The+ENIAC+&+Nuclear+Weapon+Proliferation

toc Throughout history, the defense of America has continuously ushered in new eras of scientific achievement. This is no more evident than in the evolution of the computer age. The eruption of war in Europe in 1941 prompted the U.S. Army Ordnance Department, a military establishment that provided scientific and logical support following World War I, investigation into augmenting the calculation of artillery firing tables. The Electronic Numerical Integrator and Computer (ENIAC), the world’s first digital computer, was designed during World War II at the University of Pennsylvania’s Moore School of Electrical Engineering. The commission was the ENIAC was indicative of the emergent interdependency between technology and the proliferation of nuclear warfare. Without its intricate technological support system such as nuclear material reactors and metals fabrication facilities, the Manhattan Project perhaps would not have produced an atomic device in the short period of three years. Subsequently, computing would affect the development of the hydrogen bomb.

This analysis explores the interrelationship between digital computing and the proliferation of the hydrogen bomb. This is attributed to the manner in which the development of calculation systems affected the pace of knowledge production.

=**The Escalating Problem**=

In 1938, the recovery from the Great Depression had left the United States focusing primary on domestic problems. Such introspection cultivated obliviousness within the United States to the events occurring internationally. This attitude was exemplified in the relative apathy towards military preparedness. However, as conflicts intensified in Europe, the U.S. involvement in the Second World War became more plausible. Thus a rigorous mobilization of firepower was commenced.

This effort was centralized at the Aberdeen Proving Ground in Maryland, where the U.S. Army Ordinance Department tested weapons (Goldstine, 1972). The testing of large guns required the calculation of trajectory tables to determine ballistic accuracy. These tables illustrated the horizontal distance traveled by shells for a given inclination (Reed, personal communication, 1996). Since multiple factors such as the gun angle, wind speed and direction, temperature, atmospheric pressure, and the types of guns and projectiles were accounted; the computation of such tables required three-dimensional second order differential equations (Reed, personal communication, 1996). Consequently, the tables were primarily computed with individual hand calculators where the calculation of a single trajectory often took upward of 40 hours to complete (Weik, 1961). When the vast complexity of the calculation was coupled with the fact that the Ballistic Research Laboratory (BRL) was inadequately staffed; a more efficient method had to be devised. = = =**ENIAC Predecessors**=

The Bush Differential Analyzer
An electromechanical computing device called the Bush Differential Analyzer had been installed at Aberdeen in 1935 (Weik, 1961). Dr. Vannevar Bush of the Massachusetts Institute of Technology designed the differential analyzer in 1925 (Bush, 1970). Although this device reduced trajectory calculation time to around 30 minutes, it presented a share of logical impasses (Goldstine, 1972). Foremost, for each problem solved, many of the mechanical parts had to be “retrofitted” (Goldstine, 1972). This process proved to be overwhelmingly time consuming. Though the most serious problem was the torque amplifier, an important element to maintaining high computational speed, repeated failure (Kempf, 1961). These malfunctions often occurring at the conclusion of lengthy trajectory runs resulted in the loss of the previous calculations and substantial stoppage associated with repair (Kempf, 1961). This down time proved costly.

In June of 1942, Lieutenant P. N. Gillon, supervisor of the ballistic calculations at Aberdeen, made a contract with Dean Harold Pender and Professor J. G. Brainard of the Moore School of Electrical Engineering at the University of Pennsylvania (Weik, 1961). This agreement was for the exclusive usage of the school’s Bush Analyzer, which was larger and faster than that of the BRL, for the computation of trajectory tables (Kempf, 1961). Additionally, an association of talented scientist and engineers, including Dr. John W. Mauchly and Mr. J. Presper Eckert, had begun confronting Ordinance Department’s computational dilemmas (J. Eckert, personal communication, February 2, 1988). The initial improvements included accelerating the computational speed of the school’s differential analyzer. This modification was obtained via the replacement of the mechanical torque amplifier with an electric version (J. Eckert, personal communication, February 2, 1988). Though these improvements made the differential analyzer much faster and more reliable; it remained too insufficient in meeting the demands of the Ordinance Department as Lt. Herman H. Goldstein, liaison to Moore School, conveyed in a memo to his commanding officer: “In addition to a staff of 176 computers in the Computing Branch, the Laboratory has 10-integrator differential analyzer at Aberdeen and 14- integrator one at Philadelphia, as well as a large group of IBM machines. Even with the personnel and equipment now available, it takes about three months if work on a two-shift basis to turn out the data needed to construct 2 director, gun sight, or firing table” (Campbell-Kelly & Aspray, 2004).

The ABC Computer
In 1937, John Atanasoff of Iowa State devised the schematics for an electrically powered box that could solve equations through binary math (Mollenhoff, 1988). With the assistance of grad student Clifford Berry and research grant funding, Atanasoff completed a prototype of the ABC computer that he demonstrated in October 1939 (Mollenhoff, 1988). Atanasoff progressed to develop an enhanced version, which included 300 vacuum tubes and completed calculations in several seconds, in 1941 (Mollenhoff, 1988). However, further work on this project was abandoned to focus on more urgent defense projects after the outbreak of war (Mollenhoff, 1988).

The extent to which Atanasoff’s ideas were incorporated in the ENIAC became a contentious debate. In December 1940, while Dr. John W. Mauchly was an instructor at Ursinus College, Atanasoff attended a lecture Mauchly presented on analog computing machines (Campbell-Kelly & Aspray, 2004). As similar research interest was shared; Atanasoff invited Mauchly to see his ABC prototype. Though the ABC was unexceptional technology; at the least it can be inferred that Mauchly saw the significant of the ABC (Campbell-Kelly & Aspray, 2004). Thus, he was inclined to propose a similar solution to the computational problems at the BRL.

=**The ENIAC**=

The Design Process
By August 1942, Dr. John W. Mauchly had cultivated his vision for an electronic computer such that he prepared a memorandum on The Use of High Speed Vacuum Tubes for Calculating nevertheless it was unanimously dismissed (Campbell-Kelly & Aspray, 2004). However, in the early part of 1943, Professor J. G. Brainard and Captain Herman H. Goldstein (promoted in the meantime) brought to Colonel P. N. Gillon (promoted in 1942) an outline of the technical schematics underlying the design of an electronic computer (Kempf, 1961). Although the successful completion of such an endeavor remained highly speculative, an agreement was made between the United States of America and the trustees for the University of Pennsylvania for six month of “research and development for an electronic numerical integrator and computer” dated June 5, 1943 (Weik, 1961). The U.S. Army Ordnance entrusted an investment of $61,700 under appropriation number ORD 61166 P610-07-A1005-23 in the initial contract R.A.D. 1078-W-670-ORD-4926 (Kempf, 1961). Nevertheless, nine contract supplements extended the work period into 1946 and increased the expenditure to $486,804.22 (Kempf, 1961).

As of May 31, 1943, the rigorous pursuant of implementation began at the Moore School under the supervision of Professor Brainard, with Mr. Eckert serving as the chief engineer and Dr. Mauchly as principal consultant (Kempf, 1961). As of 1944, there was a conscious that the orders for the machine could be stored in numerical form on tape. This strategy was based on a similar method used in the Babbage’s Analytical Engine (Goldstine, 1972). As a result, there was growing optimism that the project would be at completion by January 1, 1945. However, as the construction of the ENIAC progressed, deficiencies with the initial design became apparent. These were primarily related to programmability. Machines such as the Harvard Mark I could be programmed using punch card because the slow computational speed made this a feasible process (Goldstine, 1972). This was not the case for the ENIAC. Because the ENIAC performed 5,000 operations per second, card were infeasible (Goldstine, 1972). Eckert and Mauchly decided that their machine would have to be specially wired for each specific problem. The ENAIC design took 18 months (Goldstine, 1972). Subsequently, it was constructed component by component, starting with the cycling unit and accumulator in June 1944 (Weik, 1961). The initiating unit and function table in September 1945 and the divider and square-root unit in October 1945 soon followed (Weik, 1961). This project was officially concluded with a dedication at the Moore School of Electrical Engineering on February 15, 1946 (Weik, 1961).

Technical Specification
In comparison to modern computers, the ENIAC was a monstrosity. The thirty separate units occupied 1,800 square feet, exceed a weight of 30 tons, and consumed 160 kilowatts of electrical power. The central element consisted of 42 panels with each about 9 feet high, 2 feet wide, and 1 foot thick (Kempf, 1961). It was comprised of 17,468 vacuum tubes, 70,000 resistors, 10,000 capacitors, 1,500 relays, 6,000 manual switches, and 5 million soldered joints (Goldstine, 1972). Such convectional electronic circuits and elements made this design optimal for war circumstances. Because of the inadequacies of pervious machines, reliability was a primary objective for the ENIAC. Eckert ensured reliability via a rigorous testing procedure on all components (Campbell-Kelly & Aspray, 2004). Whole boxes of resistors were obtained and their tolerance was tested individually on a specialized rig (Campbell-Kelly & Aspray, 2004). Only optimal resistors were used in the most critical parts of the machine (Campbell-Kelly & Aspray, 2004). On occasions, entire boxes of resistors were rejected and shipped back to the manufacturer (Campbell-Kelly & Aspray, 2004). Ironically, the ENIAC would suffer its fair share of service interruptions (Weik, 1961). These were primarily a result of vacuum tube failure (Weik, 1961).

The ENIAC was the most sophisticated electronics system to date. It produced computations one thousand times fast than any other calculating machine of its time (J. Eckert, personal communication, February 2, 1988). In one second, the ENIAC may perhaps perform 5,000 arithmetic operations, 357 multiplications or 38 divisions (J. Eckert, personal communication, February 2, 1988).

von Neumann Modifications
By the summer of 1944, additional shortcomings were evident. Dr. John von Neumann of the Institute for Advance Study at Princeton realized while the machine was effective in solving ordinary differential equations; it would be of little use for solving his partial differential equations (Goldstine, 1972). This was a result of the storage capacity of twenty numbers being far too little (Goldstine, 1972). This problematic issue was the agenda for the first meeting between Eckert and von Neumann (Goldstine, 1972). Soon after von Neumann arrived, the stored-program concept was devised. Eckhart proposed the usage of a mercury delay-line as the alternative to electronic-tube storage (Goldstine, 1972). This innovation demanded a new and separate contract as to not hinder the progress of the ENIAC (Goldstine, 1972). This design would later become the EVAC (Goldstine, 1972). A supplementary problem was the laborious programming method. Programming new problems meant weeks of preparation. Dr. von Neumann proposed code selection via switches so that cable connections could remain fixed for standard trajectory problems (Goldstine, 1972). Finally, other programming difficulties arose from arithmetic and transfer operations occurring simultaneously (Weik, 1961). The devising of a converter code made available the storage of 600 two-decimal digit instructions for each function table (Weik, 1961). As a result, the ENIAC was converted in a serial instruction execution machine (Weik, 1961).

From the completion of the modifications in 1948, the ENIAC would led the computer field until 1952, serving as the main computational machine for the solution of the nations scientific problems (Weik, 1961). The ENIAC would operate successfully for a total of 80, 223 hours between 1946-1955 performing five thousand arithmetic operations per second (Weik, 1961). The ENIAC would perform an array of applications such as weather predictions, atomic energy calculations, cosmic ray studies, and wind tunnel design (Weik, 1961).

The Hydrogen Bomb
Because the ENIAC came post-World War II, the first program ran was for Los Alamos scientist. Edward Teller has proposed the necessity of research for a hydrogen bomb before the atomic bomb was completed. Teller immediately recognized the limitation of the widely used electro-mechanical computers (Smith, 2001).“The more complex calculations of the hydrogen-bomb simulation exceeded the capabilities of the punch-card machine operation” (MacKenzie, 1998). Hydrogen bomb calculations involved determining the behavior of deuterium-tritium system in response to various initial temperatures (Fitzpatrick, 1998). This came to be known as the “Super” problem (Fitzpatrick, 1998). Therefore, as the ENIAC construction neared completion in 1945, Stanley Frankel and Nicholas Metropolis of Los Alamos reached an agreement with von Neumann to use the new electronic computer for hydrogen bomb simulations (Goldstine, 1972). First versions of the thermonuclear calculations were run during a six-week period between December 1945 and January 1946 (MacKenzie, 1998). However, the memory limitation of the ENIAC meant some effect could not be considered (Goldstine, 1972). Thus deuterium-tritium behavior could not be predicted.

The Super problem complexity exceeded the ability of any technology available. Because the limitation of the ENIAC was realized, Teller proposed attention is shifted to enhancing the capabilities of high-speed electronic calculators (Fitzpatrick, 1998). Though, in 1946, the only large machine available to Los Alamos was the ENIAC (Bethe, 1982). As a result, Los Alamos begun working on its own electronic digital computer since the construction of large computers far from Los Alamos was developing leisurely (Fitzpatrick, 1998). However, behind schedule construction dashed optimism that a full simulation of the Super problem would be achieved by 1949 or 1950 (Rhodes, 1986). With Teller and colleagues under severe political pressure, a collection of ideas from Teller, Ulam, and Frederic DeHoffman composed a new thermonuclear system that appearing more plausible (Fitzpatrick, 1998). In choosing the Teller-Ulam configuration, nuclear scientist bypassed the hindrance of the Super problem (Fitzpatrick, 1998).

=Conclusion=

Bounded to the limitations of calculation machinery, weapons scientist amended research methods. The inadequacies of ENIAC retarded the progression of the hydrogen bomb project. Its inability to determine hydrogen bomb feasibility regressed its development until new methods permitted it to be circumvented. However, the most significant line of reason is the magnitude by which technological infrastructure determined the rate of nuclear weapon development. In the context of the hydrogen bomb, technology proved to be an impediment.

=Bibliography= Bethe, H. (1982). "Comments on the History of the H-Bomb." Los Alamos Science, Fall 1982:43-53. Campbell- Kelly, M., Asprary, W. (2004). Computer: A History of the Information Machine NewYork: Basic Books Fitzpatrick, A. (Spring, 1998). “Teller’s Technical Nemesis: The American Hydrogen Bomb and Its Developments Within a Technical Infrastructure.” Retrieved February 11, 2009, http://scholar.lib.vt.edu/ejournals/SPT/v3n3/pdf/FITZPATR.PDF Goldstine, H. (1972). The Computer from Pascal to von Neumann. Princeton: Princeton University Press. Kempf, K. (1961). Electronic Computers Within The Ordnance Corps. http://ftp.arl.army.mil/~mike/comphist/61ordnance/index.html MacKenzie, D. (1998). Knowing Machines: Essays on Technical Change. MIT Press Mollenhoff, C. (1988). Atanasoff: Forgotten father of the computer. Iowa State University Press Rhodes, R. (1995). Dark Sun: The Making of the Hydrogen Bomb. New York: Simon and Schuster. Weik, M. (1961). The ENIAC Story. http://ftp.arl.mil/~mike/comphist/eniac-story.html