Jump to ContentJump to Main Navigation
U.S. Space-Launch Vehicle TechnologyViking to Space Shuttle$

J.D Hunley

Print publication date: 2008

Print ISBN-13: 9780813031781

Published to Florida Scholarship Online: September 2011

DOI: 10.5744/florida/9780813031781.001.0001

Show Summary Details
Page of

PRINTED FROM FLORIDA SCHOLARSHIP ONLINE (www.florida.universitypressscholarship.com). (c) Copyright University Press of Florida, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in FLASO for personal use (for details see http://www.florida.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 21 June 2018

Viking and Vanguard, 1945–1959

Viking and Vanguard, 1945–1959

(p.10) 1 Viking and Vanguard, 1945–1959
U.S. Space-Launch Vehicle Technology

J. D. Hunley

University Press of Florida

Abstract and Keywords

This chapter discusses the Viking and Vanguard programs, separate American rocket efforts that contributed importantly to missile and launch vehicle technology, including the use of gimbals for steering. Although only a sounding rocket, the Navy's Viking made direct contributions to launch vehicle technology. It was also the starting point for America's second launch vehicle, Vanguard. Often regarded as a failure, Vanguard did launch more than one satellite. Together with Viking, it pioneered the use of gimbals for steering. In addition, its upper stages contributed significantly to the evolution of launch vehicle technology, since they were converted for use with the Thor-Delta series of space boosters. Additionally, a variant of its third stage was modified for use in the Scout program. This stage of Vanguard (in one of its two versions) pioneered the use of fiberglass cases and itself contributed to upper-stage technology for military missiles.

Keywords:   Viking program, Vanguard program, rocket technology, launch vehicle technology, gimbals, fibreglass cases, missile technology

Although only a sounding rocket, the Navy's Viking made direct contributions to launch vehicle technology. It was also the starting point for America's second launch vehicle, Vanguard. Often regarded as a failure, Vanguard did launch more than one satellite. Together with Viking, it pioneered use of gimbals for steering. In addition, its upper stages contributed significantly to the evolution of launch vehicle technology, since they were converted for use with the Thor-Delta series of space boosters. Additionally, a variant of its third stage was modified for use in the Scout program. This stage of Vanguard (in one of its two versions) pioneered the use of fiberglass cases and itself contributed to upper-stage technology for military missiles.1


The Viking rocket contributed to the Vanguard first stage as well as to the Vanguard guidance system. Milton W. Rosen, who was responsible for the development and firing of the Viking rockets, went on to become technical director of Project Vanguard and then director of launch vehicles and propulsion in the Office of Manned Space Flight Programs for NASA. His work developing Viking and his experiences with it prepared him for his responsibilities with Vanguard and beyond. Finally, Viking came early enough in the history of American rocketry that it illustrates a good deal about the evolution of the technologies used on launch vehicles. For all of these reasons, the history of Viking and Rosen's involvement with it deserve a place in this history. They constitute an early case study of technology transfer and the process of rocket development.2

After receiving his B.S. degree in electrical engineering at the University of Pennsylvania in 1937 and working briefly for Westinghouse and other firms, Rosen found employment at the Naval Research Laboratory (NRL) in (p.11) 1940. NRL was an interesting institution set up on July 2, 1923, “to conduct programs in the physical sciences and related fields directed toward new and improved materials, equipment, technology, and systems for the Navy.” It was a place where researchers could create areas of research “that were at once technologically important and scientifically interesting.” Rosen was later to do this after World War II, a conflict that convinced the Navy its “scientists needed to be concerned with predicting, even defining, and solving problems of the next war rather than the last one.”

Meanwhile, during World War II at NRL, Rosen worked under Ernst H. Krause on guidance systems for missiles. At the end of the war, the group under Krause began to plan its future, and Rosen, who had been reading G. Edward Pendray's proposal to use rockets for exploration of the upper atmosphere, suggested that Krause's group do just that. Krause supported the idea, and at the end of 1945 the Rocket-Sonde Research Branch came into existence.3

Rosen's background was in electronics, but if the group were to develop the sounding rocket he had proposed, it needed an expert in rocketry. Since the field was in diapers in the United States, it seemed unlikely that Krause would be able to recruit someone with that experience, so he asked Rosen to learn the field. Rosen knew that since he had proposed the idea of a rocket, Krause would not let him off the hook easily. He decided to agree but pose a condition he thought his boss could not meet: permitting him to spend a year at an organization with the most knowledge in the United States about rocketry, JPL. To his surprise, Krause agreed.

This conversation took place in November 1945. Rosen, Krause, and another colleague traveled to JPL in March 1946, and the lab agreed to add Rosen to its sta. Meanwhile, Rosen and his colleague C. H. Smith read “the JPL handbook, which was the primer of rocketry of that time.” Rosen and Smith began to plan their own rockets when they learned of the captured V-2s and the plans to fire them at White Sands Proving Ground in New Mexico, a place they had never heard of.4

Krause became the first chair of the V-2 Upper Atmosphere Research Panel and thus was involved in the research with the German rockets at White Sands. But for Rosen the immediate effect of this project was to relieve the pressure to produce an upper atmosphere research rocket until the V-2 firings were completed. On the other hand, Krause recognized that research with the V-2s would end sooner or later, so he told Rosen to continue developing NRL's own rocket. Rosen also became aware that once the V-2s reached the upper atmosphere, their stabilizing fins no longer served to keep (p.12) their noses pointed upward. Some somersaulted in the thin stratosphere and ionosphere, spoiling data scientists had hoped to gather from equipment like solar spectrographs.

Rosen and his colleagues wanted their rocket to be designed specifically for such research, unlike the V-2 missile. But Rosen's specifications for a rocket about a third of the V-2's weight with a thrust of 20,000 pounds (compared with 56,000 for the V-2) would show some influence of the German rocket as well as what little he knew of Goddard's work.5

The group under Krause actually came up with two basic designs, one by Rosen that became the Viking, and another by Smith that resembled the Aerobee sounding rocket. They sent specifications for both vehicles to five companies, and received bids from three—General Electric (GE), Douglas, and Martin. By the time the bids arrived, Krause, Smith, and Rosen had decided to go forward with the Viking concept, and the lab selected the Glenn L. Martin Company to build it. They did this because (1) the Baltimore firm was close to the Washington location of NRL, so the engineers under Rosen, who became head of the rocket research branch, could work closely with the prime contractor, and (2) they liked the confidence and optimism of the young Martin engineers, even though they could claim limited experience with rockets.6

Rosen also got the engine contract for Viking (originally called Neptune) assigned to Reaction Motors, Inc., again because of the firm's general proximity to Washington and because he and his associates liked the spirit of RMI's engineers. The contract, initiated in September 1946, was for a 20,000–pound rocket engine with the propellants fed by turbine-driven pumps. With the contracts in the works, in August 1946 Rosen left Smith as Viking “caretaker” and headed for JPL, arriving in September.7

Rosen spent about eight months at JPL. Still paid by the Naval Research Laboratory, he worked in Martin Summerfield's liquid rocket group, where Richard Canright taught him a bit about rocket testing and turned him loose on a project involving ceramic liners in rocket motors. Rosen selected it for his work because it would require him to design, assemble, and test rockets. Using nitric acid and aniline as propellants, he and his associates conducted several hundred static firings of 300–pound-thrust rockets, each test lasting three to four minutes, a long stretch for that period. The hypergolic propellants were reliable but dirty, requiring Rosen to disassemble and clean the motor after each test. Sometimes he encountered burn-throughs, and he had to analyze the causes in a paper he and Canright wrote. In the process, he learned a great deal about heat transfer in rocket motors.

(p.13) He also witnessed static firings of the WAC Corporal rocket to test the engine, propellant lines, tanks, valves, and control system. Rosen was sufficiently convinced of the importance of such tests that he decided to have every Viking rocket statically tested on the launchpad immediately before launching.

Besides his own work at JPL and what he witnessed, Rosen took two courses at Caltech. One—on theoretical physics from the Nobel Prize-winning Carl Anderson, who had discovered the positron in 1932 and later worked on Caltech's Eaton Canyon rocket project during the war8—consisted entirely of problem solving, which Rosen found useful. From Hans Liepmann, whom Rosen described as “another great teacher at Caltech,” he took a course in supersonic aerodynamics, a field that was new to Rosen and one he found highly applicable to his later rocket development.9

Rosen and others had also spent several hours with von Braun soon after the charismatic German engineer had come to the United States. Rosen did not remember much of the conversation, but he did recall that von Braun knew Rosen was developing a rocket and asked how it would be built. Rosen said they would construct it of aluminum. Von Braun believed the skin heating would be a problem, but Rosen was confident he and his team could make the aluminum work—as, in fact, they did. Early structural studies for the rocket led to selection of a simple cylindrical shape with a 25–degree cone at the nose. Wind-tunnel tests reduced the size of the tail fins, although they remained large. But mere practicality determined the initial diameter of the fuselage, a slim 32 inches. The widest available sheet of rolled aluminum was 100 inches, which yielded a cylinder 32 almost inches across. Von Braun also, according to Rosen, suggested the idea of a gimballed engine but pointed out (incorrectly) that Goddard had tried to use one and had not succeeded in making it work.

During Rosen's time at JPL, not a great deal of progress was made in developing Viking. It took a long time between letting the contract and cutting/bending metal. But both Krause and Smith were clamoring for his return, which occurred in April 1947 as soon as he had finished his project and courses.10

In some respects, Viking followed the design of the V-2, although with important variations. For example, as with the V-2, the American rocket's propellants were alcohol and liquid oxygen, pumped into the combustion chamber by turbines driven by decomposed hydrogen peroxide. But where the V-2 had used alcohol at 75 percent strength and hydrogen peroxide at 82 percent, the Viking used 95–percent ethyl alcohol and 90–percent (p.14) hydrogen peroxide. More significant, the really important elements of Viking—the nature of the guidance system, the gimballed motor, the post-cutoff stabilization—owed nothing to the V-2 but were original American designs.11

John Shesta drew up the basic RMI engine design, basing it on the firm's experience but also on recent information about the V-2. RMI's Edward A. Neu did the detailed design work on the combustion chamber and injector. Tests caused parts to fail and need replacement. Burn-throughs of the steel combustion-chamber liner (inner wall) led to the substitution of pure nickel, the first known use of this metal for such a purpose. Its superior thermal conductivity and higher melting point solved the heating problem in conjunction with the regenerative cooling in the original design. One injector caused an explosion, requiring new designs. Valves were a problem until M. E. “Bud” Parker borrowed valve designs from an MX-774B engine RMI was designing for the Air Force. Even after Rosen began launching Viking rockets from White Sands on May 3, 1949, there were component failures and redesigns. In fact, each of the twelve Viking rockets fired in the program (the last being on February 4, 1955) was different from the previous one as engineers learned from problems on one launch and applied the lessons to the next. In Rosen's view, “this was the most important aspect of the Viking program.” The result, among other things, was an engine that developed 20,450 pounds of thrust on the first flight and 21,400 on two others.12

Even so, the engine itself (apart from its steering mechanism) made no known contributions to launch vehicle technology other than the experience it afforded to Martin and NRL engineers. But Rosen claimed Viking was the first large rocket to use a gimballed engine for steering, a technology that did find widespread emulation in large missiles and launch vehicles. The validity of Rosen's claim hinges in some degree on the definitions of large and gimballed, because Karel J. “Charlie” Bossart, the “father of the Atlas,” has been credited with developing a precursor to gimballing in the MX-774 sounding rocket, which first flew on July 13, 1948, at White Sands, almost a year before the first Viking. This earlier technique was swiveling—rotating a rocket engine in a single axis for steering.13

The MX-774 project has been described in the preceding volume, Preludes to U.S. Space Launch-Vehicle Technology, in conjunction with Atlas. Here it will suffice to mention that the Army Air Forces (predecessor of the Air Force) awarded the Consolidated Vultee Aircraft Corporation (Convair) a study contract on April 2, 1946, for a missile designated MX-774. Bossart, the project manager, patterned it after the V-2 but with some radical innovations, only one of which was swiveling engines. A drastic funding cutback ordered by President Harry S. Truman in December 1946 curtailed (p.15) the project, and it was cancelled in July 1947. Meanwhile Bossart's team had built some test vehicles that were 31 feet 7 inches long, had a gross weight of 4, 160 pounds, and were powered by a four-cylinder RMI engine rated at 8, 000 pounds of thrust but actually operating over a range from 7, 600 to 8, 400 pounds. Each of the four cylinders could swing back and forth on an axis to provide control in pitch, yaw, and roll. After cancellation, the Air Force permitted Convair to fly three test vehicles. The flights, on July 13, September 27, and December 2, 1948, all had problems but did demonstrate the viability of Bossart's designs, including the swiveling of the engines.14

The swiveling engines on the MX-774 test vehicles clearly constituted an important innovation. Jet vanes like those on the V-2 and later on the Redstone were eroded by the missile's exhaust, attenuating control. Moreover, swiveling the rocket engines did not significantly decrease their thrust, whereas jet vanes produced drag that did reduce propulsion efficiency. But swiveling was merely a precursor of gimballing, not the same thing. A gimbal is a more complex device than a swivel, permitting a rocket engine (in this case) to rotate in two directions rather than on a single axis. Thus, in this respect, the Viking was more advanced than Bossart's vehicle, and gimballing became the technique of the future.15

In addition, the Viking was a larger rocket than the MX-774. Its dimensions varied, but the early Vikings were almost 43 feet long and had a gross weight of 14,912 pounds. Their thrust more than doubled the MX-774's.16 So Rosen's claim stands.

The initiative for gimballing came from Martin. The Viking specification required use of movable jet vanes to provide stability, but in 1946 the Martin engineers compared jet vanes with gimballing and recognized that jet vanes not only reduced control and thrust but also imposed a weight penalty of some 175 pounds. So in a report dated January 8, 1947, the engineers recommended gimballing. Rosen as project director concurred. The gimballing system that resulted was not as sensitive as the one later developed for Vanguard. It did not permit doing away with the fins on the Viking, and it required a separate roll control system using two aerodynamic tabs on the fins and two “steam jets” from the hydrogen peroxide turbopumps. The rest of the control system was not innovative, involving an adaptation of a commercially available Sperry A-2 autopilot. But a post-engine-cuto control system using residual gas from the pressurization system for the propellant tanks was indeed new, and provided a basis for the system used on Vanguard.17

Rosen recalled that, starting in about 1949, his team and Bossart's at Convair exchanged visits for several years until the Atlas project began in (p.16) earnest. On these visits they shared a great deal of information. The gimballed engines that were a significant feature on the Atlas missile and space-launch vehicle owed more to Viking than to MX-774, Rosen felt, pointing to the great difficulty his own team and the Martin engineers had in making gimballing work. It was not until the tenth or eleventh Viking launch that they got the bugs out of the system.

“The real contribution of Viking,” he stated, “was in making the gimballed motor work, which could not be done until the development of feedback theory at MIT during World War II.” When the gimbals got a signal from the control system, servomotors moved the engine to keep the rocket on the desired trajectory. A potentiometer then fed a signal back to the control system to stop the movement once the course was corrected. Adjusting the system so as not to overcompensate for a course deviation was a delicate process, which required negative feedback.18 Much of the credit for making the Viking system work efficiently and effectively belongs to Albert C. Hall, who wrote his Ph.D. thesis at MIT on negative feedback. As a consultant to the Martin Company, Hall set up the initial parameters for the system and proposed a method for adjusting it. This proved to be the most difficult problem the Viking team faced, but eventually they eliminated the instabilities in the control system. Others who used gimballed engines then adopted their methods, Rosen believed.19

There is no need here for a detailed description of Viking's design. Continual problems arose with each of the twelve Vikings launched during the project, and modifications to solve individual issues or to improve performance were numerous. The biggest change came between Vikings 7 and 8, when a redesign increased the capacity of the propellant tanks and lightened other parts of the rocket so that almost 80 percent of the gross weight consisted of propellant, which permitted reaching higher altitudes. Components were also rearranged to make them more accessible for repair. Whereas Viking 7 had been 47.6 feet tall, Viking 8 stood only 41.4 feet but was 45 inches in diameter rather than just 32 inches.20 In Viking 9, further improvements included modification of the post-engine-cuto control system to replace residual gas with hydrogen-peroxide rockets for better control.

Overall, the Vikings were successful. Launched from a ship at sea on May 11, 1950, Viking 4 rose to 105 miles, a record altitude at that time for a sea launch. Vikings 7, 9, and 10 (1951–54) all reached 135 or 136 miles, records for single-stage rockets until May 24, 1954, when Viking 11 rose to 158 miles above Earth's surface. More important, Viking paved the way for the use of gimbals to steer large rockets and missiles, and it helped prepare people like Rosen and the Martin propulsion engineer John Youngquist for Vanguard (p.17) and, in the case of Youngquist, for Titan. It also gave Martin valuable experience in serving as prime contractor for a large rocket, experience that would carry over to Vanguard, the Titan missiles, and the Titan III and IV launch vehicles.21

The Viking team did all of this with a handful of people. As Rosen recalled in 1971:

There were only twelve men on the Viking launch crew. The Martin Company, which built the rocket, had no more than about 50 two dozen engineers involved in the design. There were never more than about men involved in building the vehicle, and we had to borrow those from other projects because Viking didn't have enough production to hold them permanently. Finally, the government project group consisted of two men at first, and was never more than four. These two to four people wrote the specifications, negotiated the change orders, analyzed flight and test data, and wrote all of the final reports on Viking. That is why you could buy a Viking vehicle for $250,000 and conduct an entire Viking operation for $500,000.

Now, we cannot produce today's vehicles that way. Obviously, they are much more complex; they require more in the way of design and test. But we have been using thousands of people to build and launch these rockets. We have been generating tons of paper. And in many instances we have been using ten people to do the work of one. Can the ten people do a better job than the one? Yes, they can; they can do more calculations, they can try more alternatives, they can catch more errors—but they don't do the job ten times as well.22


The proposal that NRL submitted for the International Geophysical Year (IGY) satellite project—from which Vanguard was born—stemmed from an Air Force request in July 1954 to investigate use of a modified Viking to study the reentry nose cone issue that the Army later solved through ablation.23 This request led to the design of Viking versions M-10 and M-15, so designated for their capability to reenter the atmosphere at Mach numbers 10 and 15 respectively. The Air Force then pursued other avenues to study reentry. But the two designs formed the basis of NRL's proposal on July 5, 1955, of a vehicle to launch the IGY satellite.

The NRL proposal included two possible configurations. One would have the M-10 Viking as the first stage with two additional stages, both using solid (p.18) propellants. The first stage would use GE's Hermes A-3B engine,24 modified to increase its thrust from the original 22,600 pounds to some 28,000. The Atlantic Research Corporation had designed the two upper stages, which NRL believed were conservative enough to be developed in time for the IGY.

The other proposed configuration consisted of liquid-propellant first and second stages, collectively called the M-15, plus a third stage using solid propellants. This variant also used the M-10 as the first stage. Aerojet General's Aerobee-Hi sounding rocket, then still in development but scheduled to fly in August 1955, constituted the second stage. The third stage would again be an Atlantic Research solid-propellant rocket using ammonium perchlorate as oxidizer and polyvinyl chloride as fuel. When the Stewart Committee voted for Vanguard as the IGY choice, it favored the M-15 configuration even though it would take longer to develop. However, the committee recommended that NRL consider a different third stage, of the Sergeant type.25

In mid-August 1955 when the Army submitted to the Stewart Committee its revised proposal in support of Project Orbiter, the Navy put together a revised proposal of its own with a Thiokol third stage (Thiokol having built the motor for the Sergeant missile, the type of rocket the committee favored). Milton Rosen, representing the sea service, also gave the committee a revised estimate of when it could launch the first satellite. Rosen had originally stated that it would take thirty months, which proved astonishingly accurate. But Martin thought it could do the job in eighteen months if awarded the contract, and under the pressure of competition with the Army, Rosen endorsed this overly optimistic estimate against his better judgment. He bolstered the projection in his August memo with enclosures: a telegram from Thiokol promising delivery of a solid-propellant third stage nine months after a contract was let, a GE guarantee to deliver the first-stage engine in nine months, a promise by Aerojet to deliver Aerobee-Hi engines eleven months after it signed a contract, and an agreement from Martin “to put a satellite in being in approximately months” if government and industry clearly understood “the part each was to play in the program execution.” These assurances allowed the decision in favor of Vanguard to stand. In a memo to the secretaries of the Army, Navy, and Air Force, Deputy Secretary of Defense Reuben Robertson ratified the final decision of the Stewart Committee on September 9, 1955.26

As Project Vanguard developed, only Thiokol among these four contractors did not play a role. For the second stage of the launch vehicle, the Aerobee-Hi did not meet the requirements later specified, so Aerojet had to go (p.19) back to the drawing board for an engine that proved unexpectedly troublesome, although the firm had raised some red flags even for producing Aerobee-Hi in the eleven months to which it had agreed. Long after Vanguard was relegated to the history books, Rosen had occasion to comment on a pending NASA publication that said of NRL's launch vehicle, “The first stage was derived from the Viking sounding rocket, and the second stage was derived from the Aerobee sounding rocket.” Rosen, without indicating his own part in the process, wrote:

Th[is] myth, as I call it, was generated by the NRL people who were trying to sell the Vanguard project to the Stewart committee, in order to mitigate the impression that much of the project was new development, and indeed, at the time there was some hope that some of the Viking and Aerobee technology could be transferred. The truth of the matter is that Vanguard was almost entirely a new design; the engines were new, the structure was new, the guidance was new. All that was transferred was the valuable experience that the NRL, Martin, and Aerojet engineers derived from Viking and Aerobee designs and operations. The myth that the Vanguard was derived from the Viking and Aerobee sounding rockets continues to haunt those who perpetrated it.27

On the other hand, Rosen added in 2002, “I said then and I say now: without Viking there would have been no Vanguard.”28

In any event, it took from September 9 until October 6, 1955, for the Naval Research Laboratory to receive official direction from the chief of naval research to begin carrying out Project Vanguard. The objectives of the project were “(a) To put an object into orbit around the earth; (b) to prove that the object is in an orbit; and (c) to conduct at least one scientific experiment using the object.” And, according to a congressional committee report, the intent was to put this “satellite in orbit during the IGY,” which ended in December 1958.29

NRL appointed John Hagen as director of Project Vanguard. Born in Nova Scotia on July 1, 1908, Hagen had earned a B.A. at Boston University in 1929, an M.A. at Wesleyan in 1931, and a Ph.D. in astronomy at Georgetown in 1949. Between his master's and doctorate, Hagen had gone to work at NRL in 1935, devoting much of his research to improving radar techniques but also helping to develop an automatic indicator for the ground speed of aircraft. In 1954 he became head of NRL's Atmosphere and Astrophysics Division, and the next year he was named director of Vanguard.30 Rosen served as technical director under Hagen.

(p.20) NRL did not wait for official direction before getting to work on Vanguard. Knowing speed was essential, it let a contract through its parent organization, the Office of Naval Research, with the Martin Company on September 23, 1955, for the design, construction, and preflight testing of the Vanguard vehicle. On October, Martin in turn subcontracted development of the first-stage engine to GE. Once the weight, thrust, and other parameters for the first stage had been calculated, Martin could set targets for the next two stages. It determined that for the second stage to provide the velocity required, given the estimated weight of the vehicle's structure and instrumentation, it needed a thrust of 7,500 pounds and a specific impulse (measure of performance) of 278 lbf-sec/lbm (pounds of thrust per pound of propellant burned per second) at altitude. Even though Aerobee-Hi could not meet these specifications, Martin contracted with Aerojet on November 14, 1955, to develop the second-stage engine.

The third stage had to impart enough additional velocity to the payload to make it reach orbital speed. It could accelerate more easily than the overall launch vehicle because after the shedding of stages 1 and 2, it weighed much less. Also, when it began firing, it was further from the center of Earth and thus less impeded by Earth's gravitational pull. Somewhat paradoxically, therefore, it needed a higher mass fraction (the mass of the propellant divided by the total mass of the stage) than the other stages so it could take full advantage of its greater ability to accelerate. (To give an indication of how critical the third-stage weight really was, one source indicated that the addition of a pound to the first stage reduced the final velocity of the satellite by 1 ft/sec; for the second stage, the reduction produced by one pound of additional weight was 8 ft/sec; but for the third stage the reduction in speed rose tenfold to 80 ft/sec.) Thus team engineers decided to put the guidance and control equipment in the second stage rather than the third, and to use spin on the third stage to provide stability while it was burning. This in turn dictated a solid propellant with an internal-burning charge, as the spinning would make liquid propellants slosh, complicating their effective use. Martin determined that the third stage needed a weight of pounds, a mass fraction of 0.816 (compared with less than 0.7 for the first two stages), a specific impulse at altitude of 245, and a total impulse of 97,600 pounds per second. These parameters ruled out a Sergeant-type motor.

Several companies submitted proposals to furnish this stage. Since Martin and NRL expected developmental difficulties, Martin awarded a contract to the Grand Central Rocket Company in February 1956 and, in a parallel move, the Navy issued one in April 1956 to the Allegany Ballistics Laboratory (ABL), now operated by the Hercules Powder Company, to develop (p.21) third-stage motors. Having two contractors would improve the chances of getting a viable motor in time for the Vanguard launches.31

Even before Martin began awarding these subcontracts, it dealt a “disappointment, even a shock” to NRL. The Navy laboratory had learned unofficially that it had won the battle with the Army for the right to launch the IGY satellite, and early in September it had issued a letter of intent to Martin for work on the launch vehicle. Before doing so, the Navy had conferred with the Department of Defense, hoping thereby to avoid competition with other projects for Martin's rocket engineers. On September 14, Martin learned that it would win the contract to design, develop, and test the Titan I missile (as it became), with a letter contract signed with the Air Force in October and a final contract in January 1956. Since Vanguard's priority at the Department of Defense was low—the IGY satellite program, it was stipulated, must not interfere with military missile projects—while the much larger Titan program enjoyed a high military priority, Martin split up the experienced Viking team, sending only some of its Viking engineers to work on Vanguard. To be sure, the situation was not as dire as it sometimes has been portrayed. Martin's project manager for Vanguard, its chief engineer, and its chief of flight testing were all veterans of the Viking team. But NRL had counted on the entire Viking team transferring to Vanguard.32

Of more consequence, perhaps, for NRL perceptions and morale, Martin situated its Vanguard engineers in a loft of its aircraft plant that was extremely hot in summer, cold in winter, and subject to sparrow droppings. Rosen, despite being described as “intense” and “hard-driving,” had gotten along with the Martin engineers on the much smaller Viking project, but NRL's disappointment at Martin's splitting up the Viking team and operating from a balcony may have contributed to a less amicable working relationship with the Baltimore firm this time around. Another factor was a disagreement about approach. The Air Force allowed Martin considerable freedom to develop Titan I, whereas the NRL team wanted to monitor not only plans but execution. And where Martin was inclined to use shortcuts and empirical data in design and development, NRL favored greater emphasis on analysis. This contrast in approach added friction to what was already a difficult design-and-development effort.33

The First-Stage Engine

Even though the General Electric X-405 first-stage engine, as it came to be designated, was based on GE's A-3B, the modifications were numerous and difficult, making this a substantially new engine, as Rosen said. The A-3B had burned liquid oxygen and alcohol, but the X-405 replaced the alcohol (p.22) with kerosene as part of the e ort to raise the specific impulse from 225 to 254 lbf-sec/lbm.34 Available sources on Project Vanguard and GE do not indicate whether GE solved early problems with kerosene as a rocket fuel entirely on its own, but it could have learned of solutions to those problems from an Air Force program with North American Aviation.

In January 1953, Lt. Col. Edward Hall and others from Wright-Patterson Air Force Base insisted to Sam Ho man of NAA that he convert from alcohol to a hydrocarbon fuel for a 120,000-pound-thrust Navaho engine. Ho man protested, as the standard kerosene the Air Force used was JP-4, whose specifications allowed a range of densities. It also clogged a rocket engine's slim cooling lines with residues. The compounds in the fuel that caused these problems did not affect jet engines, but they simply would not work easily in rockets' power plants. To resolve these problems, Ho man initiated the Rocket Engine Advancement Program, which resulted in the development of the RP-1 kerosene rocket fuel, without JP-4's high levels of contaminants and variations in density. This fuel went on to power the Atlas, Thor, and Jupiter engines. The specifications for RP-1 were not available until January 1957, but that was still before the actual delivery date of the X-405 engine for test vehicle 2 of Vanguard.35

GE's X-405 may thus have been a direct or indirect beneficiary of NAA's work.36 In any event, other changes from the A3–B were an increase in the chamber pressure from some 450–500 pounds per square inch up to 616 psi and an extension of the burn from 50 to 146 seconds. The specific impulse rose to roughly the requisite 254 lbf-sec/lbm as well. Implementing these improvements brought the usual problems.

The higher pressure necessitated a slightly thicker chamber wall. It also pushed up the heat transfer rate from 0.4 to 0.6 BTU/in2/sec. This was offset by faster propellant flow through the regenerative cooling passages, but to raise the flow rates, the injectors had to be modified. The hydrogen peroxide in the turbopump system also had to flow more quickly to provide the higher pressure. Despite these issues, GE had the first production engine (P-1) ready for delivery on October 1, 1956. But then damage to the lining of the combustion chambers occurred during static testing of the P-2 and P-3 engines. Failures of two chamber liners in the P-4 engine delayed delivery until the problem could be solved. A redesign proved effective but required careful attention to injector specifications to preclude local hot spots and combustion instability. In fact, GE ended up testing fifteen injectors with six variations in design between January and April 1956 before finding one that would work. This was “rule-of-thumb” engineering, but it produced a (p.23) comparatively uncomplicated engine with a minimum of relays and valves that never experienced a burn-through in flight.37

The Second-Stage Engine

The development of Aerojet General's second-stage engine, designated AJ10–37, saw many more problems than GE's first stage. The stringent velocity requirements of stage 2 imposed severe weight limitations and called for a high specific impulse. The engineers selected unsymmetrical dimethyl hydrazine (UDMH) and inhibited white fuming nitric acid (IWFNA) as the propellants because they were hypergolic (self-igniting on contact, eliminating ignition issues), had a high loading density (reducing tank size), and delivered the requisite specific impulse. Another perceived advantage of hydrazine and acid was a comparative absence of combustion instability in experimental research.38

The history of the evolution from the aniline-nitric acid propellants used in the WAC Corporal39 and the first Aerobee sounding rocket40 to the UDMH and IWFNA used in Vanguard is complicated but worth relating because of what it illustrates about propellant chemistry and the number of institutions contributing to it. The basic aniline-RFNA combination worked as a self-igniting propellant pairing. But aniline is highly toxic and rapidly absorbed through the skin. A person who comes into contact with a significant amount of it faces a swift death from cyanosis. Moreover, aniline's high freezing point means it can be used only in warm weather. RFNA is highly corrosive to propellant tanks, so it has to be loaded just before firing, and when poured it gives o dense concentrations of nitrogen dioxide, which is also poisonous. The acid itself burns the skin as well.

Two chemists at JPL discovered as 1946 early as that white fuming nitric acid (WFNA) and furfuryl alcohol with aniline, while just as poisonous and corrosive, at least did not produce nitrogen dioxide. But WFNA turned out to be inherently unstable over time. Chemists in the rocket business throughout the country also found this complicated substance difficult to analyze. By 1954, however, those at the Naval Ordnance Test Station in Inyokern, California, and at JPL had thoroughly investigated nitrogen tetroxide and nitric acid and come up with conclusions that were to be used in the Titan II. Meanwhile, chemists in various places—among them JPL, the NACA's Lewis Flight Propulsion Laboratory, the Naval Air Rocket Test Station in New Jersey, the Air Force's Wright Air Development Center in Dayton, and Ohio State University—reached a fundamental understanding of nitric acid by 1951, with their findings published by 1955. The Naval Air Rocket Test (p.24) Station was apparently the first installation to discover that small percentages of hydrofluoric acid both reduced the freezing point of RFNA/WFNA and inhibited corrosion with many metals. Thus was born inhibited RFNA and WFNA, for which the services and industry representatives under Air Force sponsorship drew up military specifications in 1954. In this way the services cooperated to solve a common problem despite their continuing competition for roles and missions.

During the same period, chemists were also looking for replacements for aniline or chemicals that could be mixed with it to make it less troublesome. Hydrazine seemed to be a promising candidate, and in 1951 the Navy's Bureau of Aeronautics, through its Rocket Branch, issued contracts to a firm named Metallectro and to Aerojet to work with hydrazine derivatives and see if any were suitable as rocket propellants. The two firms found that UDMH rapidly self-ignited with nitric acid, leading to a military specification for UDMH in 1955.41

Despite the severe weight limitations on stage 2, Vanguard project engineers decided on a pressure-fed propellant delivery system rather than stage 1's pump-fed system. The pumps produced angular momentum as their turbines rotated, and for stage 2 this would be hard for the roll-control system to overcome. Concerns about reliability and development problems led to a decision to use heated helium gas as the pressurant. Aerojet convinced the Martin Company, Rosen, and Hagen to employ stainless steel instead of aluminum propellant tanks, because steel had a better strength-to-weight ratio. Aerojet argued that the lighter metal would, paradoxically, have had to be 30 pounds heavier than steel to handle the pressure.

Moreover, a “unique design for the tankage” placed the sphere containing the helium pressure tank between the two propellant tanks, serving to divide them, thus saving the weight of a separate bulkhead. A solid-propellant gas generator augmented the pressure of the helium and added its own chemical energy to the system at a low cost in weight. Initially Aerojet built the combustion chamber entirely of steel, and it logged 600 seconds of burning without corrosion. However, it was too heavy. So Aerojet engineers developed a lightweight chamber made up of spaghetti-type aluminum regenerative-cooling tubes wrapped in stainless steel, cutting the weight by more than 20 pounds. This chamber was apparently the first built of aluminum tubes to be tested with nitric acid and UDMH.42

Despite Aerojet's experience in this area, in 1956 problems arose with welding stainless steel tanks. They were resolved when, at Martin's recommendation, Aerojet made improvements in tooling and inspection. The California rocket engine firm also had to try several types of injectors before it (p.25) found the right combination of features. One with 72 pairs of impinging jets did not deliver sufficient exhaust velocity, so Aerojet engineers added 24 nonimpinging orifices for fuel at the center. This raised the exhaust velocity above specifications.43

Despite the use of inhibited white fuming nitric acid, the lightweight aluminum combustion chamber, which could be lifted with one hand, experienced gradual erosion. It took engineers “weeks of experimenting” to find out that a coating of tungsten carbide substantially added to the chamber's life. There also were problems with valves for flow control, requiring significant redesign of the valves and rerouting of the “plumbing.”44

A final problem lay in testing an engine that had to ignite at an altitude upwards of 30 miles. At the outset of the project, there was no vacuum chamber large enough, but apparently Aerojet acquired or built one, because NRL propulsion engineer Kurt Stehling wrote, “Several tests were made at Aerojet with engine starts in a vacuum chamber.” In any event, to preclude problems with near-vacuum pressure at altitude, the engineers included a “nozzle closure” that kept pressure in the combustion chamber until exhaust pressure from ignition blew it out.45

The Third-Stage Motors

One firm the Martin Company selected to develop the third stage for Vanguard was a relative newcomer to the rocket business. Charles Bartley had left JPL in 1951 and founded the Grand Central Rocket Company in 1952. Initially located in Pacoima, in the San Fernando Valley, it moved eastward in to Mentone, just past Redlands in San Bernardino County. Bartley served as the president until September 1958, when he left the firm to organize two companies that produced solid-propellant equipment such as sounding rockets and small propulsion devices to eject pilots from aircraft. At Grand Central he hired Lawrence Thackwell from Thiokol and both Larry Settlemire and John I. Shafer, who, like Thackwell, had worked for him at JPL. Thus he brought considerable talent and knowledge from JPL to the new company. Ultimately Grand Central passed through interim owners to Lockheed, becoming Lockheed Propulsion Company in 1961.46

In the meantime, the company under Bartley's direction contracted for and developed the third-stage motors that flew on all Vanguard launches except the final one (excluding test launches without a third stage). The motor used Thiokol's LP-3 polysulfide rubber as the binder and fuel (28 percent of the propellant) and ammonium perchlorate as the oxidizer (71 percent of the propellant). It was configured with a five-pointed, internal-burning star. The case was very thin stainless steel. The propellant burned at a low (p.26) chamber pressure, making the thin case possible, but Cooper Development Corporation of California, which made the case, had to spin and roll it in extremely thin sheets, then weld and heat-treat it properly for it to work. Coating the nozzle with aluminum oxide provided protection against overheating. Since ignition of this stage occurred more than 200 miles above the earth, it required a high-energy igniter plus sealing of the throat to keep in some pressure until firing began. The igniter used the same propellant as the motor.

The propellant yielded a specific impulse at altitude of 239 lbf-sec/1bm, slightly below the design figure of 245, and the metal parts initially weighed instead of the specified 53 pounds. To compensate for the underperformance of the propellant, a product improvement program decreased the weight of the parts so as to raise the mass fraction and achieve the necessary velocity. By measures that included reducing the thickness of the liner between the propellant and the case, the engineers lowered the weight by five pounds. There was also evidence of combustion instability, but it was not intense enough to be significant. Testing of the motor revealed problems with cracking of the propellant at the star points, as had occurred at JPL with the Sergeant test vehicle. Placing polyvinyl acetate cement at the star points inhibited the cracking. As of June 1957, Grand Central Rocket had completed development of stage 3 and satisfied the specifications to reach orbital velocity.47

The alternate third stage was fated to be flown on only one Vanguard mission, the last of the series, because of problems with its technical development. But it was more innovative than Grand Central's motor, and it (or variants of it) later found use in military missiles as well as both the Delta and the Scout launch vehicles. The firm that produced this motor came from a different lineage than the other solid-propellant firms discussed up to this point.

The Hercules Powder Company, which had operated the government-owned Allegany Ballistics Laboratory since the end of World War II, came into existence in 1912. An antitrust suit forced its parent company, E. I. du Pont de Nemours, to divest some of its holdings, one of which became Hercules. The spinoff firm began as an explosives company that produced more than 50,000 tons of smokeless powder during World War I and then began to diversify into other uses of nitrocellulose. During World War II the firm supplied large quantities of double-base propellants for tactical rockets, using an extrusion process. After the war Hercules began using a technique for casting such propellants. Chemists poured a casting powder consisting (p.27) of nitrocellulose, nitroglycerine, and a stabilizer into a mold and added a solvent of nitroglycerine plus a diluent and a stabilizer. With heat and the passage of time this yielded a much larger grain than could be produced by extrusion alone.

Wartime research by John F. Kincaid and Henry M. Shuey at the National Defense Research Committee's Explosives Research Laboratory, operated by the Bureau of Mines and the Carnegie Institute of Technology in Bruceton, Pennsylvania, south of Pittsburgh, had produced this process. Kincaid and Shuey along with other propellant chemists had developed it further after transferring to ABL, and under Hercules management ABL continued work on cast double-base propellants. Flight testing of a JATO using this propellant followed in 1947. The casting process allowed Hercules to produce a propellant grain as large as the castable composite propellants that Aerojet, Thiokol, and Grand Central were developing in this period but with a slightly higher specific impulse—albeit a greater danger of exploding rather than releasing the exhaust gases at a controlled rate.48

Although the Navy had let the contract to Hercules for this motor, designated JATO X241 A, it had delegated responsibility for technical coordination to the Glenn L. Martin Company. The first propellant Hercules' Allegany Ballistics Laboratory used for the motor was a cast double-base formulation with insulation material between it and the case. This propellant yielded a specific impulse of about 250 lbf-sec/lbm, higher than Grand Central's propellant and higher than the specification of 245 lbf-sec/lbm for both motors. A key feature of the motor was its case and nozzle, composed of a laminated fiberglass made from epoxy resin. ABL had subcontracted work on the case and nozzle to Young Development Laboratories, which in 1956 developed a method of wrapping threads of fiberglass soaked in epoxy resin around a liner made of phenolic asbestos. (A phenol is a compound used in making resins to provide laminated coatings or form adhesives.) After curing, this process yielded a strong, rigid shell with a strength-to-weight ratio 20 percent higher than the stainless steel Aerojet was using for its propellant tanks.49

In 1958, with its third-stage motor still under development, Hercules acquired the fiberglass winding firm. Its founder, Richard E. Young, was a test pilot who had worked for the M. W. Kellogg Company on the Manhattan Project. Kellogg had designed a winding machine in 1947 under Navy contract, leading to a winding laboratory that built a fiberglass nozzle. In 1948 the operation moved to Rocky Hill, New Jersey. There Young set up labs under his own name and tackled the problem of strength-to-weight ratios (p.28) in rocket motors through developing lighter materials. He and the firm evolved from nozzles to cases, seeking to improve a rocket's mass fraction, which was as important as specific impulse in achieving high velocities. In the mid-1950s, ABL succeeded in testing small rockets and missiles using cases made with Young's Spiralloy material.50

This combination of a cast double-base propellant and the fiberglass case and nozzle created a lot of problems for Hercules engineers. By February 1957, ABL had performed static tests on about twenty motors, with fifteen of those firings producing failures of insulation or joints. Combustion instability arose in about a third of the tests. Hercules installed a plastic paddle in the combustion zone in an attempt to interrupt the acoustic patterns (resonance) that caused the instability. This did not work as well as hoped, so the engineers developed a suppressor of thicker plastic. They also improved the bond between the insulator and the case, then cast the propellant in the case instead of just sliding it in as a single piece. Despite these modifications, nine cases failed during hydrostatic tests or static firings. The culprits were high stress at joints and “severe combustion instability.”51

In February 1958, in addition to X241, ABL began developing a follow-on third-stage motor designated X248 A2. Perhaps the lab did so in part to reduce the combustion instability, for the new motor's propellant was 3 percent aluminum, which upon burning produced particles in the combustion gases that suppressed (damped) high-frequency instabilities. But another motivation was increased thrust. The new motor was the one that actually flew on the final Vanguard mission on September 18, 1959. As of August 1958, ABL had already developed a modification of this motor, X248 A3, for use as the upper stage in a Thor-Able lunar probe. By this time, also, ABL was testing the motors in an altitude chamber at the Air Force's Arnold Engineering Development Center in Tullahoma, Tennessee, and was experiencing problems with ignition and with burn-throughs of the case.52

The X248 solid rocket motor consisted of an epoxy-fiberglass case filled with the case-bonded propellant. The nozzle was still made of epoxy fiberglass, but its coating was now a “ceramo-asbestos.” By November 11, 1958, wind-tunnel static tests showed that the X248 A2 filament-wound exit cone was adequate. The motor had a specific impulse at altitude of 256 lbf-sec/ 1bm, and its other problems had been overcome. The X248, wrote Stehling, offered “considerable improvement in reliability and performance over the X241 contracted for originally” and succeeded in launching the Vanguard III satellite weighing 50 pounds, whereas Grand Central Rocket's version of the third stage could orbit only about 30 pounds of payload.53

(p.29) Aerodynamics

As finally designed, the Vanguard launch vehicle was 71 feet 2 inches long. Its first stage was 45 inches in diameter, tapering to 32 inches for the second and third stages. Unlike the Viking, it was finless with an integral-tank construction (the tanks serving as structural support), and it had a weight at liftoff of 22, 600 pounds. These dimensions made for a slender rocket. As such, it had a low bending moment, meaning that it could bend relatively easily under the force of winds or other air “loads.” NRL and Martin consequently arranged for a wide variety of wind tunnel tests at low speeds, transonic speeds (roughly Mach 0.8 to 1.2 ), and supersonic speeds, using different tunnels for each range since there was no single tunnel that could provide accurate data over the entire velocity scale from Mach 0.0 to 3.5. (Speeds above three and a half times the speed of sound would occur at such high altitudes that the density of the atmosphere and the resultant air loading were too low to be a concern.)

Table 1.1 shows the wind tunnels used for Vanguard testing and the speed ranges tested. There were two tests at supersonic speeds because of configuration changes between the first and second test. A second series of low-speed tests was occasioned by concern about the “variable Reynolds number effect at high angles of attack.”54

The tests at MIT were ordered because Vanguard's relative slenderness made it more subject to the von Kármán vortex-shedding effect than other

Table 1.1. Wind Tunnel Testing for Vanguard

Test Date

Wind Tunnel

Speed Range


Aberdeen Proving Ground Supersonic Wind Tunnel #3

Mach 2.0–4.0


University of Maryland Low-Speed Wind Tunnel

150 mph at dynamic pressure of 57.55 lb/ft2


University of Maryland Low-Speed Wind Tunnel

50–200 mph


Naval Air Missile Test Center, Point Mugu, California, Supersonic Wind Tunnel

Mach 1.6–3.5


Wright Air Development Center, Wright-Patterson AFB, Ohio, Transonic 10-Foot Wind Tunnel

Mach 0.8–1.2


MIT Aero-elastic Wind Tunnel

40–75 mph

Sources: “Project Vanguard Report … 1 June 1957,” 2–13 to 2–15, 2–102, NHRC; Hagen, “The Viking and the Vanguard,” 130; Martin, “Vanguard Satellite Launching Vehicle,” 18, NHRC.

(p.30) rockets while on the launchpad. Named after the famous aerodynamicist who discovered it, the von Kármán effect could result from a steady wind that produced oscillatory forces or vortices; these would form periodically on the sides of the cylindrical fuselage and then disengage, imparting forces that could destroy the rocket on the pad if they were of the proper magnitude and frequency. Analysis and testing had to take into account conditions before and during propellant loading, as the amounts of propellants affected the proclivity of the long, thin rocket to bend and possibly break apart. A preliminary theoretical analysis indicated that wind velocities between 10 and 18 ft/sec (about 7 to 12 mph) could produce serious enough vibrations to destroy the vehicle under certain propellant loads. The results of the wind tunnel tests at MIT were quite complex and their scope had to be limited, but they provided the Vanguard team with rough parameters for determining safe launch conditions. They also led engineers to develop mechanisms—removable rubber spoilers on the second stage—to diminish the effects. Technicians mounted twelve of these spoilers, with drag cones attached. The spoilers suppressed the vortices, while the drag cones ensured that the spoilers fell o the vehicle after liftoff, when the oscillations were no longer a problem.55


The Vanguard structure was mainly monocoque, meaning that the outer skin bore the major portion of the bending forces, with the pressurized internal tanks also serving as structural supports. The materials for the first stage were principally aluminum and magnesium alloys. The second stage contained the propellant and helium tanks, which were made of stainless steel, as was the motor casing for the third stage except in the final Vanguard configuration where the ABL fiberglass case was used.56

Guidance and Control System

The final major concern of Vanguard's designers was the guidance and control system in stage 2. Experience with Viking had demonstrated that fins for stabilization could actually cause destabilization if they were not lined up properly. Besides, they added weight. Martin engineers had designed a gimballing system that was sensitive enough not to require fins, but the Navy's budget for Viking did not permit implementing the system. It was used later on Vanguard, which could then dispense with fins.57 Some Deltas and Titans subsequently followed the Vanguard precedent and also did not use fins.

(p.31) Controlling the gimbals on both the first and second stages was a system of electrohydraulic devices and an inertial reference system. In designing this system, project engineers had to analyze a variety of factors, using “extensive mathematical predictions as a basis for … decisions,” because until the design was completed and the vehicle built, the interacting systems could not be tested together. Design engineers had to consider not only the effects of gimballing the engines but also concerns raised by the rocket's light and flexible structure. As the structure, designed to yield a high mass fraction, bent under various aerodynamic loads, the vehicle could go off course. The gyro system would sense the displacement and send a signal to the engine gimbal. The corrective action would produce further bending of the structure, which could break under too heavy a load. Guidance and control designers also had to use differential equations to analyze data that aerodynamicists got from wind tunnel and flight tests of earlier rockets. Then they verified characteristics predicted from results of wind tunnel tests using the Vanguard configuration in models. Control system analysts looked at requirements for Vanguard and the various data to arrive at characteristics for the control system. As they learned to deal with such interactive factors, “the art of rocketry” was “emerging as a distinct branch of technology,” one engineer wrote in 1958.58

Because of Vanguard's lack of fins, controlling it was like balancing a broomstick. To keep the stick upright, a person has to anticipate any deviations and counter them in advance by moving the upturned hand on which the end of the stick rests. Similarly, the guidance and control system had to sense any deviations from the intended trajectory and move the engine's thrust vector in the proper direction to counter them before they got too large. To see if it would do this, Vanguard engineers tested the control system on an analog simulator. They also checked it on environmental testing devices for effects of such things as vibration, temperature, and humidity, then conducted in-flight tests on the Vanguard itself.59

The guidance and control system that Vanguard engineers designed and tested in this way consisted of components from several subcontractors to Martin. The Minneapolis-Honeywell Regulator Company developed the gyroscopic reference system, which was about as large as a basketball. It contained three 4.6–pound hermetic integrating gyros suspended in a viscous fluid, Flurolube, that greatly reduced the friction they encountered. Each gyro operated in a single degree of freedom to sense motions in the pitch, yaw, or roll axis. Each was arranged with a gimbal, torque generator, and spin motor (to start the gyro spinning and control its operation), plus (p.32) a heater (to maintain the proper viscosity of the Flurolube, which changed with the temperature).

As the pitch or yaw gyro sensed deviation by the rocket, the gyro precessed and made contact with a picko that sent a signal to a subminiature electronic autopilot. Developed by Martin, the autopilots (one per axis) amplified the signal and sent it to a servo system to control the gimbal-mounted first- or second-stage engine. The roll gyro sent signals to auxiliary on-off jets on the sides of the two stages via a third autopilot. Unlike the system for the Redstone missile, the Vanguard did not use a stabilized platform (that is, one that was free to rotate with respect to the rocket but was kept steady in space by gyros). Instead it was strapped down in a fixed position on the rocket. This arrangement was lighter than available stable-platform types of guidance and control systems. Yet it was accurate enough to satisfy Vanguard's requirements. Later, when small but powerful digital computers became available in the late 1970s, strapdown configurations replaced the stabilized platform in many uses. But already in 1960, unspecified ballistic missiles were adopting strapdown systems.60

Another major component of the guidance and control system was the program timer, designed, developed, and produced by the Cleveland engineering firm Designers for Industry. This small transistorized unit weighed less than 6 pounds but was highly accurate. It signaled the start of the pitch program, ignition for the second stage, stage separation, and backup ignition for the third stage. The Vanguard launched vertically, after which the program timer signaled the pitch gyro's torque generator to rotate the gyro's gimbal, leading to a gradual tilting from the vertical. To minimize the aerodynamic forces operating on the vehicle during flight powered by the first stage, the tilt was kept minimal until first-stage propellant exhaustion created a signal for the second-stage engine to ignite, resulting in separation from the first stage at about 36 miles of altitude. The first stage had provided at that point about 65 percent of the energy necessary to raise itself and the remaining stages plus the satellite to orbital altitude, and about 15 percent of the orbital velocity.

The vehicle (minus stage 1) tilted much further during the second-stage burn, which ended at about 140 miles, whereupon the Vanguard coasted to its orbital altitude of about 300 miles. During this period, control shifted entirely to jet devices that operated on pitch and yaw as well as roll. The second stage provided about 32 percent of the orbital velocity. The third stage had no guidance; solid-propellant rockets, supplied by Atlantic Research Corporation, imparted spin to maintain stability. This stage fired at orbital (p.33) altitude, by which time the second and third stages had tilted to a completely horizontal attitude. The third stage provided 50 percent of the orbital velocity—about 18,000 mph relative to Earth's surface—the remaining 3 percent coming from geophysical effects including Earth's rotation, since Vanguard was launched to the east.61

During coasting flight, another major component of the flight control system came into play: the coasting time computer. Provided by Electronic Communications, this device ascertained the velocity of the vehicle when the second-stage engine burned out and computed the proper time for the third-stage motor to ignite. An integrating accelerometer provided acceleration data throughout the Vanguard's flight up until second-stage cutoff. From the information it supplied, the computer calculated the vehicle's velocity at stage-2 burnout and energized a timing motor. This in turn signaled the spin-up of the third stage and its separation via Atlantic Research's two small solid-propellant retro-rockets. Fifteen seconds after the spinning of the third stage began, a delay fuse fired, initiating ignition of the rocket motor.62

Flight Testing

The original Vanguard schedule as of November 1955 called for six test vehicles to be launched between September 1956 and August 1957, with the first satellite launching vehicle to lift off in October 1957—by happenstance, the month of Sputnik. Had the project remained on schedule, it is conceivable that NRL could have launched a satellite at about the same time as the Soviet Union. As it was, problems with the first- and second-stage engines caused delays. On December 8, 1956, already more than two months late, Viking rocket number 13 was launched to test the Vanguard launch complex and the telemetry system, develop familiarity with range safety, and the like. The test was largely successful, but by February 1957 even Viking number 14 had not flown. Rescheduled for the end of April, that vehicle actually launched on May 1 with a prototype of the Grand Central third stage tested as the Viking's second stage for spin-up, separation, ignition, propulsion, and trajectory (the actual second stage not flying on this mission). Although the sequence of spinning up and separating for the Grand Central rocket was “untried and complicated,” the test was successful and almost in accord with the revised timetable.63

However, the third test vehicle (designated TV-2 because the first test with the Viking rocket was TV-0 ), involving a prototype Vanguard first stage with inert second and third stages, could not be launched until October 23, 1957. (p.34) This was almost three weeks after Sputnik and four months behind the June liftoff date stipulated back in February. For guidance, TV-2 contained only some of the guidance-and-control and other electronic components of a complete Vanguard vehicle. The old Viking system still provided much of the control. Before the test vehicle left the Martin plant, the contractor's crew found that the roll jet and pressurization systems were not performing according to specifications and had to be partially redesigned. Despite numerous hardware difficulties along the way, TV-2 finally launched successfully and met all of its objectives. The first stage operated as designed, and conditions appeared favorable for successful separation of the first and second stages.64

By the time of this launch, pressures were mounting on John Hagen. The project director had been up on Capitol Hill to brief the staff of the formidable Senator Lyndon B. Johnson and his Senate Preparedness Subcommittee, which was scrutinizing the nation's missile and satellite programs in the wake of Sputnik. Hagen had also briefed President Eisenhower on both TV-2 and TV-3, the second of which was now scheduled for launch on December. Already in July, before the furor attending the Soviet satellite, the Navy had decided to launch TV-3 and TV-4 with minimal satellites weighing 3.4 pounds instead of an instrumented nose cone for the original 21.5–pound satellite. Hagen had so indicated to Eisenhower, pointing out that there was no guarantee they could be put into orbit. Despite this warning, the president's press secretary informed reporters that during December the Vanguard project would launch a test vehicle with a satellite on board. The press, of course, seized upon the mere test as America's answer to Sputnik.65

TV-3 would be the first Vanguard launch with all stages operational. Its mission was not just to launch a small satellite but to test and evaluate all three stages. More particularly, it would be the first test flight of the problem-ridden second stage as well as the complete guidance and control system. It was rare in those days for any rocket to have three successful tests in a row, as Vanguard had. The odds were certainly against a fourth. Still, TV-3 had passed its preflight “functional, instrumentation, and restrained-firing tests” and, following delays for various reasons, sat poised on the launchpad on December 6, 1957, ready to send the grapefruit-sized satellite into orbit to join its larger Soviet cousin. Then the odds caught up with Vanguard. The first stage ignited, but the test vehicle rose only slowly, “agonizingly hesitated a moment … and … began to topple [as] an immense cloud of red flame from burning propellants engulfed the whole area.” The press did not (p.35)

Viking and Vanguard, 1945–1959

Figure 1. Vanguard test rocket TV-3 exploding on the launchpad. A malfunction in the first stage caused the vehicle to lose thrust after two seconds, topple over, and burn. U.S. Navy photo courtesy of NASA.

react charitably but called the vehicle “Kaputnik, Stayputnik, or Flopnik,” while Americans, in one historian's words, “swilled the Sputnik Cocktail: two parts vodka, one part sour grapes.”66

In the somber wake of the colossal failure in public relations, technicians and engineers from GE and the Martin Company pored over records from (p.36) ground instrumentation, films of the failed launch, and the two seconds of telemetered data from the toppling inferno that was to have orbited America's first satellite. Not surprisingly, the Martin people blamed the problem on an “improper engine start” caused by low fuel tank pressure, whereas those from GE said there was no improper start and blamed a loose fuelline connection. The Martin engineers turned out to be correct, although there were more problems than low fuel tank pressure. To solve that one, GE negotiated a specification change with Martin to increase the minimum pressure in the fuel tank by 30 percent. But telemetry data also showed a high-pressure spike on engine start that the GE engineers had not caught in testing because of low-response instrumentation. The pressure spike had ruptured a high-pressure fuel line, causing destruction of the rocket. The remedy was to increase from 3 seconds to 6 seconds the period when oxygen was injected into the combustion chamber ahead of the fuel. The modified engine worked without problems in fourteen static and flight tests following this disaster.67

Meanwhile, waiting in the wings was a backup (BU) version of TV-3, virtually identical to the vehicle that had failed to launch. Following repairs to the launchpad, on February 5, 1958, stage 1 of TV-3BU fired properly and the test vehicle lifted off successfully to the cheers of the Vanguard team. However, 55 seconds into the launch, at an altitude of about 1,500 feet, the control system malfunctioned. Investigation later showed that a broken wire or connection had caused the vehicle to go out of control and break apart at a point between the first and second stages because of structural loads on the fuselage.68

TV-4 contained modifications made to the stage-1 engine following the failure of TV-3, but it did not yet incorporate the tungsten-carbide coating in the aluminum combustion chamber of the stage-2 engine. And it was still a test vehicle. On March 17 the slender Vanguard launch vehicle lifted o on the first actual flight test of the troubled second stage. It performed well enough this time to place the small 3.4–pound Vanguard I satellite in a highly stable orbit. The strapdown guidance system proved itself by producing an error of less than one degree in the satellite's angle of injection. It did this despite problems with the roll-control jets and a rough start by the second-stage engine. The mission was a huge, if belated, success. The Vanguard team had developed a complex, high-performance launch vehicle in a mere two years, six months, and eight days—only eight days longer than Rosen's initial estimate, though considerably longer than the revised one submitted to the Stewart Committee the second time around.69

The remaining Vanguard launches went as shown in table 1.2. Midway (p.37) through the launching cycle, on November 30, 1958, Project Vanguard transferred to NASA. Project personnel remained at NRL for a time, but NASA delayed further launches until a committee of rocket experts not associated with the project looked at it. The committee proposed some slight changes in test procedures and circuitry, and the project personnel continued with the remaining four launches.70

Table 1.2. Vanguard Launches


Description and Results

TV-5 4/28/58

Final test vehicle, like a satellite launching vehicle but with additional instrumentation. Attempted to launch a 21.5-pound satellite. First two stages satisfactory, with stage 2′s below-normal thrust offset by stage 1′s better-than-normal performance. Stage-2 cutoff sequence interrupted electrically, preventing coasting-flight control system from igniting stage 3. Upper stages coasted to 358 miles of altitude but, having insufficient speed to go into orbit, fell into the ocean.

SLV-1 5/27/58

First nontest satellite launching vehicle. Attempted to launch a 21.5-pound satellite. Normal flight through stage-2 cutoff, when a pressure switch apparently malfunctioned, disrupting the pitch gyro and sending stage 3 to an altitude variously estimated between 1,850 and 2,200 miles but not into orbit.

SLV-2 6/26/58

Stage-2 engine cut off after eight seconds of burning, probably because corrosion of the oxidizer tank caused clogging of filters in the inhibited white fuming nitric acid lines.

SLV-3 9/26/58

Attempted, after flushing the oxidizer tanks of stage 2, to launch a 23.3-pound satellite. Stage-2 performance was below normal, and the satellite narrowly missed orbital speed. Again the problem seemed to be clogging, but this time of a fuel filter, not an oxidizer filter.

SLV-4 2/17/59

Placed the 23.3-pound Vanguard II satellite into orbit.

SLV-5 4/13/59

Suffered violent yaw on stage-2 ignition, apparently from flame oscillation, causing stages 2 and 3 with the satellite to tumble in the pitch plane and fall into the ocean.

SLV-6 6/22/59

In stage 2, with a hydraulic system changed by Aerojet and separation reprogrammed to occur earlier, a previously reliable regulating valve failed after ignition. Helium pressure could not vent, and an explosion sent the vehicle into the Atlantic about 300 miles downrange.

TV-4BU 9/18/59

Test vehicle backup, equipped with ABL's X248 A2 third stage. Launched 52.25-pound X-ray and environmental satellite Vanguard III into orbit.

Sources: Stehling, Project Vanguard, 223–42, 269–81; Green and Lomask, Vanguard, 283–87; U.S. Congress, “Project Vanguard,” 66; memo, Chief of Naval Research to Director, Advanced Research Projects Agency, September 25, 1958, in RG 255, box 1, ARPA folder, NA; Newell, “Launching of Vanguard III,” NHRC; and seven Space Activities Summary documents: for SLV-1, SLV-2, SLV-3, and SLV-6 in Vanguard folders 006637–006639 and 006642, NHRC; for SLV-5 in “Vanguard Launch Vehicle 5,” folder OV-106121–01, NASM; for Vanguard II in “Vanguard II (Feb. 17, 1959),” NHRC; for Vanguard III in “Vanguard III Satellite Launch Vehicle 7,” NASM.

(p.38) Analysis and Conclusions

Given this inconsistent performance, what, finally, can be said about Project Vanguard, as well as its predecessor, the Viking rocket? Were they successes or failures? What did they contribute to the evolution of launch vehicle technology? Viking was a small program that contributed more than often is realized. Its early use of aluminum was a step toward that metal's extensive use in later missiles and launch vehicles. Viking's gimballing was an advance beyond the swiveling technology used on the MX-774 test vehicle,71 and designs that Martin drew up during the Viking project led to a further evolution of the technology in Vanguard that prepared the way for extensive use of it on many other missiles and launch vehicles. Viking also afforded experience that helped with Project Vanguard.

Given the limited expectations initially invested in Project Vanguard, it succeeded much better than often has been thought. It did put a satellite into orbit before the end of the IGY, although the grapefruit-sized Vanguard I was smaller than the satellite NRL had intended to orbit. After the end of the IGY, the project put two other satellites into orbit that were of the requisite size, one of them more than double the originally intended size. The Vanguard team did this despite the low priority the Department of Defense allotted to the project and the injunction not to interfere with high-priority missile projects. It also launched the three satellites in the midst of the enhanced expectations that came with the launch of Sputnik and the public clamor it produced. Another obstacle was the splitting up of the Viking team by Martin after it won the Titan I contract. Developing a substantially new launch vehicle of three stages at a time when the technology was still not mature and doing this in a comparatively short time were also notable achievements.72 They were, however, diminished by the salesmanship that had won Project Vanguard the approval of the Stewart Committee and the Department of Defense. Giving the impression that the project would use stages that were nearly developed and saying it could do so in eighteen months created goals that the team could not meet, as Rosen sensed.

Measured against the contributions of Vanguard to launch vehicle technology, rather than NRL's own initial marketing, the project appears much more successful. The Air Force's Thor-Able launch vehicle used the Thor intermediate range ballistic missile as a first stage and modified Vanguard second and third stages, the last being the original third stage developed by Grand Central Rocket Company. The air arm had better success with the second stage than did the Vanguard Project, for two reasons. One was special cleaning and handling techniques for the propellant tanks that came (p.39) into being after Vanguard had taken delivery of many of its tanks. Also, Thor-Able did not need to extract maximum performance from the second stage as Vanguard did, so it did not have to burn the very last dregs of propellant. This residue contained a disproportionate amount of scale from the tanks, but the Air Force could close the valves before the scale entered the fuel lines in the regenerative cooling jacket.73 Thus the important Thor-Able system not only benefited from two stages of Vanguard and took over their technology; it also learned from problems that Vanguard experienced and avoided them.

In January 1959, Rosen proposed to Abe Silverstein, NASA's director of space flight programs, that the Thor-Able be evolved into what became the Delta launch vehicle. Drawing on his experience with Vanguard, Rosen suggested substituting the ABL third stage for the one from Grand Central Rocket, designing more reliable control electronics than Vanguard's, replacing the aluminum combustion chamber in the Vanguard second stage with a stainless steel one, and adopting Bell Telephone Laboratories' radio guidance system then being installed in the Titan ballistic missile. Silverstein commissioned Rosen to develop the Delta launch vehicle along those lines, and it became highly successful.74

A variant of the ABL third stage for Vanguard, known as the Altair I (X248 A5), later became the third stage for Delta and the fourth stage for the all-solid-propellant Scout launch vehicle. A follow-on, also built by Hercules Powder Company (at ABL), became the third stage for Minuteman I. And the fiberglass casing for the ABL third stage was a feature of these later stages and found many other uses in missiles and rockets.75 The strapdown guidance and control system, finally, although it had imperfect electronics, was another contribution to missile technology and to later launch vehicles.

These were significant legacies to American rocketry. Clearly, if they are taken into consideration, Vanguard was an important success. Impressively, the industry-Navy team made this contribution with some 15 people at NRL and a total team of only 180. Its cost was $110 million. “It wasn't first in orbit, but it did its job and lived up to its name of being the vanguard of many space projects to follow.”76


(1.) See Hall, “Vanguard and Explorer,” 101–4; Hall, “Earth Satellites,” 111; and sources cited in this chapter.

(2.) Biographical information on Rosen is from his official NASA biography, February 6,1962, in “Rosen, Milton W.,” NHRC.

(3.) Ibid.; Rosen, OHI; Rosen, Viking, 18–20. The Pendray book Rosen mentioned was The Coming Age of Rocket Power (1945), which Rosen described as “a proposal that rockets be used for exploring the upper atmosphere” (18). On NRL, the first quotation is from a pamphlet with the NRL seal on its cover and “1923–1973” inscribed inside, in “Naval Research Laboratory,” folder 012164, NHRC; the others are from Hevly, “Tools of Science,”221–22.

(4.) Rosen, OHI,29–30(quotation, 30); Rosen, Viking, 20–21.

(5.) Rosen, OHI,31; Rosen, Viking, 22–23,66 (quotation). Here and below in this chapter I have made corrections based on comments by Rosen written and telephoned in May 2002.

(6.) Rosen, Viking,26; Rosen, OHI,31.

(7.) Rosen, Viking, 26–27; Martin, “Design Summary,”5, NASM.

(8.) See Hunley, Preludes,153.

(9.) Rosen, OHI, 38–40; Rosen, Viking, 27–28; Rosen, comments, UP. On Anderson, see Goodstein, Millikan's School, 105–7,252,271.

(10.) Rosen, OHI,44, 52–53; Rosen, Viking,28,64; Martin, “Design Summary,” 36–37, NASM; Harwood, Raise Heaven and Earth, 256; Rosen, comments, UP.

(11.) Harwood, Raise Heaven and Earth,253; Martin, “Design Summary,”,5,6,99 NASM. For V- 2's alcohol percentage, see Hunley, Preludes, chap. 2; for hydrogen peroxide concentration, Kennedy, Vengeance Weapon, 2,77; Rosen, comments, UP.

(12.) Rosen, Viking, 58–62, 236–37; Winter and Ordway, “Pioneering Commercial Rocketry,” 162–63; Martin, “Design Summary,” 104–10, NASM; Scala, “Viking Rocket,” 34, which contains a handy launch chronology; Rosen, comments, UP (quotation).

(13.) For Rosen's claim, see his OHI,44. On Bossart's relationship to swiveling, J. Neufeld, Ballistic Missiles,47, says the Germans (meaning the von Braun group) first conceived the idea and discarded it but then Bossart came up with the design independently. Swenson, Grimwood, and Alexander, This New Ocean,22, says, “Bossart and associates proposed a technique basically new to American rocketry (although patented by Goddard and tried on some German V-2s)—controlling the rocket by swiveling the engines, using hydraulic actuators responding to commands from the autopilot and gyroscope. (p.339) This technique was the precursor of the gimballed engine method employed to control Atlas and other later rockets.” Chapman earlier wrote, “Bossart's ‘swinging’ powerplant, forerunner of the gimbal control system (movable in any direction) prevalent in today's liquid-propellant rockets, was the first of its type anywhere” (Atlas, 34), then noted, “He was to learn a few years later that the Germans had experimented with the idea on the V-2” but had discarded it because of the “plumbing problem” with the 18-topf V-2 engine, which would have been difficult to rotate.

(14.) J. Neufeld, Ballistic Missiles,26, 45–49; Winter and Ordway, “Pioneering Commercial Rocketry,” 161–62; Convair, “MX-774 Flight Test Report,” NASM.

(15.) See quotations in note 13; cf. Convair, “MX-774 Flight Test Report,” NASM, with diagrams of swiveling cylinders; Rosen, Viking,63, with diagram of Viking gimballed motor.G. Sutton, Liquid Propellant Rocket Engines,220, agrees that the Viking had the first gimballed rocket engine to fly.

(16.) Hagen, “Viking and the Vanguard,”124.

(17.) Martin, “Design Summary,” 5, 60–68,210, NASM; Harwood, Raise Heaven and Earth, 254.

(18.) According to Harold F. Klock, “Feedback Circuit,” in McGraw-Hill Encyclopedia of Science and Technology,7:36, “In automatic control systems, feedback is used tocompare the actual output of a system with a desired output, the difference being used as the input signal to a controller.” And, “When the feedback signal is of opposite phase to that of the input signal, the feedback is negative …”

(19.) Rosen, comments, UP (quotation, May 8). He called the methodology for adjusting the system the Nyquist Diagram.

(20.) Rosen, Viking, 172–73 and passim; Rosen, Bridger, and Jones, “Viking 8 Firing,” 2–5, NASM.

(21.) Rosen, Viking,132,236; Rosen, Bridger, and Snodgrass, “Viking 9 Firings,” 5–6, NASM; Harwood, Raise Heaven and Earth,25.

(22.) Rosen, “Viking and Vanguard,”7, NHRC.

(23.) Rosen wrote (comments, UP): “This was our idea and I proposed it to Gen. Schriever, but we could not possibly meet [the Air Force's] time schedule, as he patiently explained to me.” For background on IGY, see Hunley, Preludes, chap. 6.

(24.) See Hunley, Preludes.

(25.) Green and Lomask, Vanguard, 43–47, 49–51; “Project Vanguard Report … June 1957,” 2–6, NHRC; [U.S. Army Ordnance/GE], “Hermes,”68, NASM, and note 34 below for thrust of original A3 –B propulsion system. See Hunley, Preludes, chap. 6, on the Stewart Committee.

(26.) Green and Lomask, Vanguard,54,262; M. Neufeld, “Orbiter,” 249. Rosen's memo was to the NRL director, but Green and Lomask make it clear the director forwarded it to the Stewart Committee.

(27.) Letter, Milton W. Rosen, Executive Secretary, Committee on Underground Coal Mine Safety, National Research Council, to Monte D. Wright, Director, NASA History Office, May7,1981, in “Rosen, Milton W.,” NHRC. Despite Rosen's comments, the original statement remained unchanged in the published volume, Ezell, NASA Historical Data Book2:86.

(28.) Rosen, comments, UP.

(29.) “Project Vanguard Report No.1,” 1(first quotation), NHRC; U.S. Congress, “Project Vanguard,” 60(second quotation), whose authors say they could not document the quoted intent but were assured by “a representative of the Secretary of Defense” that this was one objective of the project, as indeed it was.

(30.) U.S. Congress, “Project Vanguard,” 61; NASA bio, “John P. Hagen,” in “Vanguard II (Feb.17,1959 ),” NHRC.

(31.) U.S. Congress, “Project Vanguard,” 61; “Project Vanguard Report … 1 June 1957,” 2–6 to 2–7, NHRC; Stehling, Project Vanguard, 301. Mass fraction is defined in “Rocket Performance: Mass.” Cf. the discussion of these issues and others in Green and Lomask, Vanguard, 57–90. Stehling was a propulsion engineer whom Rosen lured from Bell Aircraft to be chief of propulsion at NRL. His is thus an insider's history, and he is a clear example of an engineer who contributed to technology transfer by moving from a commercial firm to a government project. His book (and Rosen's on Viking) also abetted technology transfer.

(32.) Hagen, “Viking and the Vanguard,” 127–28 (quotation, 128); Hall, “Vanguard and Explorer,” 109; Harwood, Raise Heaven and Earth, 286–87; Green and Lomask, Vanguard, 66–67; Stumpf, Titan II,15, for dates of Titan I letter and final contracts; Stehling, Project Vanguard, 64–65.

(33.) Hall, “Vanguard and Explorer,” 109; Green and Lomask, Vanguard, 62–68; Stehling, Project Vanguard, 64–66.

(34.) Early measurements of specific impulse were not standardized. [U.S. Army Ordnance/GE], “Hermes,”63, NASM, says the A3B (as it is rendered there) had a sea-level specific impulse of 242 lbf-sec/lbm, but Vanguard sources list it at.225 See, e.g., “Project Vanguard Report … 1 June 1957,” 2–25, NHRC.

(35.) Heppenheimer, Countdown, 72–73; Gibson, Navaho Missile,40; “Project Vanguard Report … 1 June 1957,” 2–34, NHRC; J. Clark, Ignition! 32–33, 104–5. C. William Schnare, who worked with Hall at Wright-Patterson AFB, told the author by telephone on March 14,2002, that he was involved in insisting that NAA convert to kerosene fuel because it was more readily available than alcohol and more energetic. Clark, a propellant chemist working for the Naval Air Rocket Test Station, Lake Denmark, N.J. (taken over by the Army in 1960 as the Liquid Rocket Propulsion Laboratory of Picatinny Arsenal), until his retirement in 1970, wrote:

A turbojet has a remarkably undiscriminating appetite, and will run, or can be made to run, on just about anything that will burn and can be made to flow, from coal dust to hydrogen. But the services decided, in setting up the specifications for … jet fuel …, that the most important consideration would be availability and ease of handling. So since petroleum was the most readily available source of thermal energy in the country, and since they had been handling petroleum products for years, and knew all about it, the services decided that jet fuel should be a petroleum derivative—a kerosene. … But the permitted fractions of aromatics and olefins [were] 25 and 5 percent respectively [in JP-4]. (32–33)

Aromatics and olefins were what caused the problems in rockets, and as Clark points out (105), the RP-1 specifications limited olefins to 1 percent and aromatics to 5 percent, solving the problems with kerosene use in rocket engines.

(36.) Martin, “Vanguard,” 4, suggests this interpretation is possible by saying of the “state of the art” in 1955, “Engines were under development, using hydrocarbon fuels, which would increase specific impulse to the 240- to 250-second level.” On the other hand, the report lists the fuel for the GE engine as jet fuel, suggesting that it probably was not RP-1.

(37.) “Project Vanguard Report … 1 June 1957,” 2–25 to 2–32, NHRC; Stehling, Project Vanguard, 129–30; Martin, “Vanguard,” 49–50, NHRC; Stehling, “Aspects of Vanguard Propulsion,” 45–46. In his comments of May 8,2002, Milton Rosen wrote, with regard to the burn-through problems, that “GE wanted to redesign the entire cooling system, which would have caused a six month delay in deliveries. Kurt [Stehling] recommended running a ¼ inch d[iameter] copper wire through the existing cooling passages, and that solved the problem.” Stehling does not discuss his own contribution.

(38.) “Project Vanguard Report … 1 June 1957,” 2–37, NHRC; “Vanguard Vehicle Characteristics,” n.d. [after December 16,1959 ], in “Vanguard Project, History,” NASM.

(39.) Red fuming nitric acid (RFNA) with 6.5 percent nitrogen dioxide plus aniline with the addition of 20 percent furfuryl alcohol.

(40.) With 35 percent furfuryl alcohol instead of 20 percent.

(41.) For WAC Corporal's propellants, see Hunley, Preludes, chap. 3; for Aerobee's, Newell, Sounding Rockets,63; for rest of last three paragraphs, J. Clark, Ignition! 21,26,41–45, 47–65.

(42.) “Project Vanguard Report … 1 June 1957,” 2–37 to 2–38 (quotation, 2–37 ), NHRC; “Project Vanguard Report No.9,”5, NHRC; Martin, “Vanguard,” 51–55, NHRC; Green and Lomask, Vanguard,89, 204; Stehling, Project Vanguard, 132–33; George E. Pelletier, Director of Public Relations, Aerojet General, “Details of the Aerojet-Gen-eral Second-Stage Propulsion System for Vanguard Launching Vehicle,” February 11,1959, in “Vanguard II (Feb.17,1959 ),” NHRC. The coolant for the aluminum tubes was IWFNA.

(43.) “Project Vanguard Report … 1 June 1957,” 2–41 to 2–42, NHRC; Green and Lomask, Vanguard,89; Martin Company, “Trip to Aerojet-General, Azusa, California, December 18–20,1956,” in NASA, Vanguard Division records, box 1, NA.

(44.) Green and Lomask, Vanguard, 204 (also first quotation); Stehling, Project Vanguard, 134–35 (second quotation, 135); and three items in NASA, Vanguard Division records, box 1, NA: memo, Director, U.S. Naval Research Laboratory, to Martin Company, July 25,1957, signed J. P. Hagen; memo, Martin Company to Director, Naval Research Laboratory, July 29,1957, signed N. E. Felt Jr.; memo, Director, U.S. Naval Research Laboratory to Bureau of Aeronautics Representative, Azusa, California, November 27, 1957, signed M. W. Rosen. Stehling refers to “burnouts” rather than erosion, but U.S. Congress, “Project Vanguard,” 62, specifically states that the problem was “gradual erosion (not burn-through)” and that the “problem was not solved until October.1957” In his comments of May 8,2002, Rosen pointed out that Stehling was the one who suggested “coating the inner side of the liner with a ceramic and he found the ceramic we used,” but as with the first-stage burnout problem (see note 37), Stehling is silent about his own contribution in this regard.

(45.) Stehling, Project Vanguard, 134 (first quotation); Rosen, “Rocket Development (p.342) for Project Vanguard” (undated, annotated “Dod approval 23 Jan 57”), 5 (second quotation), in “Rosen, Milton W.,” NHRC.

(46.) Bartley and Bramscher, “Grand Central Rocket Company,” 267–68,271, 273–76; Bartley, letters to American Heritage of Invention and Technology and to Hunley, UP. See also Bartley, OHI, passim. In the article and the paper from which it was derived, Bartley spelled Settlemire as Settlemine, but in the letter to Hunley and in the OHI, it is spelled with an r.

(47.) “Project Vanguard Report … 1 June 1957,” 2–45 to 2–49, NHRC; Bartley and Bramscher, “Grand Central Rocket Company,” 273–74; Green and Lomask, Vanguard, 287. On cracking problems in the Sergeant, see Hunley, Preludes.

(48.) Hall, “Vanguard and Explorer,” 111; Moore, “Solid Rocket Development,” 5–6; Hunley, “Evolution,” 27–28 and sources cited there, including Dyer and Sicilia, Modern Hercules, 2, 9, 257–58, and Dembrow, OHI. Dembrow, who worked at ABL, said Kincaid and Shuey had relocated from ERL to ABL and done the actual development work there. On this, see also Noyes, Chemistry,17, 26–33, 127–28. My understanding of the cast double-base process has been greatly expanded by e-mails from Ronald L. Simmons, who first became familiar with the propellants used in Hercules' upper stage for Vanguard in 1958. He calls the cast double-base process the “casting powder/solvent process.” See esp. his e-mail of July 9, 2002, and résumé, UP.

(49.) “Project Vanguard Report … 1 June 1957,” 2–50, NHRC; “Project Vanguard Report No.9,”7, NHRC; report, A. H. Kitzmiller and E. J. Skurzynski, Hercules Powder Company, to R. Winer, September 17, 1956, subject: MPR—JATO Unit X241 A (Project Vanguard) Problem 4–a-81, in NASA, Vanguard Division records, box 6, NA.

(50.) Hunley, “Evolution,” 28; Dyer and Sicilia, Modern Hercules, 9, 318–20; B. Wilson, “Composite Motor Case Design”; “Lightweight Pressure Vessels,” portion of Ritchey, “Technical Memoir,” UP, kindly provided by Ernie Sutton.

(51.) “Project Vanguard Report … 1 June 1957,” 2–52, NHRC.

(52.) For the presence of the 3 percent of aluminum, ABL Report 40, “JATO X248 A2, A Solid Propellant Thrust Unit with High Impulse, High Performance, Wide Applications,” November 1958, 7, in NASA, Vanguard Division records, box 6, NA; on the effect of aluminum on combustion instability, see e.g., Povinelli, “Particulate Damping,” 1791–96. For the rest of the paragraph, memo, Code 4120 to Code 4100, August 13, 1958, subject: Summary of Conference at ABL and Recommendations, in NASA, Vanguard Division records, box 6, NA; Green and Lomask, Vanguard, 287; Holmes, “ABL's Altair,” 29. After I wrote this account, Ronald L. Simmons sent me (e-mails of July 10 and 12, 2002, UP) a breakdown of the X248 propellant, which included almost 39 percent nitrocellulose, almost 43 percent nitroglycerine, and almost 12 percent triacetin (a diluting agent) plus the aluminum mentioned in the narrative and lesser ingredients to serve as stabilizers, etc.

(53.) Attachment to a letter, Director, U.S. Naval Research Laboratory to Commander, Arnold Engineering and Development Center, September 12, 1958, in NASA, Vanguard Division records, box 2, NA; ABL Report 40 (see note 52), which reports the specific impulse in terms of lbf-sec/lbw, the w indicating weight rather than mass, but I have substituted the more normal designation lbf-sec/lbm; Stehling, Project Vanguard, 128–29n. (p.343) Incidentally, Stehling notes that the Stewart Committee had to approve use of the improved X248 motor instead of the X241.

(54.) The Reynolds number is a nondimensional parameter representing, roughly, the ratio of momentum forces to the viscosity of the fluid through which a body is passing, taking into account representative length; among other uses, the ratio is vital for scale-model testing in wind tunnels, as it provides a basis for extrapolating the test data to full-sized vehicles. For a fuller, more technical definition, see Braslow, Suction-Type Laminar-Flow Control, 2; for further discussion, see Hansen, Engineer in Charge, 69, 72, 74–76, 318. It was named after Osborne Reynolds (1814–1912) ofthe University of Manchester who, in Hansen's words, “identified this crucial scaling parameter” (69).

(55.) “Project Vanguard Report … 1 June 1957,” 2–13 to 2–15, 2–102, NHRC; Hagen, “Viking and the Vanguard,”130; Martin, “Vanguard,”18, NHRC.

(56.) “Project Vanguard Report … 1 June 1957,” 2–103 to 2–104, NHRC.

(57.) Harwood, Raise Heaven and Earth, 254; Furth, “Vanguard,” 3.

(58.) Dominic Edelin, Guidance Control Division, Martin Company, “Separation, Stabilization Final Report X208377,” vol. 3, September 1958, 1 (quotation, my italics), and “Systems and Components,” vol. 1, August 1, 1958, 1, both in NASA, Vanguard Division records, box 5, NA; Freeman, “Vanguard Control System,” 2, 4, NASM.

(59.) Freeman, “Vanguard Control System,” 5–11, NASM.

(60.) “Project Vanguard Report … 1 June 1957,” 2–58 to 2–65, 2–69 to 2–75, 2–95, NHRC; Minneapolis-Honeywell Regulator Company, “Vanguard Guidance Sidebar,” news release, January 23, 1958, in “Vanguard II (Feb. 17, 1959),” NHRC; Steier, “What Guides the Vanguard,” 70, 72; “In the Vanguard.” MacKenzie, Inventing Accuracy, 182, is the source for the use of strapdown systems in conjunction with small but powerful digital computers; Martin, “Vanguard,” 175, NHRC, is the source for their use by 1960 in “many ballistic missile systems.” Neither source is specific about which particular rockets or missiles used the strapdown systems. The Minneapolis-Honeywell platform weighed 30 pounds, 70 less than any other system. Yet the Martin report says it provided “accuracy better than required” (29). Steam exhausting from the turbopump in stage provided the thrust in the roll jets for that stage. Stage 2 had no turbopumps, so pressurized gases provided the thrust for its roll jets.

(61.) Designers for Industry, “Designers for Industry Produced Guidance Subsystem for Vanguard Rocket,” news release, n.d., in “Vanguard II (Feb. 17, 1959),” NHRC; “Project Vanguard Report … 1 June 1957,” 2–15, 2–17, 2–58, 2–63, NHRC; Green and Lomask, Vanguard, 59, schematic of Vanguard trajectory; “In the Vanguard”; Steier, “What Guides the Vanguard.”

(62.) “Project Vanguard Report … 1 June 1957,” 2–77 to 2–88, NHRC; Stehling, Project Vanguard, 266. The integrating accelerometer used the same type of gyro as the Minneapolis-Honeywell platform, but in a pendulous assembly. See Martin, “Vanguard,” 34, NHRC.

(63.) U.S. Congress, “Project Vanguard,” 62, 65–66; Green and Lomask, Vanguard, 176, 283; Stehling, Project Vanguard, 82–83.

(64.) Green and Lomask, Vanguard, 177–82,; Stehling, Project Vanguard, 106–22, (p.344) with many details of problems encountered at Cape Canaveral. For comparative coverage of U.S. and Soviet space-launch efforts in this period, see Bille and Lishock, First Space Race; for more detailed coverage of Soviet developments, Siddiqi, Challenge to Apollo.

(65.) Green and Lomask, Vanguard, 196–98; Stehling, Project Vanguard, 103, 123; satellite weight from “Space Activities Summary, Vanguard,” December 6, 1957, in “Vanguard Test Vehicle 3,” NHRC; substitution of small satellite for instrumented nose cone from Martin, “Vanguard,” 4, NHRC.

(66.) Green and Lomask, Vanguard, 283; “Preliminary Report on TV-3,” December 18, 1957, 1 (first quotation), in “Vanguard Test Vehicle 3,” NHRC; Stehling, Project Vanguard, 24 (second quotation); McDougall, Heavens and the Earth, (last two quotations); Heppenheimer, Countdown, 127, quoting the London Daily Herald for Flopnik and the London Daily Express for Kaputnik. McDougall gives no sources for Stayputnik or the other two wordplays on Sputnik.

(67.) Green and Lomask, Vanguard, 210; Rosen, comments, May 16, 2002, UP.

(68.) Green and Lomask, Vanguard, 213–14, 217, 283; Stehling, Project Vanguard, 157–81, with details of launch efforts leading up to the failed mission.

(69.) Martin Information Services, “Project Vanguard,” [November 23, 1959 ], in “Vanguard Project, History,” NASM; Green and Lomask, Vanguard, 204, 219, 285; Stehling, Project Vanguard, 156, 182–222, 274.

(70.) NRL, “Transfer,” NHRC; Green and Lomask, Vanguard, 223; Stehling, Project Vanguard, 237–38; NASA Management Instruction.1052.1 with Attachment A (Executive Order 10783) and Attachment B (Agreement between DoD and NASA), October 1, 1958, effective date October 13, 1959, in “Vanguard Satellite Launching Vehicle 3,” NASM; Martin, “Vanguard,” 96–98, NHRC. Although President Eisenhower's executive order was dated October 1, NRL notice 5400 says personnel did not transfer to NASA until November 30. Harry Goett headed the NASA committee, which, according to the Martin report, made valuable suggestions but found “no major deficiencies in the rocket design” (98).

(71.) On these two points, see Rosen, “Viking and Vanguard,” 4–5, NHRC.

(72.) See U.S. Congress, “Project Vanguard,” 67–80; Green and Lomask, Vanguard, 253–54; Rosen, “What Have We Learned from Vanguard?”

(73.) Rosen, “Brief History,” NHRC; Stehling, Project Vanguard, 233–34; Hagen, “Viking and the Vanguard,” 139.

(74.) Rosen, “Brief History,” NHRC; Hagen, “Viking and the Vanguard,” 139; Green and Lomask, Vanguard, 254–55. For a Silverstein bio, see Levine, Managing NASA, 30.

(75.) Hagen, “Viking and the Vanguard,” 140; CPIA/M, unit 412, Minuteman Stage 3 Wing I, January 1964, and unit 427, Altair I, March 1964; Bedard, “Composite Solid Propellants,”.9 Thanks to CPIA for permission to cite the manual and to Tom Moore of that agency for pointing out the Altair connection. Thanks too to Steve Benson for pointing out Bedard's article online.

(76.) “Vanguard, Historical Significance,” n.d., in “Vanguard Project, History,” NASM. Cf. Hagen, “Viking and the Vanguard,” 132, 135, 139–40.