The Perfect Machine

by Ronald Florence

  The negotiations between Bush and Mason went on indirectly. Max Mason had one advantage. As the former president of the Rockefeller Foundation, he enjoyed enormous respect and support from the foundation. Warren Weaver sent copies of his diary records of meetings with Bush to Mason and made it clear that the Rockefeller Foundation would not intervene in support of Bush’s plan.

  When Bush and Mason finally met in May 1939, they were able to split many of their differences. Bush was willing to agree to a joint committee to run the observatory provided its chairman was a member of the Mount Wilson staff and had “reasonable power.” Members of the Mount Wilson staff would have honorary appointments at Caltech, where they could conduct seminars and accept graduate students. Astronomy-minded members of the Caltech faculty would be named associates of the Carnegie Institution—an honor that hadn’t yet been accorded to Mason himself. Bush thought that by consolidating basic services of the two observatories, the Palomar facility could be run from an endowment of $1.5 million. He tentatively agreed to include the sum in his requests for funds from the Carnegie Corporation.

  Mason liked the plan. The day after he met with Bush, he met with Weaver and Raymond Fosdick at the Rockefeller Foundation to explain that the plan would require no additional commitment from the Rockefeller Foundation, even though “some individuals might interpret this arrangement to mean that the RF had given an observatory to CIT and that CIT was in effect giving it away to the Carnegie Institution.”

  In the offices of the Caltech astrophysics building, named Robinson Hall after the death of Henry Robinson, Caltech physicists, astrophysicists, and cosmologists had already begun talking about establishing their own program in astrophysics to take advantage of the new observational facilities. The Mount Wilson people were quick to proclaim their expertise and experience at running large telescopes, and to point out that Caltech did not have the funds to provide the operating expenses for the telescope. It was hard for many at Caltech to shake the feeling that the Carnegie Institution was trying to steal their telescope.

  Bush and Millikan were both ambitious men, with great appetites for power and publicity. The control of the most expensive and most famous scientific instrument in the country was a plum that neither was willing to give up easily. Millikan had slowed down—“By actual clocking,” the Pasadena wags reported, “it takes M[illikan] twenty-two minutes to ask whether the R[ockefeller] F[oundation] would object if CIT named the astrophysics building Robinson Hall”—but he was still a fierce defender of Caltech. After a year of talk Mason and Bush were close enough that “less than one sentence stood between them, and they could write out the partial sentence in a way which would be satisfactory to both if they could keep Rob Millikan out of the room while they wrote it.”

  In return for having a Mount Wilson man as director, the Carnegie Institution was willing to allocate up to three million dollars of endowment for the telescope, if not in one grant, in a series of maintenance grants and progressive endowments. Bush was convinced that in the end the Rockefeller Foundation would contribute toward the endowment of the telescope. When Mason offered to bet that they wouldn’t, Bush backed down. Carnegie would pay to run the telescope that Caltech built.

  At Corning, once a year, the accounting department would bring up the question of when it would be time to bill Caltech for the 10 percent “profit” on acceptance of the mirror. When the question came up in 1937 and 1938, the issue was shunted aside. In 1939 Eugene Sullivan, second only to Amory Houghton at the glassworks, decided that Corning would send a bill when the figuring was completed, “which probably will be sometime next year.” Corning had been paid $329,347.27 by Caltech for the entire series of mirrors. The unbilled “profit” for their work on the mirrors was $29,129.14, 10 percent of the expenses other than depreciable major equipment.

  McCauley had successfully cast the last of the disks for auxiliary mirrors, and work was underway in both the Caltech and Mount Wilson optics shops to grind and figure the difficult convex secondary mirrors. His analysis of the striae that had emerged in the two-hundred-inch disk had led to a decision that all future disks would be cast by a new process. Instead of ladling the molten glass and risking contamination from the dropping level in the melting tank, he would mine blocks of cullet from glass that had been melted in a tank built of unused refractory brick. An appropriate weight of this clear glass would then be “sagged,” melted in place, in a mold. To ensure against striae, if there wasn’t a single block of the correct size, smaller blocks of cullet would either be ground and polished on every face or chosen from cleanly fractured blocks, then fused together to achieve the required mass of glass before they were sagged into a mold.

  In April 1938 McCauley got an opportunity to try the new process when Caltech ordered a seventy-two-inch mirror for the big Schmidt camera.* McCauley mined blocks of Pyrex from a fresh batch of cullet, including blocks of 3,200 and 2,372 pounds. The larger block was used to sag the Schmidt mirror. The specifications called for a five-and-one-half-inch hole in the center of the disk, which would have required a core in the center of the mold. Since that core would interfere with the placing of the huge block of glass, McCauley had the disk sagged without the hole and Corning technicians later ground a hole through the nine inches of glass. The disk was shipped in November 1938 and was soon on a grinding machine at the Mount Wilson optical labs. McCauley’s new process was reliable, and another wave of orders poured in from observatories around the country.

  Don Hendrix, chief optician at the Mount Wilson optics shop, was put in charge of figuring the mirror and corrector plate for the Schmidt telescope. For the thin corrector plate he ordered a carload of three-eighths-inch-thick plate glass from the Fuller Plate Glass Company, inspecting the sheets of glass one by one until he found one optically clear enough to be figured as the corrector. The fine ridge he ground into the corrector plate was so subtle, varying in contour less than five-thousandths of an inch over the surface of the fifty-inch-diameter plate, that it could not be seen with the naked eye or felt with the fingers. The spherical shape required for the primary mirror of the Schmidt camera required substantially less work than the paraboloid of the two-hundred-inch disk, so despite a delayed start, progress on the Schmidt was soon ahead of the bigger telescope.

  When Ray Fosdick originally approved the expenditures for the Schmidt telescope, in May 1937, on the grounds that the original grant for the “construction of an observatory, including a 200-inch reflecting telescope with accessories” was broad enough to include the Schmidt, the estimate for the construction of the forty-eight-inch Schmidt telescope was fifty thousand dollars. A year later Max Mason admitted that their original estimate was low. With the machine shop busy on drive gears, machining tracks for the fifty-millimeter ball bearings on the polar axis, welding the huge cannonlike tube out of 5/16-inch plate, machining interior supports of Invar to keep the distance from the photographic emulsion to the mirror fixed, and constructing equipment to prebend the glass emulsions, and mandrils to hold the emulsions in the required curved plane for exposures, the estimate for the cost of the telescope was now closer to one hundred fifty thousand dollars. Max Mason knew the Rockefeller Foundation well enough to know that they wouldn’t object to the cost overrun. “If it were $250,000,” he told Warren Weaver, “the group would unanimously endorse it.”

  The pieces of the observatory were rapidly coming together. Anderson believed that another year of polishing and tuning of the supports would lick the astigmatism problems. One more year after that to parabolize the mirror, deepening the curvature from the spherical figure by about 1/200 of an inch, and the mirror would be ready to go up to the telescope. The drive and control mechanisms would be completed in the machine shop by early 1941. One right ascension drive gear was already on the mountain, and the machinists were almost finished with the fine cutting of the declination gear and the second right ascension gear. Once the gears and the rest of the drive and control system were installed, B
yron Hill and Bruce Rule estimated that it would take approximately eight months to tune the control system with the dummy mirror so the telescope would be ready for the real one.

  The plans all came together. Everything would be finished at the same time. The tentative date for taking the mirror up to the mountain was January 1, 1942. A few months more for final tuning, and the telescope would be ready for the astronomers.

  30

  Impossible Circumstances

  No one wanted war. Opinion in the United States covered the spectrum from Lindbergh, Henry Ford, and the America Firsters, who saw German victories as the triumph of civilization over the “red hordes,” to the president and some of his close advisers, who thought America’s entry into the European conflict inevitable and were biding their time until events turned the opinion of the nation. Even without open conflict America began gearing up for war production. Industries that had been moribund in the depression were soon working at capacity to supply the needs of Lend-Lease and the increased buying and production for the U.S. Army and Navy.

  Though no one wanted war, the talk in the United States was about war. The stories that had once monopolized the news—the debutante parties of Brenda Frazier, Wrong-Way Corrigan’s comic flight to Ireland, or the superlatives of dams, skyscrapers, and telescopes—faded alongside the banner headlines and photographs of marching armies. Newspaper readers and radio listeners could say that the rape of Nanking or the fall of Paris didn’t affect them, but when England hung on by virtue of a handful of fighter pilots, and no nation seemed ready to restrain the Japanese, even Americans who felt safely isolated by two oceans asked what would come next. Would the Japanese attack Singapore? The Dutch East Indies? The Philippines? How long would England hold out alone? Where would Hitler, now master of the entire European continent, turn next? Another assault on England? Or would he attempt what even Napoleon had failed and attack the Soviet Union?

  Even on Palomar Mountain, connected to the world outside only by the daily radiotelephone calls to Pasadena and visitors, the war intruded. Five years before, in 1935, there had been no shortage of labor: men had only to hear a rumor of a job, and they would drive, hitchhike, or walk up the mountain road for the chance of work. For fifty cents an hour they were willing to live on an isolated mountaintop, sleep in a rude bunkhouse, put in a six-day week, and spend Saturday night with the boys and a bottle of cheap wine. By 1940 men weren’t so desperate for work. The New Deal was no longer new. Nine and one-half million Americans, 17.2 percent of the total work force, were still unemployed, but it was a far cry from the depths of the depression, when some heavy industries, like steel, had slowed to the point where more than half of all workers were out of work, and companies like U.S. Steel had so parceled and split jobs that they could honestly declare that they had no full-time workers on their payrolls.

  The outbreak of full-scale war in Europe and Asia completed the economic fix that the New Deal had begun. War orders lit the furnaces and started smoke up the idle stacks. In 1939 the aircraft plants in San Diego and the Los Angeles area were again hiring. By the end of the year Douglas Aircraft, bristling with $18 million of back orders, had ten thousand men on the payroll. Factories that built airplanes or tanks or trucks bought parts and raw materials from hundreds of other companies, bringing work to aluminum smelters in upstate New York, rubber factories in Ohio, and engine plants in Connecticut. Building trades picked up. Companies and contractors competed for electricians, construction workers, and mechanics—exactly the jobs that were needed on Palomar Mountain to finish the telescope.

  Some workmen left the mountain, especially family men tired of commuting once a week to wives and children in Escondido, and men who had come to Palomar only because there were no other jobs. Working on the mountain suited some men. Whether they were fascinated by the machine they were building and proud of their role in bringing it to life, or enraptured by the beauty and mystery of the mountain, many stayed even when better-paying jobs came along. The mountain had its own rewards, the beauty of the trees and snow and air at five thousand feet, the evening skies, and the pride of working on a machine so huge and famous that the word “Palomar” evoked pride in those who built it and immediate recognition in others. Like the windowless optics shop, the mountain was a place men loved or hated.

  Through 1940 and 1941 the work went on. In August 1941 one of John Anderson’s Saturday tests determined that the mirror surface had achieved a satisfactory sphere with a radius of curvature of 1335.7 inches. On August 30 Marcus Brown and his men started the final stage of figuring—parabolization—polishing the shape to a slightly deeper curve. By the end of September the parabolization was 90 percent complete. What remained was the final figuring, bringing the entire surface to ultimate smoothness, so that the deviations from a perfect optical surface would be measured in fractions of a wavelength of light.

  The original plans for figuring the mirror, drawn up when George Hale first turned to Corning for glass blanks, included the 120-inch disk both as a trial run for the casting of the two-hundred-inch, and to use as an optical flat mirror for use in testing the big mirror. The machine shop had built a special grinding machine for the 120-inch mirror, and the dimensions of the optical shop had originally been chosen to permit room to test the telescope mirror with a flat mirror at the proper distance.

  Figuring a large flat mirror is a difficult task. The normal procedure for grinding mirrors, rotating one disk over the other with a slurry of carborundum and water between the disks, produces a concave curve in the lower disk. Creating an optically flat surface on the lower disk requires special care. A flat used for testing had to be figured to tolerances as demanding as the big mirror, lest imperfections in the flat mirror confuse the tests. Brownie had estimated that it would take a year or more to figure the flat.

  Anderson didn’t want to wait. In 1941 he and Frank Ross, who had been developing corrector lenses for the prime focus of the telescope, asked whether it wasn’t possible to test a big mirror accurately without using a large flat mirror. They fiddled and finally came up with a scheme that would use a much smaller half-silver flat mirror and a special lens to focus the light from a light source in a way that would precisely test the zones of the big mirror. It looked right on paper, so they had Brownie’s men build the lens and the small flat mirror—relatively simple tasks. Their scheme—they never claimed much credit, because they weren’t sure no one else had ever tried it—worked beautifully. In the hands of a skilled optician like John Anderson, the test showed patterns of dark and light in the eyepiece that told him which zones needed additional polishing.

  All that remained was to polish down those zones, in tiny increments, until the mirror took the perfect paraboloid figure. Some weeks Brownie would polish with a small tool for only a few minutes each day, resting the mirror for hours to let the heat of the polishing dissipate before the mirror was tilted up in test position to measure the effects of the polishing. When asked, Anderson and Brownie would predict that a few more months of polishing could finish the mirror. The figure on the mirror had reached the point at which Anderson was doing much of the testing late on Saturday night, so traffic on the street outside would not disturb his measurements. Sometimes alone, often with Brownie there, Anderson would study the shadows and write notes. Some Saturday night, Brownie knew, Anderson would finish his tests and announce that the mirror was ready.

  The point when an optician stops figuring a mirror is arbitrary. The small mirror of an amateur telescope might be figured to one-fourth of a wavelength of light, a precision of hundreds of thousandths of an inch. The surface of the two-hundred-inch mirror would be figured to one-fortieth of a wavelength of light, a precision of approximately one two-millionth of an inch. It was the most formidable optical task ever attempted. The closer they got, the more obsessed Marcus Brown and the men in the windowless room became. They were polishing smaller and smaller areas, narrowing the zones of imperfections, bringing the entire surface c
loser and closer to the elusive goal of a perfect paraboloid.

  In the machine shop next door, the machinists on the big-gear-grinding machines worked toward a similarly elusive goal. The smoothness of the motions of the telescope, and the freedom from backlash that would enable the control and drive systems to keep the telescope precisely guiding on faint stars, depended on the machined precision of those gears. The design goal was for the drive mechanism not to vary by more than one arc second per hour at the normal drive rate of one revolution per day, and for short period errors to be limited to one-tenth second of arc per five seconds of motion.

  There were three big gears, one for the declination axis, tipping the telescope tube in the great yoke, and two for the right ascension axis, turning the yoke on the great horseshoe bearing. One right ascension gear was for slewing the telescope, moving at relatively rapid speeds when the observer wanted to point the instrument at a new area of the sky; the second gear was for the slow-motion drive that would keep the telescope moving synchronously with the motion of the earth. By using a separate gear for the relatively rapid slewing motion, wear and tear on the main drive gear would be reduced.

  Earle Buckingham of MIT had done much of the design work on the gears, working without compensation. Each gear was fourteen feet in diameter and weighed ten tons. The gears were cast by Westinghouse and rough-machined at their South Philadelphia Plant, then shipped west. Special gear-cutting machinery had been designed and built in the astrophysics machine shop.

  These were the largest high-precision gears ever made. One second of arc equals 0.000445 inch on the pitch circle of the gears, which meant that the 720 teeth on each gear had to be machined to a tolerance of one ten-thousandth of an inch, crude stuff by optical standards, but demanding for mechanical work on that scale. The process, from the rough-cutting to the fine-polishing of the gears, took two and one-half years. To limit the expansion and contraction of the gears, the work was done in a subroom inside the machine shop, where air-conditioning could maintain the temperature at an even seventy-five degrees. The mechanics who ran the machines needed patience to match that of the opticians in the building next door. They examined the mating surfaces of the gears with microscopes. Soon only special measuring gauges could record the fine change in the surfaces.