks after President Kennedy committed the nation to achieve a lunar landing by 1970. When Kennedy announced that decision, he already knew that the F-1, key to the lunar rocket, would soon pass a critical test.
The F-1 project began fully three years before the design was chosen for the rocket that would use it: the Saturn V first stage. The NASA leaders knew that an early start on new rocket motors was essential; their decisions were to develop powerful new engines, without necessarily having a commitment to specific boosters that would use them. In this way, they laid much of the groundwork for the Apollo program even before Kennedy took office, in the days when Eisenhower was pooh-poohing suggestions that the nation expand its space initiatives.
The F-1 was not the only new motor begun in this way. In the summer of 1960 Rocketdyne won another contract: to develop a 230,000-pound thrust engine burning hydrogen and oxygen, to be known as the J-2. As with the F-1, the J-2 was begun well in advance of the space vehicles that would use it. The first successful ground test of a J-2 came in late January 1962. At about the same time preliminary designs were developed for the second and third stages of the Saturn V. The second stage, designated S-II, used five J-2 motors, much as the first stage used five of the more powerful F-1's. The third stage, the S-IV B, used a single J-2.
This S-IV B functioned well in its first flight late in February 1966. A second flight, in July, disclosed no major problems. The first flight of the complete Saturn V took place on November 9, 1967, and again all was successful. But on the second test flight of Saturn V, in April 1968, a problem arose which threatened to bring the progress of the Apollo program to a dead halt.
The first stage functioned successfully, and the second-stage cluster of five J-2 's started properly and at first operated normally. After 70 seconds, the engine compartment in the area of the No. 2 engine began to chill. This chilling was followed by a slight reduction in engine performance; its combustion chamber pressure fell off. Then, 193 seconds later, the No. 2 engine lost all pressure and shut down. This shutdown caused the adjacent No. 3 engine to shut down also. The onboard computer adjusted for the loss of these two engines by computing a modified flight path and longer burn time for the surviving three engines, and the spacecraft reached orbit successfully.
The single J-2 in the S-IV B third stage started normally and operated for 68 seconds. Then its engine compartment also began to chill, followed after another 40 seconds by a slight falloff in engine chamber pressure. The engine functioned in this fashion for the full 170 seconds of its planned burn. But when the engine was commanded to restart, it could not be made to do so.
This meant that despite its earlier successes, the J-2 could not be relied on for manned flight. Yet it was critically important that the problem be found and fast. Amid the euphoria of a string of successes in 1966, NASA planners had set the first lunar landing for February 1968. The tragic Apollo fire in January 1967 cost the lives of three astronauts and left NASA deeply shaken; many key managers were transferred or fired outright. That fire also cost the program a year and a half of delay, as the Apollo command module had to be redesigned. Now, the new 1-2 problem threatened further delay.
Yet such delay was quite intolerable. The Apollo program had from its start held the goal: Beat the Russians to the Moon. By the spring of 1968 the achievement of this goal was in doubt. The Soviets were preparing their Zond spacecraft to carry a cosmonaut on a looping flight around the Moon. Such a flight would have been far short of a lunar landing or even of orbiting the Moon. Indeed, when the U.S. undertook such a single-loop circumnavigation, in the Apollo 13 mission, it was only to save the astronauts' lives; the flight was written off as a failure. But a successful Soviet flight would have given them the highly prized achievement of the first man to the Moon. Later Apollo achievements—first to orbit the Moon, first to land on it—then would have been merely qualifications.
The threat to the J-2 was real. Extensive ground testing had failed to turn up situations that would duplicate the failures which had occurred in space. There was no way to recover the failed engines orbiting the Earth, and only the most meager data was sent back from onboard instruments.
Rocketdyne's J-2 project manager, Paul Castenholz, had the responsibility for solving the problem. Normally an intense, hard-driving man, Castenholz drove himself to new heights of intensity as Wernher von Braun called for a round-the-clock, seven-day-a-week effort.
The eventual solution turned out to be elusive but simple. Liquid hydrogen flowing through an auxiliary fuel line had set up a vibration, causing the line to rupture. In turn the supercold liquid was released causing the temperature drops in the engine casings. The failure had been missed in earlier ground tests because they had been conducted in normal atmosphere. The frigid liquid hydrogen had actually caused air to solidify around the fuel line and to form an icy sludge that protected it from vibration. In airless space, there was no such protection.
To track through the few clues to the solution took thirty days. Thanks to the ingenuity and hard work of Castenholz and his group, Apollo went forward on schedule. In the wake of this success, NASA officials announced that the first manned Apollo flight using the Saturn V would orbit the Moon. This was the famous Apollo 8 Christmas flight, whose astronauts read from Genesis and televised views of the Earth and Moon to the world. As for the Soviets, their unmanned test craft Zond 5 successfully circumnavigated the Moon, but entered Earth's atmosphere along a wrong trajectory and experienced re-entry forces sufficient to kill a man. The Soviets then abandoned their effort.
Even before Neil Armstrong's "one small step" in July 1969, NASA planners had begun to look beyond the Apollo program to the project that was to become the space shuttle. Here again, advanced rocket development was pointing the way. As early as 1965 the Air Force has given a contract to Pratt & Whitney to develop a new experimental motor, the XLR-129. Burning hydrogen and oxygen and developing 250,000 pounds thrust, the XLR-129 incorporated important advances over the J-2. It was to operate at the high combustion chamber pressure of 3,000 psi, to produce high thrust and improved performance from a lightweight, compact engine. Also, it was to be reusable.
Preliminary studies of space-shuttle designs began in earnest in 1968. Many of the early designs called for direct use of the XLR-129 or of closely similar engines. By 1970 NASA requirements had jelled, and called for a two-stage shuttle, each stage of which was to burn hydrogen-oxygen. Both stages were to be designed as winged rocket airplanes, flying back from their missions for frequent reuse; and both stages were to use what was by then designated the Space Shuttle Main Engine, or SSME. The SSME was to develop 550,000 pounds thrust, with twelve engines in the first stage, two. in the second stage or orbiter.
Detailed studies of such shuttle designs went forward at North American, General Dynamics, and McDonnell Douglas during 1970-71. At the same time other detailed studies were under way on the SSME at Rocketdyne, Aerojet, and Pratt & Whitney. Most observers gave the edge to Pratt & Whitney, whose XLR-129, already successfully tested, gave them a decided leg up. At Rocketdyne, they had their hands full with existing contracts. It fell to Paul Castenholz, fresh from his J-2 triumph, to grasp the nettle.
He persuaded his corporate management to grant him $3 million of company funds to undertake the building of an experimental motor. His motor would go the XLR-129 one better since it would develop 400,000 pounds thrust, to the 250,000 of the competition. With his usual hard-driving approach, he set his group to work in December 1970. Two months later, he had his motor.
It featured a copper combustion chamber cooled with hydrogen in the manner of a space-rated engine and incorporated an injector design similar to the one that would finally fly. It also used preburners, as would the final design, in an important demonstration of SSME technology. In February 1971 the motor was taken to the Rocketdyne facility at Reno, Nevada, rigged to a test stand, and fired for one-half second—enough to measure performance, yet not so much as to risk burning a hole in the side.
This test completed, Castenholz had the motor taken on a tour of NASA centers. At Lewis, Marshall, and NASA Headquarters in Washington, NASA people could see for themselves that an SSME was just around the corner. That effort of Castenholz put Rocketdyne back in the race for the SSME contract; in July 1971 Rocketdyne won the contract.
In the summer and fall of 1971, as the restrictions of future tight budgets became evident, NASA completely changed the shuttle design. In a matter of months, it evolved from being a two-stage fully reusable design to the current design, which uses solid boosters and carries its propellants in an expendable external tank. These changes influenced the SSME. When the shuttle was redesigned, NASA proposed that it use not two but four engines in the orbiter. But the shuttle was to be built for frequent reuse, and it was expected to operate in the fashion of an airliner. Rocketdyne sought the counsel of TWA, whose consultants pointed out that they had experienced great savings in bringing in such three-engine airliners as the Boeing 727, as compared with four-engine craft like the 707. With fewer engines, aircraft required less maintenance. As a result, Castenholz recommended that the shuttle have not four but three engines, and the shuttle has remained so to this day.
The decision to build a three-engine shuttle fixed the SSME thrust at 470,000 pounds. The first complete engine was tested in May 1974; again the test was for one-half second. Then, over most of the next four years, the SSME program struggled through one major problem after another.
The problems involved the propellant turbopumps, which were driven by preburners. In the preburners hydrogen was burned with a minimal flow of oxygen, producing a very fuel-rich exhaust that was cool enough not to destroy the turbopumps. These gases drove the turbines, then fed through the injector into the combustion chamber, where they were burned with the rest of the oxygen, producing temperatures of 6,000°F. Hydrogen, flowing in many small tubes close to the main lining of the engine, cooled the combustion chamber and nozzle before being pumped to the preburners.
The fuel turbopump was designed as a compact package, about four feet long by eighteen inches in diameter, the size of a large outboard motor. It was to develop 76,000 horsepower. This was more than had run such huge ocean liners as the Mauretania early in the century in an era when ship engine rooms covered an acre and more of space below decks. Few comparisons can better illustrate the astounding advances in engine design that have occurred in this century. Of course, marine engines have been built to last for decades of service; the SSME was designed for 7.5 hours of total operation. Yet even this was a huge advance over the few tens of minutes of operation required of earlier rockets, which were only to be flown once. [Author's footnote: In this, there is considerable hope for the future of low-cost space flight. One of the early jet engines, the J-33 used in the F-80 fighter, in 1946, had an average operating lifetime of fourteen to twenty-five hours. Modern jet engines have lives up to 10,000 hours. Since engines are expensive, long life helps greatly in keeping costs down. Similar improvements in rocket engine lifetimes may be in store in decades to come.]
The first problem with the turbopump was that the turbine shaft was not held solidly enough by its bearings; it tended to jiggle with a circular motion. At 37,000 rpm, the jiggling quickly wore out the bearings. The solution was to stiffen them but it took eight months to figure out how to do it properly.
Next, there was the problem of cooling and lubricating the bearings. No oil could be used; thus the bearings were cooled with liquid hydrogen. Since this method was rather like lubricating an auto engine with water, it took some ingenuity. It turned out that one set of bearings was not getting enough hydrogen, was overheating, and then failing. The solution: redesign the channels that fed the bearing its hydrogen. That took another six months.
A third major problem involved the turbine blades. There were sixty-three of them in one section of the turbine, each the size of a postage stamp; and each of them was generating six hundred horsepower, the power of an Indianapolis racing car. Naturally, the blades were under some strain, and they tended to crack. The problem was traced to vibration, a well-known scourge of engineers, and finally corrected. The time: another six months.
While work was progressing (slowly) on the fuel turbopump, there was concurrent development of the oxygen turbopump. Here too there were serious problems but the problems were more difficult to find and correct. When turbopump problems made their presence felt, the turbopump in question would fail and shut down. A fuel turbopump failure simply caused the engine to lose power. It could then be disassembled, and the cause of the failure sought. But failures in the oxygen turbopump brought severe damage to the engine. In earlier rocket development projects, oxygen pump failures often caused the motor to blow up. The SSME was built of sterner stuff; to contain its high pressure, it was built so stoutly that it would not (usually) explode. But it could, and did, catch fire. It was built of metals like copper, nickel, and steel, which we do not regard as fire hazards; but at the temperatures and pressures of an SSME, in the presence of liquid oxygen, virtually anything will burn. These fires often burned so much of the engine that it was difficult to discover what parts had failed or in what sequence the failure had spread.
The oxygen turbopump suffered two major problems. The easy one took six months to find and fix. It involved a rotating seal, which served to separate liquid oxygen from the hot gases in the turbine. The seal was designed to rotate without friction, but it tended to rub against another engine part. This rubbing then produced heat by friction—heat sufficient to ignite the metal, as when Boy Scouts start a fire by rubbing sticks together. In the SSME such fires burned more than marshmallows. Eventually, by choosing a different type of seal, the problem was licked.
The most difficult problem, though, was that the oxygen turbopump bearings repeatedly failed and burned up. In the end, a variety of expedients was used. The rotating shaft of the turbine was redesigned to give better balance. Just as an unbalanced auto tire wears rapidly, the inadequately balanced turbine shaft, rotating at 31,000 rpm, had worn so quickly as to cause the unit to fail. The bearing supports were stiffened. Finally, the bearings and their races, or holders, were made bigger and built to carry heavier loads. After a year and a half, the turbopump bearings finally passed their tests.
The last pump bearing failure was in March 1977, and from that point the designers could proceed with engine tests at low power. They could not proceed to tests at the rated power level till they were quite sure that there were no problems at the lower levels, and the testing proceeded cautiously. By mid-August 1978, engine tests had accumulated a total of only 17,000 seconds of operating time, much of which had been at low power.
Late in that month, Rocketdyne officials met with NASA administrator Robert Frosch and stated that they were prepared for a dramatic step-up in their rate of testing. The goal to be reached by September 1979 was 65,000 seconds of test time, at which point the engine would be pronounced qualified for shuttle flight. The Rocketdyne people were as good as their word, and better. By Thanksgiving they had reached 32,000 seconds, and most of the new testing was at 90 percent or more of rated power. The accomplishment put them three-and-a-half months ahead of schedule. Thus, they set themselves a new goal: to reach the 65,000-second mark by early June. That would qualify the SSME for a first shuttle flight date of September 28, 1979. Very soon the testing was a month ahead of even the new schedule.
Then just after Christmas 1978, disaster struck again. An SSME motor, under test at the major NASA facilities at Bay St. Louis, Mississippi, blew up. This time it was not a problem with the turbopumps; the new problem areas were the main oxidizer valve and a heat exchanger. The valve problem, once realized, was quickly fixed by a redesign. The heat exchanger problem was less simple. In the words of a senior manager, "The failure of the heat exchanger remains unexplained, and it gives you a very soggy feeling. These incidents occurring so late in the test program just do not inspire confidence."
Further testing was held up for weeks, and Frosch slipped the first shuttle flight to November 9, 1979. As 1979 progressed, continued SSME testing proceeded, and no new problems were uncovered. Confidence in the engine again began to grow.
Why these delays and overruns? In the views of senior program officials, the reason is that the program had been too thinly supported from the start, its funding levels held too low to assure success on time. With more funding, there would have been better analysis and testing of designs, which would have uncovered flaws earlier, allowing them to be corrected more easily. As it was, the program was managed with a strong "success orientation": it lacked the sort of managerial defense-in-depth which could readily accept failures and errors, then work to correct them. Such technical surprises were only to be expected in view of the unprecendented requirements of the shuttle program, and in the equally demanding Apollo program, it was just this defense-in-depth which Castenholz had relied on in rescuing the J-2 engine. Because of this, Apollo had reached the Moon on time and within budget. But by skimping on funding early in the shuttle program, problems had been left to build up, until they could be fixed only at considerable expense and delay.
For all this, there was no doubt the shuttle soon would fly; and by mid-1979, the SSME had long since shown the true meaning of the phrase, "engineering development. " The long and weary years of turbopump shutdowns and engine fires were now a receding memory. Meanwhile, interest in power satellites was steadily increasing. In short, it was possible to look ahead to a new, major rocket engine development project.
In reality, rocket specialists had never entirely ceased their thinking about the future, and such projections had gained great encouragement in 1971. In that year Robert Salkeld, a consultant with Systems Development Corporation, announced his discovery of mixed-mode propulsion. His discovery raised eyebrows among rocket experts, for he had found a new wrinkle in the very familiar equations which govern the performance of rockets. First derived around the turn of the century, these equations are taught in freshman physics classes, and have been worked over so thoroughly that it was difficult to believe anyone could find anything new in them.
Salkeld had studied the performance of a rocket stage carrying two different fuels and two types of rocket motors, each burning its own fuel. He showed that a judicious choice of the fuels and of the sequence in which they were to burn would allow a rocket stage of given size to carry more payload than with any single choice of fuel, even if the fuel was of high energy like hydrogen.
For instance, a space shuttle might be designed as a single stage, burning only hydrogen and oxygen. But if part of the fuel tank was given over to a different fuel such as kerosene, then the shuttle could be made to carry much more cargo, or else the craft could be built smaller, more compactly, and still carry the same payload.
The obvious application of this idea was to a long-held dream of rocket designers, the single-stage-to-orbit rocket, or SSTO. Such a craft would take off (possibly from a runway, in the fashion of an airplane), fly to orbit in one piece and perform a mission, then return through the atmosphere to a landing. It would bring true airplane-like simplicity to space operations by avoiding the costs and complexities of booster stages. And it would bring a certain tidiness. After all, every airplane that flies today is a single-stage craft. [Author's footnote: This was not always so. In the late 1930s the task of commercial flight across the Atlantic was thought by some to be so difficult as to require a two-stage aircraft. This, the Short-Mayo Composite of 1938, was a four-engine flying boat, which took off with a smaller four-engine seaplane on its back. It worked. But by 1939, the larger Boeing 314 had introduced single-stage transatlantic service.
Salkeld went on to work with Rudi Beichel, an expert in the design of rocket motors, to propose SSTO craft capable of a most diverse range of missions: space rescue, cargo transport to orbit, even transport of people by rocket to any point of the world in two hours. An important key to these ideas was a novel rocket motor designed by Beichel, the dual-fuel motor. It would start off at liftoff burning kerosene or a similar fuel, but would later switch to hydrogen.
In 1975 NASA undertook a study, "Outlook for Space, 1980-2000." An attempt to forecast the nation's future in space, the study called on the nation's best rocket designers to come forth with their predictions as to what rockets of the future would be able to do. In particular, they called for suggested designs for a class of rockets capable of use in building power satellites. Such rockets would carry two hundred tons to orbit. The resulting designs included both single-stage and two-stage craft, water landings and aircraft-type landings by winged craft. The SSME figured prominently in many of the designs, but the designers were not reticent in calling for new engines.
In 1977 two Langley scientists (Beverly Z. Henry and Charles Eldred) reviewed their center's work on advanced rockets in a paper given at the Princeton Conference on Space Manufacturing Facilities and Space Colonies. They were concerned with the advantages to be gained from a novel rocket motor design proposed by Beichel: the dual-expander engine. This engine would make use of Robert Salkeld's principle of mixed-mode propulsion to offer performance gains of up to 25 percent over the SSME.
The dual-expander engine is actually two rocket motors in one. An inner combustion chamber bums oxygen and a fuel such as kerosene at the very high pressure of 6,000 pounds per square inch—twice that of the SSME. Around the inner combustion chamber is an outer one, which surrounds it to form a ring. It burns hydrogen and oxygen at the SSME's pressure of 3,000 pounds. The general arrangement thus resembles a bell mounted inside a larger bell. An ingenious flexible seal or bellows permits the combustion chambers to shift up or down with respect to each other, despite the hot, high-pressure gases. It serves to vary the size of the rocket's throat or narrow part of its nozzle, so as to give the best performance at any altitude. An important feature is that at any time it is possible to thrust with either chamber, or with both, thus providing unparalleled flexibility in operation. Equally important, this high performance is achieved with a much shorter and simpler rocket nozzle than had previously been thought possible. The short nozzle makes it easier to mount in a rocket and easier to steer from side to side when controlling the rocket's flight. Such a motor can then produce the thrust of an SSME, yet have less than half the weight. A dual-expander motor weighing only 4,350 pounds could produce 608,000 pounds thrust at liftoff, compared with a 6,250-pound SSME, which produces 375,000. In the vacuum of space, where thrust increases, the SSME will give 470,000 pounds thrust, but the dual-expander design will give 693,000.
As Henry and Eldred emphasized, the dual-expander engine is far from the only advance that may be expected. Their paper looked ahead to continued progress in the design of lightweight structures: wing and tail surfaces, propellant tanks, engine mounts. They emphasized that modest levels of funding for research could produce large payoffs in reducing the weight of these structures. Other advances could be expected in protecting against the heat of atmosphere re-entry.
The lowest-cost rocket they discussed was a winged single-stage craft, which would take off vertically like the space shuttle, carry a 250-ton payload to orbit, then re-enter and land on a runway. With conventional SSME-type engines, it would be so tail-heavy that it could not fly in the atmosphere. It would lack the stability needed for safe operation. The lightweight dual-expander engines would overcome this problem by reducing the tail-heaviness. The result would be a space transport craft offering a cost to orbit of $7.20 per pound.
Still, such a craft would have to operate from Cape Canaveral or a similar site; it could never fly from such airports as JFK or Dallas-Ft. Worth. Precisely this type of operation would be possible with an entirely different type of propulsion—the scramjet or supersonic combustion ramjet. A type of jet engine, it would permit an aircraft to fly at speeds and altitudes that today can be reached only by powerful rockets.
Can a jet engine fly at close to orbital speeds? A turbojet cannot fly at much more than three times the speed of sound, or as an aerodynamicist would say, Mach 3—no mean achievement. The reason for this performance is that incoming air must be slowed down and compressed in order to burn the fuel by means of a carefully shaped inlet channel that creates shock waves. Air passing through the shocks slows and compresses. In doing so, it also heats up, the heating growing intense at high speeds. Above Mach 3, the heat is sufficient to cause engine parts to soften or weaken. Nor can the engines be built to use cooler air flowing at higher speeds; the rotating fan or compressor of a jet then could not work properly.
Since the end of World War II propulsion specialists have been tantalized by the ramjet. The simplest type of jet in current use, it flies at high speed and rams or compresses air by its speed alone. It has none of the complex turbines and compressors of a conventional jet. It cannot take off from the ground, but requires a turbojet or rocket to get up to the several hundred miles per hour needed for it to work properly. Its simplicity allows it to fly higher and faster than a turbojet; so far from being limited to Mach 3, it reaches peak performance at that speed. It then can fly on to higher speeds called hypersonic. However, above Mach 6 its performance falls off drastically. By contrast, orbital velocity is Mach 24.
The reason is that the airflow in a ramjet still must be subsonic in order for the flame not to blow out. Again it is shock waves that serve to slow and compress the outside airflow to the desired internal speed. Though fierce heating is produced, a ramjet is simple enough that it can be cooled by flows of fuel. Above Mach 6, even this fails. The need for subsonic combustion sets the limits to performance of ramjets.
Since the late 1950s it has been appreciated that if a ramjet was capable of burning fuel in a supersonic airflow to provide thrust, very great gains in flight speed and performance would become possible. If the resulting scramjet would burn hydrogen, it would give better fuel economy at Mach 6 than a conventional turbojet at Mach 2. As early as 1965 scientists at the Marquardt Corporation operated an experimental scramjet at better than Mach 10. Since then, it has become a certainty that scramjets can be developed to permit routine flight well above Mach 6.
Then why are there no scramjets today? The reason is that they pose peculiar problems not encountered in developing jet engines and rockets. A rocket can be tested on the ground and its performance in space reliably predicted. The same is true when a turbojet is ground-tested or is fed with high-speed air from a wind tunnel. A scramjet can also be run in a wind tunnel, and indeed many have been. But the test results cannot reliably be used to understand how a scramjet would actually perform in hypersonic flight.
The reason is that a scramjet has a curious feature. Unlike a turbojet or rocket, it cannot be designed and built separately from its aircraft and then mounted. Instead, aircraft and scramjet must be designed together. The forward portion of the aircraft underbelly actually must behave somewhat as an inlet channel, which feeds air to the scramjet. The aft portion of the underbelly acts as a nozzle, allowing the scramjet exhaust to expand and produce more thrust.
Such a complex system cannot be tested as a scale model in a wind tunnel. Only a full-size craft will do, and only in hypersonic flight. In the mid-sixties NASA had a Hypersonic Research Engine project that aimed at building an experimental scramjet to fly aboard the X-15 rocket plane. However, the X-15 program was cancelled (in 1968) before the engine was ready to test. If scramjets are to be developed, there will be need for an entirely new Hypersonic Research Aircraft, which indeed is stongly advocated by the senior management of NASA-Langley. Like the earlier research aircraft such as the X-1, X-2, and X-15, this craft would continue the well-proven practice of pushing back the frontiers of flight by means of research aircraft flown by daring test pilots.
It took a quarter-century to advance from the first experimental flight at Mach 2 to the double-sonic Concorde commercial jet. It would probably take nearly as long to build and test this new research craft and to apply its lessons to a new generation of single-stage-to-orbit vehicles. Such scramjet craft would still need more time for development. So for the task of building powersats, scramjets will loom too far in the future to be of interest. Instead, the emphasis will remain fixed upon that tried and true engine, the rocket. It is this which will drive the powersat's space freighters.
By 1978 further studies of power satellites had defined more precisely what the post-shuttle space freighters might look like. Such projected advances as mixed-mode propulsion, which made SSTO attractive, were seen to make two-stage freighter craft even more attractive. The single-stage craft offered the lower operating costs, though not by much; and two-stage craft still looked easier and less risky to design. With proposed payloads now approaching five hundred tons, the emphasis had to be on avoiding risk.
Even so, Salkeld's work had had its influence, and it was clear to all that the first stage of such two-stage freighters would have to burn a fuel such as liquid methane, or perhaps liquid propane. Thus it would be necessary to go forward with developing a new rocket motor.
This fact implies that advocates of the power satellite will have a clear signpost to look for, an indication that the nation truly is preparing to go ahead with building powersats. The new engine will have to be contracted for well in advance of any other major powersat activity. Like the F-1 and J-2, it will likely start as a project and be well under way even before there is a commitment to go forward with the space freighter that will use it. Like the SSME and the earlier XLR-129, its development may be well advanced before firm designs are set for the freighter, or even for the powersat itself.
The sign to look for will be a NASA contract, probably to Rocketdyne, to proceed with a new and large rocket motor. Its thrust will be some two million pounds, perhaps more. It will burn oxygen and methane or propane and may be capable of dual-fuel operation. It will be designed for high chamber pressure and consequent compact size, and for operating lifetimes of many hours.
On the day that contract is announced, probably in the early 1980s, it will be evident to those who are knowledgeable that whatever the president may say, whatever that year's space budget may show, a commitment to build powersats will be coming along. It will come as certainly as the digging of a foundation presages the building of a house.