The Perfect Machine

by Ronald Florence

  The equipment had to run twenty-four hours a day, spraying layer after layer to build up the clear quartz surface. The operation went well until the blank was half glazed. One of the three heating elements in the furnace burned out. Two elements could maintain the needed 1700° temperature, so work continued. Then another element burned out. Ellis ordered the furnace partially cooled, repaired, and refired. Before it was hot enough to restart the spraying equipment, the repaired elements failed again.

  Ellis and Thomson concluded that the furnace had to be rebuilt. Ellis took advantage of the shutdown to do some planning for the big disks, extrapolating from their experience on the smaller one. The figures he came up with were shocking: A surface layer 2.5 inches thick on a two-hundred-inch-diameter disk would require seven million cubic feet of hydrogen fuel—enough to fill two Graf Zeppelins. To surface the disk, GE would need either an enormous hydrogen plant on the premises, or a gasometer one hundred feet in diameter and seven hundred feet tall. In an era long before the creation of the Occupational Safety and Health Administration (OSHA) or the explosion of the dirigible Hindenburg at Lakehurst, New Jersey, in 1937, no one gave much thought to the danger of storing that much hydrogen near an industrial plant.

  The hydrogen consumption was so daunting that Ellis tried calculations for alternative fuels. A nearby plant produced dissociated (chemically separated) ammonia. A rail spur connected the plants. Ellis calculated that 3,700 tank carloads, each of 2,500 cubic feet of ammonia, would be enough for the two-hundred-inch mirror. A shuttle train could carry the tanks, if the trains didn’t break down and if they kept enough men on duty to load and unload the continuous stream of tank cars. With either fuel the lab would need a huge supply of oxygen for the furnace. An on-site plant could be built to produce it. The project was beginning to seem bigger than anyone had anticipated.

  Ellis’s mechanics finally got the electric furnace rebuilt. During testing the furnace broke down with nagging regularity, but Ellis persisted and produced two twenty-two-inch fused-quartz disks by mid-August 1929. The disks were suitable for testing, but Ellis warned Anderson in California that the quartz had emerged from the furnace with mysterious black specks embedded in the surface.

  With the disks packed and shipped off to Pasadena, Ellis again shut down the furnaces, and had the room sealed off until he could figure out what caused the specks. Chemical analysis was inconclusive. He tried introducing traces of potential contaminants into the flame of the torch to see if he could produce comparable spots. After weeks of testing, iron turned up the likely culprit. When tests of the refractory bricks of the furnace, the piping and burner components, and the oxygen and hydrogen gases that had been fed to the burner could not detect quantities of iron greater than 0.0001 of 1 percent—too small to account for the specks—that left only the quartz itself as the source. The contaminations in the supposedly pure quartz were proving as troublesome and incurable as George Hale’s demons.

  One correlation that emerged from the testing was that the specks seemed to be most numerous when the quartz had been deposited onto the disk at relatively low temperatures. That meant an end to Ellis’s idea of using dissociated ammonia as a fuel for the spraying. Ammonia would have been cheaper, safer, and easier to transport and store, but it would fire the burners at a lower temperature than hydrogen. It looked as if they still needed two Graf Zeppelins of hydrogen.

  Charges for overtime and new equipment piled up. While they waited for the results of the optical tests on the disks that had been shipped to Pasadena, Ellis and Thomson planned for the fabrication of a sixty-inch disk that would actually be used for one of the secondary mirrors in the telescope. The telescope would ultimately require two or three sixty-inch secondary mirrors, on the optical paths for the Cassegrain and Coudé foci. The laboratory at West Lynn was big enough to house the furnace for the mirrors, but Thomson confidently authorized construction of a new building for the next stage of the spraying operation.

  No one had ever seen a building quite like what Thomson designed: sheet iron sides and roof over a structural steel frame. And Anderson, in Pasadena, had never seen a bill like the one Thomson forwarded. With the connections for heat, light, water, steam, and gas, the electrical equipment to regulate the furnaces, and the construction of a furnace large enough for a sixty-inch mirror, the building cost more than $115,000, close to half of the entire original GE budget for producing all the mirror blanks for the telescope. When Anderson questioned the expense, Thomson pointed out the advantages of his design: The structure could easily be expanded to accommodate the spraying of the two-hundred-inch mirror, and the steel walls of the building would have a high salvage value when the operations were finished. Nothing was too good for a perfect machine.

  At the Mount Wilson optical labs on Santa Barbara Street, an optician rough-ground one of the twenty-two-inch quartz disks to a spherical curvature, the first stage in the shaping of a telescope mirror. It didn’t take long for his grinding disk to cut through the transparent quartz layer to the rough quartz underneath, in an area just off the center of the disk. Anderson reported to Ellis that the fused layer was only one-eighth inch thick.

  “Impossible,” Ellis answered, insisting that the transparent layer had to be consistent in thickness over the entire surface of the disk because the rough quartz had been ground flat before they began spraying on the transparent layer. A week later he admitted that while the edges of the disk had been ground flat, the center might have been “a little high.” In any case, Ellis wrote, it was impossible to reglaze the test disk because the furnaces were already being reconfigured for the next mirror.

  There was little hope of making a successful mirror out of the disk, but Anderson ordered the opticians to keep grinding. At least they could learn more about how the fused quartz behaved under the grinding tools. As the tool ground deeper into the quartz disk, layers of bubbles appeared. The deeper the tool worked, the more numerous the bubbles. The fused quartz was extremely hard, and the edges of the exposed bubbles were so sharp they tore the polishing tool. Each bubble had to be dug out with hand-grinding tools. It wasn’t a good omen.

  The progress on the disks was so unpromising that as a precaution Anderson quietly kept his fingers on sources of alternative materials for the mirror, as a “second line of defense.” U.S. Steel, eager to promote its newest stainless-steel alloys, lobbied for a steel mirror. The Philips Lamp Works in Holland promoted its process of fusing a layer of glass to a metal base. Under contract from the Bureau of Standards, the English firm of Parsons, experienced opticians and telescope makers, were trying to build up mirrors from thin glass plates in a cellular mosaic. None of these alternative materials was accorded much of a chance by the opticians at Mount Wilson.

  From his new base in France, George Ritchey was eager to get funding for a vertical three-hundred-inch reflector with a fixed mirror. A movable coelostat and secondary mirror would feed light to the primary mirror and back to a fixed focus, allowing the observers to remain at a fixed observation position. He wanted to build the telescope on the edge of the Grand Canyon, which he announced was the ideal site.

  Ritchey designed cellular mirrors for his proposed telescope, built up of small sections that would fit together like a symmetrical jigsaw puzzle. He argued that a composite mirror could be lighter than a solid disk, the individual cells of glass could be thinner and hence less subject to thermal distortions, and the frame holding the cells in relation to one another would provide great rigidity to the mirror.

  Adams, Anderson, Pease, and Seares all studied the Ritchey scheme and concluded that while a perfect cellular mirror might work, the demands on the framework that would have to hold the sections in alignment with one another were so great that the design couldn’t be realized. Ritchey experimented with cellular mirrors, cementing the sections of glass together with specially fabricated presses. He made great claims for the process but was never able to eliminate the “quilt” pattern in the surface of the
glass from the cemented cells. A coelostat telescope, like his designs, also has a problem of limited declination range, which limits the amount of sky the telescope can see even more than the English mount of the one-hundred-inch telescope.

  Whatever the merits of Ritchey’s ideas, no one in Pasadena wanted much to do with one of his designs, whether for mirror construction or the Ritchey-Chrétien optics with its promise of a wide field of sharp focus. Ritchey’s last years at Mount Wilson had alienated almost everyone. His projects, Adams wrote, “seem to be a case in which Professor Ritchey’s wish is father to the thought.”*

  There were just enough signs of progress from West Lynn that Anderson kept the alternatives on hold.

  One year after the public announcement of the grant for the construction of the telescope, there were already signs of the work in locations around the country, not only Pasadena, but in architectural offices in New York, at the GE laboratories at West Lynn, and at sites where astronomers were researching the observational conditions. But despite the flurry of activity, the great “eye” that had captured headlines when it was announced had already faded from the news. Occasionally a flamboyant Southern California evangelist would gather temporary notoriety by blasting the telescope project as an enterprise of blasphemers, destined to bring the curse of perdition on the builders, and perhaps on mankind, for their arrogance. From time to time GE or another company with a contract for some part of the project would time a press release on its work to draw attention to its company in the newspapers. But the slow developments in Pasadena and elsewhere weren’t really headline stuff—not when the nation was suddenly overwhelmed with much bigger news.

  The stock market had been front-page news even before the telescope project started. The market had been edgy all through the latter half of 1928. In December 1928 a sharp “correction” was followed by a brief panic that had newspapers pulling out the 144-point type. The market rallied and pulled through, but when the Federal Reserve refused to extend speculative credit, the market collapsed again in February 1929. Call money soared to 20 percent as the New York banks poured money in at 15 percent to rescue the market.

  Despite the frightening volatility, speculators still saw optimistic signs. Every slump, the analysts observed, was followed by a recovery, each time to ever greater heights. Those who had been successful with their investments, or who had borne their losses with aplomb, chided those who were still watching from the sidelines. Before long Americans who had never invested in stocks decided that they, too, had to have what the slangsters now called “a piece of the action.” “It can’t last,” the voices of doom cautioned, but remarkably, the predictions of the optimists proved true. Between corrections, the market went up and up and up. Savings and borrowed money poured into the market. No one wanted to miss out on the apparently free bounty. By the summer of 1929, close to one million Americans held stock on margin.

  The crash came on the first anniversary of the announcement of the telescope project, October 28, 1929. Overnight, paper fortunes collapsed. The falling market took good money with the bad, old money with the new: Once-solid accounts, committed to the market at the last minute in an effort to restore order, fell as hard as the margin accounts of speculators. The cautious few who could afford their losses painfully picked up the pieces. Those who had invested everything, often on margin, were ruined. Bankers weren’t the only ones to jump off the fine bridge over Arroyo Seco in Pasadena. Before long the city of Pasadena would be spending twenty thousand dollars per year guarding the bridge.

  There were brief signs of hope in the following months, as the market recovered a portion of the lost ground, but it was soon obvious to all but the most stubbornly optimistic that the crash had inflicted mortal wounds on the market and the American and world economies. Even the pessimists who had predicted the collapse hadn’t imagined the consequences of an economic crash on a nation in which one-half of 1 percent of all Americans possessed one-third of the national wealth, and 80 percent of American families had no savings at all.

  The real impact of the crash on the national and world economy wouldn’t be felt for months or even years. But Black Tuesday shattered the last remnants of the optimism of the 1920s. A decade of mania ended overnight. In the months following the crash, there was no world-class prizefight, no great new athletic feat, no murder trial of national interest, no ticker tape parade, no spectacular flight to match Lindbergh’s achievement. The Atlantic City beauty pageant was canceled. Shipwreck Kelly came down from the summit of the Paramount Building because no one was watching anymore. The stunts that had fueled public interest in the mad 1920s gave way to desperation. Two men drove a Model A Ford across the country in reverse, but no one paid much attention. Russia’s Five-Year Plan and the reports of recent visitors to the Soviet Union were suddenly of more interest than the amusing stunts, as those who suffered sought relief in dreams of a different political and economic system.

  Even those who had resisted the lure of quick money and margin speculation weren’t immune to the infectious spread of the crash. Henry M. Robinson, the chairman of the Board of Trustees of the California Institute, who had personally guaranteed the endowment of the telescope, was ruined. A year before, a personal guarantee from him was beyond question; now his investments were in shambles. When the matter was brought up at a meeting of the Board of Trustees, A. H. Fleming, the president of the board, reported Hale’s original estimate of $150,000 per year or an endowment of $3 million. After the meeting, a board member leaned over to the president to say, “Fleming, you’re not worrying about this, are you? I’m going to take care of this matter myself.”

  The trustee who made the generous promise never came through with the funds. Progress on the telescope had become so halting by late 1929 that the council and the Board of Trustees stopped worrying where an endowment for the instrument would be found.

  The grant to build the telescope was still secure, but California Street, where the new astrophysics building and machine shop were going up at the edge of the Caltech campus, was in the midst of its own depression.

  The Mount Wilson optical laboratories were still testing pieces of fused quartz from GE. Anderson tried holding his hand in contact with the surface of a disk for a minute. With a glass disk, the heat of his hand would have raised the temperature of a portion of the disk measurably, distorting the optical surface. On the quartz disk Anderson and the opticians repeated the experiment three times and could detect no measurable distortion. Fused quartz still seemed an ideal material for a telescope mirror.

  But after almost a year and a half of experimentation, GE hadn’t produced a single satisfactory disk. Ellis sent samples of quartz bars off to the Bureau of Standards for testing. The native quartz was cheap but full of imperfections; Brazilian quartz was pure, but expensive. He came up with another idea: If he could find quartz supplies for the backing and the sprayed surface with nearly identical coefficients of expansion, Ellis suggested, they could use the cheap native quartz for the molded backing and the pure imported quartz for the sprayed surface. It would save money and was sure to produce a satisfactory disk.

  Like most of the plans coming from West Lynn, this one sounded good. But with expenditures already over the original budget for the mirrors, the date when the two-hundred-inch mirror was to have been finished long gone, and not a single usable mirror finished, there were doubts in Pasadena that the West Lynn laboratory would ever produce a usable mirror disk.

  Hale called Russell Porter over to the solar laboratory and showed him the letter of agreement for the grant from the IEB, pointing out a clause they all hoped would never be invoked: “PROVIDED, That if at any stage of the project it be decided that the construction of the telescope is not feasible, any remainder of the amount hereby pledged by the Board, according to the terms above prescribed, shall be and become null and void.” Hale then sent Porter back to West Lynn to determine just how much truth lay under the promises and assurances Ellis and Thoms
on offered almost weekly.

  Privately Hale questioned exactly what Gerard Swope had meant when he agreed that GE would do the work “at manufacturing cost.” Swope’s original telegram had said that there would be no charges for “commercial or administrative expenses,” but it wasn’t clear who was paying for the time and expense of filing patent specifications, in the name of General Electric, on every step of the process. Hale never put it in so many words, but there was a clear conflict between the commercial style of GE, which saw the project as prestigious publicity and an opportunity to develop processes with future commercial potential, and the scientists in Pasadena, who were concerned only with getting a working telescope before they spent their entire budget.

  Efforts to pin GE down on costs got nowhere. Hale and Anderson raised the question with Thomson and Ellis. Robinson brought it up with Swope. Everyone at GE hewed to a consistent party line: Fused quartz was an experimental process; many bugs had to be worked out before it would work successfully; GE was committing valuable personnel and resources and making no profit; and any budgets were only estimates. As long as GE was making some progress, Hale was reluctant to press too hard on the budget. Given the record of the temperamental one-hundred-inch Hooker telescope, which still hadn’t reached the theoretical resolution the Mount Wilson designers had predicted, Hale and his colleagues weren’t ready to settle for a second-rate mirror for the Big Telescope, no matter what the difficulties.

  In West Lynn, Ellis—despite the consistently optimistic reports he had sent to Pasadena—began to have doubts of his own. The whole process had grown too complex. The costs of building and fueling a furnace large enough to melt quartz sand for the base of a sixty-inch mirror were enormous. After the base was molded, they would face the experimental spraying procedure for the surface. The slightest mismatch in the temperature coefficients of the batches of quartz used for the base and the surface risked distortions or strains in the disk. For an engineer used to production processes, there were too many ifs.