The Perfect Machine (58 page)

As long as astronomers asked new questions, there would always be dreams of newer and bigger telescopes. This one felt different. It wasn’t just the size and already-growing fame of the two-hundred-inch telescope: The questions of astrophysics were so ripe, and the answers to the eternal cosmic riddles seemed so close, tantalizingly just beyond the grasp of the one-hundred-inch, that it was hard not to feel incredible impatience for the new telescope.

27
Passing the Torch

Sinclair Smith, a bright young astronomer, came to Mount Wilson in the early 1920s. A year of study in England “settled him down,” and he was soon a regular member of the staff, investigating the physical constitution of nebulae and star clusters. By 1931 he and Fritz Zwicky were independently studying clusters of galaxies. Smith scrutinized all the available data on the Virgo cluster, measuring the differences of velocities of galaxies in the cluster, and concluded that the galaxies were moving too fast to stick together; a group with that little mass should fly apart. Zwicky coined the name “dark matter” for the missing mass. Cosmologists today are still searching for the dark matter to balance their equations.

Smith was recruited to the two-hundred-inch telescope project to work on electronics, particularly instrumentation and the drive system for the telescope, the combination of motors, gears, and black-box magic that keeps the telescope accurately tracking objects as the earth turns through the evening. At first glance the task seems easy. The rotation of the earth makes objects appear to move across the night sky. Move the telescope to compensate for the earth’s motion and the objects appear to stand still. The equatorial mounting of the telescope, with its polar axis exactly parallel to the axis of the earth, made the basic motion a simple rotation. The basic task of the drive mechanism was to turn the telescope at the precise speed of the earth’s rotation, but in the opposite direction so that the sky appeared stationary.

But turning the telescope the equivalent of one full rotation in twenty-four hours to match the rotation of the earth isn’t enough. The bigger the telescope, the more exacting the requirements for a drive mechanism. Atmospheric refraction, a minuscule offset of images as the azimuth of the telescope was raised or lowered from the horizon to the zenith, is magnified in a large telescope. The big telescope would be sensitive to eccentricities in the bearings, gears, or mounting structure. A bearing surface as true and smooth as the machine shop could
produce, the surfaces honed to the precision of a watch, would still introduce periodic errors, “bumps” in the motion of the telescope. Over the course of a long exposure, an uncorrected bump would ruin an image or degrade the quality of a spectrum by moving the light from the object off the slit of the spectrograph.

The sixty- and one-hundred-inch telescopes, after decades of use, still had balky quirks in their drive and control systems. Baade used to test budding observers on the sixty-inch telescope at Mount Wilson by seeing if they recognized and compensated for the periodic error in the gear train that had the telescope creep ahead of the stars it was tracking every eighty seconds. By listening for the relays, Baade could tell if the observer was pushing the East and West buttons to compensate. Good observers, like Milton Humason and Walter Baade, were magicians with the machines, with more tricks, bumps, and grinds than a burlesque queen in their repertoire of techniques to get the telescopes to behave. The tricks distracted from the primary task facing the observer, and as the size of an instrument reached the scale of the two-hundred, the idea of
horsing
a five-hundred-ton machine into position became absurd.

Smith wasn’t an electrical engineer by training, but he was a veteran of many nights on telescopes, who understood what astronomers would want from a drive system. At a conference the Pasadena astronomers decided that the drive system should not have any periodic error of more than 1/10 of a second of arc (a second of arc is
1/60
of
1/60
of a degree, or 1/1,296,000 of a circle) for periods of five seconds or more. They also wanted automatic and manual controls of the dome and telescope, with repeater stations so the telescope could be controlled at the various foci.

Their ideal was for the night assistant to dial in coordinates for the precise area of the sky the astronomer had selected, and for the control mechanism of the telescope to do the rest. The astronomers’ dream list included simple control panels at each observing station, with indicators of right ascension and declination, buttons for guiding and slewing the telescope, switches for adjusting the focus, and a telephone for communicating with the night assistant. Instead of horsing with the telescope or contorting himself into strange positions, the astronomer would use as much as possible of his precious time on the telescope for actual observing. Smith understood the wish list.

Robert McMath, who had designed an operating and control system for telescopes in Michigan, and later for telescopes at the Lick Observatory and the new eighty-two-inch telescope of the McDonald Observatory in Texas, was invited to join the project as a consultant. “In the very nature of things,” he wrote,

A project like the 200-inch telescope forces the engineer to extrapolate…. In this case, we are extrapolating many important items,
such as the 200-inch mirror, the pedestal truss, the oil pad bearings, no polar axis defining bearings, the gimbal declination axis connections, the declination axis radial and thrust bearings, the polar axis torque tube, etc. Doubtless most of these items will prove satisfactory in service. Unfortunately, just one of them can spoil the job…. I again urge concentration of your available personnel on the problem of building the simplest possible telescope, considered as a whole.

McMath particularly thought the plans for the electric drive system unnecessarily complex. The system he had designed took care of most corrections, except that it required that the observer manually change the rate of drive from time to time, based on what he observed in the guide scope eyepiece. His much simpler scheme, he pointed out, added “one-half of one percent more burden on the observer.” McMath’s control systems were good, but for this telescope the astronomers wanted even more.

It wasn’t just the astronomers and engineers who had ideas for the control systems. In 1934 Max Mason had recommended that Hale and his colleagues study the “automatic curve-following mechanisms” which were being designed at MIT under Vannevar Bush. Hale had written to Bush, who thought his analog-computer mechanisms might be more accurate than any manual control of the telescope. The idea was to use photoelectric cells to track a guiding star and to trigger signals to control screws that would move the plate holder at the heart of the telescope in response to the apparent motion of the guide star. It was a superb idea, decades ahead of its time, but calculation showed that the field of view of the two-hundred-inch telescope was so narrow that the best of the guide stars would be magnitude 10 or 12, too faint for the photocells then available.

Sandy McDowell, with his long years of experience supervising the construction and installation of naval gun turrets, also considered himself an expert on control mechanisms. He had worked for years with Hannibal Ford, whose small company on Long Island had developed pioneering servo control systems for big naval guns. McDowell wanted to use the Ford work in the telescope. The accuracy required by the drive mechanism for the telescope was orders of magnitude more demanding than the needs of naval guns—someone once calculated for publicity purposes that an error the size of a quarter at three miles would have been unacceptable—but to McDowell it was only an incremental difference. Ignoring the work of the Ford company, he said, was reinventing the wheel.

On an early trip east, McDowell took Sinclair Smith with him to meet Hannibal Ford. Smith had spent his working life at Mount Wilson and had little experience dealing with large corporations or specialized defense contractors. McDowell insisted that no work take place on the drive controls until the Observatory Council had a proposal
and estimate from Ford. Smith was given the job of conveying specifications to Hannibal Ford. Hannibal Ford and his company were not familiar with astronomical telescopes or equatorial mountings, so Smith faced a formidable task in explaining the requirements of pointing a telescope accurately to engineers accustomed to the much simpler task of moving a gun turret. The proposal and estimate from Ford were repeatedly delayed.

While McDowell waited for the Ford proposal, Smith worked on his own ideas for a control system. He worked alone much of the time, and colleagues didn’t notice that he was often pale, easily tired, and sometimes in apparent pain. When the Ford proposal seemed stalled, McDowell wanted Smith to go east again to “consult” with Hannibal Ford. Smith was too ill to make the trip. By September he was in the hospital. Mark Serrurier and others on the project visited him. Serrurier arranged to give Smitty a blood transfusion. Smith didn’t talk about it, even with close friends, but a doctor had told him he had cancer. He worried that he might not see the drive system completed.

By October 1937 Smith was out of the hospital, in remission. The Ford proposal and estimate finally arrived, and Smith had the job of studying the drawings and estimate. McDowell was eager to sign a contract with Hannibal Ford. Smith reported to the construction committee that the Ford proposal didn’t really meet the needs of the telescope. He estimated that a better drive system could be built locally for one-fourth of Ford’s estimate. The committee of Caltech engineers and astronomers, overruling McDowell, voted to have Smith develop his own plans and estimate. They gave him three months, until January 1, 1938.

He met the schedule, and by mid-January, the committee voted to accept Smith’s plans. He pushed ahead on working drawings. As soon as he had a sketch finished, it would go to the draftsmen and off to the machine shop. He reported to the construction committee that the controls would be finished in two months. He was often short of breath and pale, but he pushed on, converting ideas to working drawings and control systems. He told no one that the doctors had given him only a few months to live.

A young Caltech graduate in electrical engineering, Bruce Rule, was recruited to work with him, but the control system was so complex that only Smith understood its full workings. The heart of the system was a corrector unit that would make compensating adjustments in the alignment of the telescope. The “errors” that Smith had isolated were minuscule, far smaller than had ever been worried about on a telescope. He had designed corrections for atmospheric refraction, flexure of the mounting of the telescope, misalignment of the polar axis, and a sinusoidal (cyclical, like a sine wave) skewing of the yoke. He had isolated the remaining uncompensated errors—a nonsinusoidal element in the skew of the yoke, a potential eccentricity of the
telescope bearings, and a slight rotation of the field of view when the telescope was pointed near the pole. The first of those, he concluded, could be corrected by the machining of the horseshoe.

Smith was working frantically on the last remaining problems when he was again hospitalized. After two months in the hospital, Sinclair Smith died on May 18, 1938. He was thirty-nine years old. Max Mason got permission to pay six months’ salary to his widow. John Anderson took time off to write an article on Sinclair Smith for the astronomy journals.

Smith’s death was a terrible loss. He was a bright young astronomer at the prime of his career. He had temporarily suspended his promising astrophysics research to work on the control system for the telescope. He had lost the race to finish and document his work by weeks.

Three months before Smitty’s untimely death, at the beginning of February, Francis Pease was hospitalized for cancer surgery. It had seemed a routine operation, but surgery before antibiotics—sulfanilamide and penicillin were not yet generally available—was never routine. On February 11 Pease died from complications of peritonitis and septicemia. Like Smith, he had put astronomy research aside to work on the telescope. The project had begun with his drawings and models, and Pease had worked to the end to refine the design of the telescope. His work had increasingly been shunted aside by Sandy McDowell, who favored the work of engineers he had known from his navy days and personally disliked Pease. To Pease’s credit, he recognized when alternatives were better than his own designs. He had championed roller bearings, but when Rein Kroon showed that oil bearings would work for the telescope, it was Pease who moved for the adaptation of the radical new design.

Francis Pease’s life had spanned two generations of big telescopes. He had designed the one-hundred-inch telescope almost alone. Others contributed details, but the design and the drawings were his. Twenty years later the new telescope belonged to an era of Big Science, of projects so complex that it was no longer possible for one man to understand all that was involved.

George Hale, who had been confined to his dark room for months, hadn’t been able to attend Pease’s funeral. His nervous affliction had been compounded by a new symptom, attacks of violent vertigo coming without warning and severe enough to confine him to bed. With this, as with his other medical conditions, Hale was not candid, even with trusted friends. From his terse descriptions, the symptoms sound like Ménière’s syndrome, a disorder of the inner ear. It was not treatable, and the unpredictability of the attacks left Hale confined to his house, frequently unable even to go to his beloved solar laboratory.

A few days after Pease’s funeral, Hale felt well enough to be
wheeled outside. He looked up at the sky and said, “It is a beautiful day. The sun is shining and they are working on Palomar.” It was his last word on the telescope. He died a few days later, on February 21, 1938.

More than any other man, George Hale’s name had been synonymous with big telescopes in the United States. He was known everywhere for his research in solar astronomy, for the Yerkes and Mount Wilson Observatories and telescopes, as a cofounder of the California Institute of Technology, a longtime supporter and officer of the National Academy of Sciences and the National Research Council, and the founder of journals of astrophysics. For years, as the public had read the continuing saga of the design and building of the great two-hundred-inch telescope, they were reminded that George Hale had conceived the idea and found the people, the funding, the companies, and the institutions that would contribute to the cooperative project. It was his web of academic, government, and business friends who constituted the old-boy network that had made science and technology on a national scale possible; his prestige that had persuaded the Rockefeller Foundation to commit the largest grant ever made for a science project; his ideas that had pushed the technology beyond limits; and his leadership that had kept the project from foundering when the demands of the telescope grew too large.