The Perfect Machine

by Ronald Florence

  The obvious solution was heavier ladles. Corning had little experience with ladles, and queries indicated that even the big window-glass-makers had never used ladles thicker than ¾ inch. When even the window-glassmakers shifted to machines for plate glass, ladle technology had stopped. Fortunately, the purchasing office queries led to Bethlehem Steel in Buffalo, which still had the dies they had once used to press ladles for the window-glass industry. The depression had left their mill slack, and they were willing to bring out the old dies and press one-inch-thick ladles—all within fifteen days.

  By early April 1933, McCauley was ready to cast a 120-inch disk. The masons and electricians built a new annealing oven, large enough for the two-hundred-inch disk. McCauley’s conservative calculations came up with 516 for the number of oversize GE heating elements needed to maintain the temperature of the oven. To avoid the possibility of shorts, the heating elements were mounted in rows on the sides and bottom of the oven cover, with heavy-enough wiring and masonry to survive an annealing period, at high temperature, of up to one year. The thirty-five-volt circuitry reduced the insulation problems, but the increased current required large-diameter wiring from the panels of reactors and theater-dimming controls. Ten thermocouples were spaced around the upper, lower, and wall surfaces to continuously monitor the internal temperature. The controls were set up so an operator could control the internal temperature to within 1 degree in the range between 400° and 550°C.

  The cores for the new mold were secured with cold-rolled steel rods to avoid an accident like the floating core that had marred the 60-inch disk. Just in case, McCauley had the masons prepare a spare set of cores. If the mold failed during the pour, they could quickly build a new mold for another try.

  Corning masons converted one of the large melting tanks in A Factory to what the glass industry called a “day tank” for the 715-CF glass. Normally melting tanks had a bridge wall that separated freshly introduced materials in the back of the tank from the ready glass in the front. With the bridge wall removed, the tank could melt a single large batch of glass. Millwrights installed overhead tracks to support the heavy ladles, leading from three openings in the tank to three openings in the igloo over the mold, each 120 degrees apart around the mold. McCauley designed the tracks so the “tank and casting oven operated like the moon, with the same surfaces facing away from the workmen.” In the heat and panic of filling a mold, he didn’t want workmen confused by having to spin around and change directions while they were guiding ladles filled with 750 pounds of molten glass at 1500°C. McCauley also had three special wheelbarrows fabricated for the ladle heels, with two wheels in front, to make it easier to avoid sideways tipping when the workmen were moving around the tank in the heat and confusion of a long pour.

  At the end of April the tank was lit, and the filling began with 715-CF batch. Wheelbarrow after wheelbarrow went up the ramps to the loading shovel at the opening. Thirty tons of glass requires an enormous supply of sand, borax, soda, and lime. The level of molten glass inside the tank rose four inches per day. After a few days a workman discovered a break in the tank lining. McCauley ordered the tank cooled and rebuilt. There weren’t enough of the special low-contamination refractory bricks on hand, so the masons cannibalized bricks from the floor of the tank to fix the walls and built a new floor from common bottom bricks like those used in the normal production tanks, with a four-inch layer of clay tamped over it. No one was sure a tamped bottom would work for a melting tank, but the alternative was to wait months for the Pot and Clay Department to prepare a batch of new bricks. After the months lost searching for a new glass formula, delay wasn’t a welcome word.

  Within the world of astronomy, observatory directors followed every move of the group at Caltech, hoping they could piggyback on the research efforts to obtain disks and technology for their own planned telescopes. Dr. O. A. Gage, head of the Aircraft and Instruments Division at Corning, took advantage of the talk to solicit more orders for telescope blanks for Corning. Telescopes would always be a small portion of Corning’s business, but once a portion of the facilities had been set aside for the bulky casting and annealing ovens, a large melting tank had been dedicated to the special glass for the telescopes, and personnel—not only the ladlers but electricians, millwrights, carpenters, masons, tinsmiths, and glassmakers—had been assigned temporarily from other work to the telescope project. Aside from the juggling of the annealing ovens, more orders just meant an economy of scale for the operation.

  Once he found out that Caltech was abandoning all work at GE, Harlow Shapley canceled his order for two 60-inch disks from GE and ordered them from Corning. The David Dunlap Observatory at Toronto asked for a 76-inch disk, which would ultimately become a 74-inch telescope blank, and the McDonald Observatory of the University of Texas ordered an 82-inch disk. Heber Curtis, Shapley’s old adversary in the great debate, who had left Lick Observatory for the University of Michigan,* ordered a large disk of undetermined size. Each order for a primary mirror was accompanied by orders for one or two auxiliary disks.

  In all it was enough work to keep the tanks and glassmakers busy for over a year. And since the available annealing tanks and casting igloos were limited to casting one large mirror at a time, and the 120-and two-hundred-inch disks for the Observatory Council would keep those busy—between casting and annealing time—for at least the next two years, McCauley had another set of molding and annealing equipment built south of the existing set, but close enough to the big 3A tank to use the same batches of glass. The new kiln was big enough for an 84-inch mirror. The smaller molds and kiln required less elaborate equipment for raising and moving than the big molds and kiln for the 120-and 200-inch mirrors.

  By the third week of June the 3A tank was filled and the glass mixture had set (the glassmakers would say it had been allowed to “plain”) for nine days until the glass was deemed ready. It was a Wednesday when the Pyrex was ready to pour. McCauley decided to use the 76-inch mirror for Toronto as a trial for the molding procedure, and to postpone the 120-inch disk for the Observatory Council until Saturday morning, when the regular blowing room personnel would not be in the factory to distract the ladlers and other crew.

  The pour of the 76-inch disk went without a hitch, and it was consigned to the smaller annealing oven. The melting tank was topped up with enough batch mixture to replace the 2½ inches of glass that had been used from the tank. The crew to pour the 120-inch disk—the largest piece of glass ever poured—would be the same crew that had poured the 76-inch, augmented by the additional personnel needed for an operation that would keep three ladling positions busy. Two additional men operated switches on the overhead tracks that supported the heavy ladles. Two more men worked the additional ladles, and four men ran two more wheelbarrows to carry the ladle skins back to the tank. Because the journey from the tank to the mold was longer, three men with backpack spray tanks carried ring-shaped nozzles they could hold over the lips of the ladles while they sprayed water to keep the lip of the ladle cool. After each ladle was dumped, the sprayers would retreat to a recharge station to get their backpacks refilled with water and compressed air. From the first simple pours with one man handling a ladle, the operation had become a precisely choreographed ballet with a company of dozens of men.

  On Wednesday evening, June 21, McCauley ordered the burners of the casting oven igloo lit in preparation to cast the disk the next Saturday, only to discover that the expansion of the metal anchor bolts that had been added to hold down the cores had dislocated the tops of the cores. The masons came in for repairs and finished on Saturday morning at 5:00 A.M., just as the pouring crew assembled. Except for a furnace staff to attend the fire under the tank, the factory was empty when McCauley gave the order to begin.

  With no audience to distract the crew, the procedure went smoothly. Everyone worked according to the script. The area around a fired glass tank—with the roar of the furnace, the clanging of the metal doors, the rumbling of ladle carrier tracks o
n the overhead rails, and the rattle of the metal barrow wheels against the steel floor—was far too noisy for verbal cues from a stage manager. And the sheer level of activity—along with the danger of moving ladles, each filled with 750 pounds of molten glass, heavy equipment, and barrows of glass slag—was far too dangerous for visual cues.

  As each ladle of Pyrex emerged from the tank, the switchers had to route it to the correct port on the mold. A sprayer stood ready, avoiding the two men on the ladle arm and the heat of the full ladle of molten Pyrex as he held the ring of his sprayer centered around the ladle’s lip. The threesome walked under the track together, the ladle handlers maneuvering the ladle with its load of molten glass as the carrier rolled along the tracks to the opening over the mold. The sprayer pulled back just before the ladlers lifted the cup of the ladle into the opening in the igloo over the mold. When the molten glass was tipped into the mold, half would remain in the ladle, cooled enough by the journey to stick to the steel boilerplate walls of the ladle. As the ladle came back out of the mold, men were ready to empty the ladle skin into a waiting wheelbarrow, which was wheeled to the rear of the tank to be returned to the cauldron of molten glass.

  Cycle after cycle, the pour went flawlessly. By 7:30 the mold was filled. The ladles and wheelbarrows were moved away as the crew gathered around the mold, staring at the disk of molten glass through the pour openings. When the last ladles of glass were added, the glass on the top of the mold was hotter than at the bottom. The cooled glass around the molds stood out, outlining the geometric pattern of the ribs and cores. While McCauley and the glassmakers watched, the glass at the surface began to cool, and the molds became radiant against the dark ribs of glass. The changing image was so beautiful that the crew held off, mesmerized by the spectacle, before they finally lifted the igloo off and maneuvered the mold down its tracks to the annealing oven.

  By four o’clock in the afternoon the mold had been lowered from beneath the igloo and slid down its tracks to a position under the annealing oven and raised again. An insulating seal of Sil-o-Cel powder was added between the edge of the mold and the oven cover. McCauley had calculated and recalculated his annealing schedule. For 11 days the disk would be held at 520°C. Then it would cool by 1.6°C per day for 140 days, and finally be allowed to cool at its own rate, limited only by the insulation surrounding the disk. The crew went ahead pouring smaller disks during that period, but the 120-inch disk, the largest glass casting ever poured, was the test of McCauley’s procedure and the glass formulation. If it worked McCauley was certain they could cast the two-hundred-inch disk. If not? McCauley didn’t have contingency plans.

  Theodore Dunham, of the Mount Wilson Labs, was at Corning to witness the pour, which he called a “magnificent spectacle.” Hostetter was Dunham’s guide in Corning, and he gave both Dunham and Hale the strong impression that much of the scientific and engineering work on the disk project was his doing. “Dr. Hostetter,” Dunham wrote, “does not trust a crane for such delicate work and thinks his experience with this table might be useful in designing equipment for the optical lab.” Hostetter announced that he was anxious to cast all the secondary mirrors at the same time with the two-hundred-inch in order to avoid the great expense of heating the furnace a second time with another melt of the special glass. Observers in Corning noticed how often Hostetter was calling the telescope project—which he oversaw as a project manager, but on which he had done no scientific or engineering work—his project.

  While Hostetter claimed the credit, McCauley checked the annealing oven every day, including Saturdays and Sundays, for six months. As a senior engineer/scientist, he had no production responsibilities for the disk. He could have asked the production crew foremen to assign someone to monitor the equipment. But McCauley was an orderly man. He considered himself responsible for the telescope disk project, and he insisted on checking the oven himself.

  George Hale fled to Europe again that fall, to get away from the pressures and the demons. On his way home he arranged his travel to stop in Corning while the disk was still in the annealer. Hostetter, who enjoyed the visits of the famous to Corning, joined McCauley in showing Hale the annealing oven in A Factory that held the disk, the casting equipment and molds, and the panel of electrical controls that regulated the heat. Hale was “delighted with everything I saw.”

  At Christmas 1933, six months almost to the day after the disk had been poured, McCauley lifted the cover of the annealing oven to peek at the disk. There were no cracks, no pieces of the kiln broken away and embedded in the disk, no displaced cores to mar the ribbed pattern. The glass was clear, with the characteristic yellow color of Pyrex. Under close examination the strains in the glass were minimal, close to what he had expected after annealing. The only flaws in the disk were small bubbles (what the glassmakers call “vacuum bubbles”) near the tops of some of the round cores. McCauley knew the bubbles would not affect the disk.

  “We were obliged to admit,” McCauley wrote, “that our product, while wholly suitable for the service for which it was to be used, was not the sleek object, without blemish, for which we dreamed. We could only accept our disappointment and try for greater perfection in the 200 inch disc, the next chance to produce a flawless disk.” McCauley’s prose doesn’t quite conceal the pride of achievement. To avoid bubbles on the tops of the cores in the next disk he decided to make the large cores hollow, so the surfaces in contact with the glass would cool rapidly.

  In Pasadena the optical shop on the Caltech campus was under roof and work had already started on a grinding machine for the 120-inch disk, which was to be ground as an optically flat mirror to use in testing the two-hundred-inch disk. Hale was encouraged, but had been steeled by long experience to anticipate the unexpected. “Large telescopes,” he wrote, “as I have learned before, are secular phenomena. But fortunately the Corning estimates of cost do not increase beyond their original figures.” He had been surprised too many times not to keep his guard up.

  Still, when he heard the good news from McCauley, he couldn’t help a tone of confidence. “The point of doubt as to the possibility of getting a satisfactory 200” Pyrex disk,” Hale reported to the Observatory Council, “had been passed.”

  17

  “The Greatest Item of Interest … in Twenty-five Years”

  At the annual meeting of the American Ceramic Society in Cincinnati, in February 1934, Arthur L. Day delivered the Edward Orton Jr. Memorial Lecture. Day, the director of the Carnegie Institution Geophysical Laboratory, began with the requisite pseudo-Shakespearean paraphrase—“All the world is a ceramic product”—as he traced the history of ceramics from the formation of natural glass in the earth’s crust to the current frontiers of ceramics research and to the most difficult of optical challenges, the mirror disk for the planned two-hundred-inch telescope. The ideas that had been pursued for the telescope disk, he reported to the audience, included everything from sawing off slabs of the obsidian cliffs at Yellowstone Park to fabricating disks of fused silica. There were problems with both extremes, which was why Pyrex-brand glass had become the compromise choice. Everyone at the meeting knew that Pyrex meant the Corning Glass Works.

  Reporters hurried to ask Day questions. He confirmed that Corning was casting mirror disks for the big telescope, including the two-hundred-inch disk for the primary mirror: “This disk in all its details is a whale! Every detail of the process is on a scale so much larger than anything heretofore attempted that the setup is already somewhat appalling to contemplate.” When would the telescope be ready? a reporter asked. “So far as astronomy is concerned,” Day answered, “the existence of a 200” disk will remain a dream until such a disk emerges from the annealing furnace at ordinary temperature in one piece and free from strain. After that I am at your service for any account of the disk or the manufacturing operation you may wish to publish.”

  Corning—the company and the town—had not been accustomed to the attention of the outside world. There had been occasi
onal hoopla when a new product, like the first Pyrex utensils, was announced, but marketing publicity was very different from the persistent questions of newsmen. When a reporter called Corning for details on Day’s comments, Leon V. Quigley, the newly appointed director of publicity for Corning Glass, dutifully explained the status of the project, the successful casting of the 120-inch mirror, and the preparations for the casting of the two-hundred-inch mirror. When the time came, he explained, Corning glassmakers would ladle the 20 tons of glass into the mold to create a two-hundred-inch-diameter, 26-inch-thick disk of Pyrex for the telescope. The reporter’s story became one more newspaper item, lost among the reports of Roosevelt’s frustrated efforts to deal with the deepening depression, Hitler’s first stabs into the maelstrom of European politics, and the morbid daily details of the Lindbergh kidnapping.

  The story of the mirror might have remained buried in the news if an NBC researcher hadn’t picked up the item and put it into a script that was sent over to Lowell Thomas’s office in the new Empire State Building.

  In 1934 America listened to the radio. Whole families gathered around a console in the parlor, or perhaps a tabletop unit in the kitchen, in the hours after supper. The immediacy of radio meant that for the first time, an entire nation could be focused on a single program, event, or news item. Rich and poor, black and white, men and women, recent immigrants and Mayflower descendants—all heard the same broadcasts.