The Atlas, Thor, and Jupiter Missiles

Following Redstone and Vanguard, the Atlas, Thor, and Jupiter mis­siles brought further innovations in rocket technology and became the first stages of launch vehicles themselves, with Atlas and Thor having more significance in this role than Jupiter. All three pro­grams illustrated the roles of interservice and interagency rivalry and cooperation that were both key features of rocket development in the United States. They also showed the continued use of both theory and empiricism in the complex engineering of rocket sys­tems. “It was not one important ‘breakthrough’ that enabled this advance; rather, it was a thousand different refinements, a hundred thousand tests and design modifications, all aimed at the develop­ment of equipment of extraordinary power and reliability," accord­ing to Milton Rosen, writing in 1962.61

Atlas was a much larger effort than Vanguard, and it began to create the infrastructure in talent, knowledge, data, and capability necessary for the maturation of launch-vehicle technology in the decade of the 1960s. However, until the air force became serious about Atlas, that service had lagged behind the army and the navy in the development of purely ballistic missiles.62

The process began in a significant way on January 23, 1951, when 32 the air force awarded the Consolidated Vultee Aircraft Corporation Chapter 1 (Convair) a contract for MX-1593, the project that soon became Atlas. (MX-1593 had been preceded by MX-774B and a number of other air force missile contracts in the late 1940s, with a total of $34 million devoted to missile research in fiscal year 1946, much re-

duced in subsequent years.) But the specifications for the MX-1593 missile changed drastically as technology for nuclear warheads evolved to fit more explosive power into smaller packages. This new technology plus the increased threat from the Soviets provided one condition for greater air force support of Atlas and other bal­listic missiles.

But it also took two heterogeneous engineers to nudge the new­est armed service and the Department of Defense (DoD) in a new direction. One of them was Trevor Gardner, assistant for research and development to Secretary of the Air Force Harold Talbott in the Eisenhower administration. The other key promoter of ballistic missiles was the “brilliant and affable" polymath, John von Neu­mann, who was research professor of mathematics at Princeton’s Institute for Advanced Study and also director of its electronic com­puter project. In 1953, he headed a Nuclear Weapons Panel of the Air Force Scientific Advisory Board, which confirmed beliefs that in the next six to eight years, the United States would have the capability to field a thermonuclear warhead weighing about 1,500 pounds and yielding 1 megaton of explosive force. This was 50 times the yield of the atomic warhead originally planned for the Atlas missile, fit in a much lighter package. This and a report (dated February 1, 1954) for von Neumann’s Teapot Committee set the stage for extraordi­nary air force support for Atlas.63

In May 1954 the air force directed that the Atlas program be­gin an accelerated development schedule, using the service’s high­est priority. The Air Research and Development Command within the air arm created a new organization in Inglewood, California, named the Western Development Division (WDD), and placed Brig. Gen. Bernard A. Schriever in charge. Schriever, who was born in Germany but moved to Texas when his father became a prisoner of war there during World War I, graduated from the Agricultural and Mechanical College of Texas (since 1964, Texas A&M Univer­sity) in 1931 with a degree in architectural engineering. Tall, slen­der, and handsome, the determined young man accepted a reserve commission in the army and completed pilot training, eventu­ally marrying the daughter of Brig. Gen. George Brett of the army air corps in 1938. Placed in charge of the WDD, Schriever in es­sence took over from Gardner and von Neumann the role of het­erogeneous engineer, promoting and developing the Atlas and later missiles.64

Gardner had been intense and abrasive in pushing the develop­ment of missiles. Schriever was generally calm and persuasive. He selected highly competent people for his staff, many of them be-

coming general officers. An extremely hard worker, like von Braun he demanded much of his staff; but unlike von Braun he seemed somewhat aloof to most of them and inconsiderate of their time— frequently late for meetings without even realizing it. Good at plan­ning and organizing, gifted with vision, he was poor at management, often overlooking matters that needed his attention—not surprising because he spent much time flying back and forth to Washington, D. C. His secretary and program managers had to watch carefully over key documents to ensure that he saw and responded to them. One of his early staff members (later a lieutenant general), Otto J. Glasser, said Schriever was “probably the keenest planner of any­body I ever met" but he was “one of the lousiest managers."

Later Lt. Gen. Charles H. Terhune Jr., who became Schriever’s deputy director for technical operations, called his boss a “superb front man" for the organization, “very convincing. . . . He had a lot of people working for him [who] were very good and did their jobs, but Schriever was the one who pulled it all together and represented them in Congress and other places." Glasser added, “He was just superb at. . . laying out the wisdom of his approach so that the Congress wanted to ladle out money to him." Glasser also said he was good at building camaraderie among his staff.65

To facilitate missile development, Schriever received from the air force unusual prerogatives, such as the Gillette Procedures. Designed by Hyde Gillette, a budgetary expert in the office of the secretary of the air force, these served to simplify procedures for managing intercontinental ballistic missiles (ICBMs). Schriever had complained that there were 40 different offices and agencies he had to deal with to get his job done. Approval of his annual develop­ment plans took months to sail through all of these bodies. With the new procedures (granted November 8, 1955), Schriever had to deal with only two ballistic missile committees, one at the secretary – of-defense and the other at the air-force level. Coupled with other arrangements, this gave Schriever unprecedented authority to de­velop missiles.66

Another key element in the management of the ballistic mis­sile effort was the Ramo-Wooldridge Corporation. Simon Ramo and Dean Wooldridge, classmates at Caltech, each had earned a Ph. D. there at age 23. After World War II, they had presided over an 34 electronics team that built fire-control systems for the air force at Chapter 1 Hughes Aircraft. In 1953, they set up their own corporation, with the Thompson Products firm buying 49 percent of the stock. For a variety of reasons, including recommendations of the Teapot Com­mittee, Schriever made Ramo-Wooldridge into a systems engineer-

ing-technical direction contractor to advise his staff on the man­agement of the Atlas program. The air force issued a contract to the firm for this task on January 29, 1955, although it had begun working in May 1954 under letter contract on a study of how to redirect the Atlas program. This unique arrangement with Ramo- Wooldridge caused considerable concern in the industry (especially on the part of Convair) that Ramo-Wooldridge employees would be in an unfair position to use the knowledge they gained to bid on other contracts, although the firm was not supposed to produce hardware for missiles. To ward off such criticism, the firm created a Guided Missile Research Division (GMRD) and kept it separate from other divisions of the firm. Louis Dunn, who had served on the Teapot Committee, became the GMRD director, bringing sev­eral people with him from JPL. This arrangement did not put an end to controversy about Ramo-Wooldridge’s role, so in 1957, GMRD became Space Technology Laboratories (STL), an autonomous divi­sion of the firm, with Ramo as president and Dunn as executive vice president and general manager.67

Some air force officers on Schriever’s staff objected to the con­tract with Ramo-Wooldridge, notably Col. Edward Hall, a propul­sion expert. Hall had nothing good to say about Ramo-Wooldridge (or Schriever), but several engineers at Convair concluded that the firm made a positive contribution to Atlas development.68

The Ramo-Wooldridge staff outnumbered the air force staff at WDD, but the two groups worked together in selecting contrac­tors for components of Atlas and later missiles, overseeing their performance, testing, and analyzing results. For such a large under­taking as Atlas, soon joined by other programs, there needed to be some system to inform managers and allow them to make decisions on problem areas. The WDD, which became the Air Force Ballistic Missile Division on June 1, 1957, developed a management control system to collect information for planning and scheduling.

Schriever and his program directors gathered all of this data in a program control room, located in a concrete vault and kept under guard at all times. At first, hundreds of charts and graphs covered the walls, but WDD soon added digital computers for tracking infor­mation. Although some staff members claimed Schriever used the control room only to impress important visitors, program managers benefited from preparing weekly and monthly reports of status, be­cause they had to verify their accuracy and thereby keep abreast of events. Separate reports from a procurement office the Air Force Air Materiel Command assigned to the WDD on August 15, 1954, pro­vided Schriever an independent check on information from his own

managers. The thousands of milestones—Schriever called them inchstones—in the master schedule kept him and his key manag­ers advised of how development matched planning. All of the infor­mation came together on “Black Saturday" meetings once a month starting in 1955. Here program managers and department heads presented problem areas to Schriever, Ramo, and Brig. Gen. Ben I. Funk, commander of the procurement office. As problems arose, discussion sometimes could resolve them in the course of the meet­ing. If not, a specific person or organization would be assigned to come up with a solution, while the staff of the program control room tracked progress. Sometimes, Ramo brought in outside experts from industry or academia to deal with particularly difficult problems.69

Because the process of developing new missile systems entailed considerable urgency when the Soviet threat was perceived as great and the technology was still far from mature, Schriever and his team used a practice called concurrency that was not new but not routinely practiced in the federal government. Used on the B-29 bomber, the Manhattan Project, and development of nuclear vessels for the U. S. Navy, it involved developing all subsystems and the facilities to test and manufacture them on overlapping schedules; likewise, the systems for operational control and the training sys­tem for the Strategic Air Command, which took over the missiles when they became operational.

Schriever claimed that implementing concurrency was equiva­lent to requiring a car manufacturer to build the automobile and also to construct highways, bridges, and filling stations as well as teach drivers’ education. He argued that concurrency saved money, but this seems doubtful. Each model of the Atlas missile from A to F involved expensive improvements, and the F models were housed in silos. Each time the F-model design changed, the Army Corps of Engineers had to reconfigure the silo. There were 199 engineering change orders for the silos near Lincoln, Nebraska, and these raised the costs from $23 million to more than $50 million dollars—to give one example of costs added by concurrency. What concurrency did achieve was speed of overall development and the assurance that all systems would be available on schedule.70

A further tool in WDD’s management portfolio was parallel de­velopment. To avoid being dependent on a single supplier for a sys – 36 tem, Schriever insisted on parallel contractors for many of them.

Chapter 1 Eventually, when Thor and Titan I came along, the testing program became overwhelming, and Glasser argued that Ramo-Wooldridge just ignored the problem. He went to Schriever, who directed him to come up with a solution. He decided which systems would go on

Atlas, which on Titan and Thor, in the process becoming the deputy for systems management and the Atlas project manager.71

A final component of the management structure for Schriever’s west-coast operation consisted of the nonprofit Aerospace Corpora­tion. It had come into existence on June 4, 1960, as a solution to the problems many people saw in Space Technology Laboratories’ serv­ing as a systems-engineering and technical-direction contractor to the air force while part of Thompson Ramo Wooldridge (later, just TRW), as the company had become following an eventual merger of Ramo-Wooldridge with Thompson Products. STL continued its operations for programs then in existence, but many of its person­nel transferred to the Aerospace Corporation for systems engineer­ing and technical direction of new programs. Further complicat­ing the picture, a reorganization occurred within the air force on April 1, 1961, in which Air Force Systems Command (AFSC) re­placed the Air Research and Development Command. On the same date, within AFSC, the Ballistic Missile Division split into a Ballis­tic Systems Division (BSD), which would retain responsibility for ballistic missiles (and would soon move to Norton Air Force Base [AFB] east of Los Angeles near San Bernardino); and a Space Systems Division (SSD), which moved to El Segundo, much closer to Los Angeles, and obtained responsibility for military space systems and boosters. There would be further reorganizations of the two offices, but whether combined or separated, they oversaw the development of a variety of missiles and launch vehicles, ranging from the Atlas and Thor to Titans I through IV.72

To return specifically to the Atlas program, under the earlier (1946-48) MX-774B project, Convair had developed swiveling of en­gines (a precursor of gimballing); monocoque propellant tanks that were integral to the structure of the rockets and pressurized with nitrogen to provide structural strength with very little weight pen­alty (later evolving into what Convair called a steel balloon); and separable nose cones so that the missile itself did not have to travel with a warhead to the target and thus have to survive the aerody­namic heating from reentering the atmosphere.73

Other innovations followed under the genial leadership of Karel (Charlie) Bossart. Finally, on January 6, 1955, the air force awarded a contract to Convair for the development and production of the Atlas airframe, the integration of other subsystems with the airframe and one another, their assembly and testing. The contractor for the At­las engines was North American Aviation, which built upon earlier research done on the Navaho missile. NAA’s Rocketdyne Division, formed in 1955 to handle the requirements of Navaho, Atlas, and

Redstone, developed one sustainer and two outside booster engines for the Atlas under a so-called stage-and-a-half arrangement, with the boosters discarded after they had done their work. Produced in

1957 and 1958, the early engines ran into failures of systems and components in flight testing that also plagued the Thor and Jupiter engines, which were under simultaneous development and shared many component designs with the Atlas.74

But innovation continued, partly through engineers making “the right guess or assumption" or simply learning from problems. De­spite repeated failures and (trial-and-error) modifications to elimi­nate their causes, development proceeded from Atlas A through At­las F with a total of 158 successful launches for all models against 69 failures—a success rate of only 69.6 percent. The Atlas D became the first operational version in September 1959, with the first E and F models following in 1961. All three remained operational until 1965, when they were phased out of the missile inventory, with many of them later becoming launch-vehicle stages.75

Meanwhile, fearing (unnecessarily) that an ICBM like the Atlas could not be deployed before 1962, a Technology Capabilities Panel headed by James R. Killian Jr., president of MIT, issued a report in mid-February 1955 recommending the development of both sea – and land-based intermediate-range ballistic missiles (IRBMs). In Novem­ber 1955, the Joint Chiefs of Staff recommended, in turn, that the air force develop the land-based version while the army and navy collaborate on an IRBM that could be both land and sea based. Thus were born the air force’s Thor and the army’s Jupiter, with the navy eventually developing the solid-propellant Polaris after initially try­ing to adapt the liquid-propellant Jupiter to shipboard use.76

Arising out of this decision was the “Thor-Jupiter Controversy," which the House of Representatives Committee on Government Operations called a “case study in interservice rivalry." The Thor did not use the extremely light, steel-balloon structure of Atlas but a more conventional aluminum airframe. Its main engine consisted essentially of half of the booster system for Atlas. In 1957 and 1958, it experienced 12 failures or partial successes out of the first 18 launches. Before the air force nevertheless decided in September

1958 that Thor was ready for operational deployment, problems with the turbopumps (common to the Atlas, Thor, and Jupiter) and

38 differences of approach to these problems had led to disagreement Chapter 1 between the Thor and Jupiter teams.77

Von Braun’s engineers, working on the Jupiter for the army, di­agnosed the problem first and had Rocketdyne design a bearing re­tainer for the turbopump that solved the problem, which the Thor

program would not admit at first, suspecting another cause. Once the Jupiters resumed test flights, they had no further turbopump problems. Meanwhile, failures of an Atlas and a Thor missile in April 1958, plus subsequent analysis, led the air force belatedly to accept the army’s diagnosis and a turbopump redesign. The first Thor squadron went on operational alert in Great Britain in June 1959, with three others following by April 1960. When the Atlas and Titan ICBMs achieved operational readiness in 1960, the last Thors could be removed from operational status in 1963, making them available for space-launch activities.78

While the Western Development Division and the successor Air Force Ballistic Missile Division were developing the Thor in con­junction with contractors, von Braun’s group at what had become the Army Ballistic Missile Agency (ABMA) in Alabama and its con­tractors were busily at work on Jupiter without a clear indication whether the army or the air force would eventually deploy the mis­sile. At ABMA, the forceful and dynamic Maj. Gen. John B. Medaris enjoyed powers of initiative roughly analogous to those of Schriever for the air force. On December 8, 1956, the navy left the Jupiter program to develop Polaris, but not before the sea service’s require­ments had altered the shape of the army missile to a much shorter and somewhat thicker contour than the army had planned. With Chrysler the prime contractor (as on the Redstone), Medaris reluc­tantly accepted the same basic engine North American Rocketdyne was developing for the Thor except that the Jupiter engine evolved from an earlier version of the powerplant and ended as somewhat less powerful than the air force counterparts.79

With a quite different vernier engine and guidance/control sys­tem, the Jupiter was a decidedly distinct missile from the Thor. The first actual Jupiter (as distinguished from the Jupiter A and Jupiter C, which were actually Redstones) launched on March 1, 1957, at Cape Canaveral. Facing the usual developmental problems, including at least one that Medaris blamed on the thicker shape resulting from the navy’s requirements, the Jupiter nevertheless achieved 22 sat­isfactory research-and-development flight tests out of 29 attempts. The air force, instead of the army, deployed the missile, with initial operational capability coming on October 20, 1960. Two squadrons of the missile became fully operational in Italy as of June 20, 1961. A third squadron in Turkey was not operational until 1962, with all of the missiles taken out of service in April of the following year. Three feet shorter, slightly thicker and heavier, the Jupiter was more accurate but less powerful than the Thor, with a comparable range. The greater average thrust of the Thor may have contributed

to its becoming a standard first-stage launch vehicle, whereas Jupi­ter served in that capacity to only a limited degree. Another factor may have been that there were 160 production Thors to only 60 Ju­piter missiles.80

Although much has rightly been made of the intense interservice rivalry between the army and the air force over Thor and Jupiter, even those two programs cooperated to a considerable extent and exchanged much data. Medaris complained about the lack of infor­mation he received from the air force, but Schriever claimed that his Ballistic Missile Division had transmitted to ABMA a total of 4,476 documents between 1954 and February 1959. By his count, BMD withheld only 28 documents for a variety of reasons, including con­tractors’ proprietary information.81 This was one of many examples showing that—although interservice and interagency rivalry helped encourage competing engineers to excel—without sharing of infor­mation and technology, rocketry might have advanced much less quickly than in fact it did.