Book Read Free

The Perfect Machine

Page 24

by Ronald Florence


  Ellis had the shop fabricate a brick of recrystallized carborundum, the only material that would withstand the extreme heat inside the furnace, fastened to a metal bar wide enough to run across the top of the furnace chamber. Men in protective clothing, pressing down at each end of the bar, could exert a force of two hundred pounds on the brick to flatten the sprayed surface. Like most experiments in the deadly heat of the furnace, it didn’t work. If the brick was as much as six inches from the flames of the spraying torch, the surface was already cooled enough to be rigid. If they got closer to the nozzles that were spraying the quartz, the brick took the full brunt of the flames. Chunks of the carborundum would chip off and embed themselves in the quartz surface. The resulting imperfections were worse than the ridges.

  Despite the continuing problems, the GE publicity department was still sending off press kits, including a major release that found its way, almost verbatim, into the science section of the New York Times in mid-April. The article crowed that the successful fusing of the sixty-six-inch disk was undoubtedly the biggest step in the production process, with the next jump to a two-hundred-inch telescope only a magnitude of 22 to 1, after the leap of 225 to 1 in magnitude of the previous step. The GE press release and the article never mentioned that the disk was severely cracked and not usable.

  After Mason and Noyes reported what they had seen and heard in West Lynn, confirming Hale’s worst suspicions, Hale wired Thomson ordering that all work on the sprayed quartz mirrors cease on June 1.

  Thomson, semiretired and involved in a dozen other projects, had lost interest in the project, but Ellis was reluctant to give up. As the June 1 deadline approached, Ellis reported that he had conducted tests on new experimental disks, using an interferometer to measure deflection of the disk under pressure on the two axes. The device was so sensitive that he could only run the tests between midnight and five in the morning on a weekend, when the other machinery in the factory was shut down and there was no traffic on local streets that might introduce a measurable vibration. Ellis’s measurements, accurate to five millionths of an inch, confirmed that deformations of the disk on either axis were minimal.

  “It is also evident… that a further development of the burner and spraying equipment is necessary to produce a surface layer,” he wrote. He hoped, but couldn’t promise, to have the additional work done by June 1, at which point they could spray another sixty-inch mirror, this one usable for the telescope and the final prelude to beginning work on the two-hundred-inch mirror disk. Trying to hang on to his lab when there was little work around, Ellis pleaded:

  We have been using a large part of our laboratory force directly on this mirror work day and night, without regard to union hours or personal inconvenience, and to the exclusion of other work, having taken on no new problems while the mirror work was in hand. As a result of this condition and the business depression, to comply with your telegram means that we will have to discharge the men on June 1. If this is done it will be very difficult to take up the work again months later should you then wish to go on with it. I am stating this case frankly, believing that you might have thought we could stop all work at a given time, place the men on other work while a review of the whole situation is being made, and pick up the work where we left off at some later date.

  It was a sound argument, but too late. Sometime in the midst of the promises and frustrations, Hale had lost faith in GE, Thomson, and the whole fused-quartz project. It was no longer clear that the two-hundred-inch telescope needed fused quartz for the mirror, and the cost estimates, once reasonable, had escalated to the point where the budget of $6 million for the telescope no longer seemed adequate to keep up with the GE bills. After reading one report from Ellis, Hale asked Anderson whether—if Ellis did find a reliable method for selecting the quartz—the cost of “merely selecting enough quartz for a 200-inch mirror,” would come within the budget.

  Hale, Robinson, Millikan, and Noyes held another long meeting, this time to draft an ultimatum, over their four signatures, to Gerard Swope, the president of GE, reminding him that the original estimate was for the first sixty-inch disk to cost $75,000 and be ready in May 1929, and that after three years’ work and $600,000, “We are not yet in position to know with any certainty whether a sixty inch disk can be successfully made and we have no definite information as to cost.”

  Swope telegraphed back immediately: “As far as we are concerned, this is an adventure into the unknown and outside our line of business. We have been giving our facilities and the best brains we have to this work to the neglect of other work and without any profit or hope of building up a business along this line.”

  Mason urged Hale to give GE the go-ahead for another disk, even at the estimated cost of $60,000 to $80,000, so they could see some return—if only a single disk—from the money that had been spent at GE. Hale thought more work at GE was throwing good money after bad, but he authorized the second try at a sixty-inch disk. “I don’t see how we can go further unless some miracle occurs.”

  Ellis immediately started spraying. The base portion of the new disk went well. For the surfacing Ellis raised the disk on a platform of crushed quartz fragments mixed with binder clay to reduce the risk of cracking. With the familiar delays for breakdowns and repairs to leaking pipes and burners, Ellis completed the second sixty-inch disk before the end of June.

  The disk emerged from the furnace uncracked. Thomson triumphantly wrote: “We fully believe that the completion of this 60” disk will point the way most assuredly to the building of the 200” disk of fused quartz.” Later Ellis reported that the new disk was filled with “a good many bubbles” varying in size from a few thousandths of an inch up to a tenth of an inch in diameter; he was convinced that they would not interfere with the intended purpose of the disk. Ellis also suggested that they could saw the earlier sixty-inch disk in half. The cracked half could be “welded” together, and the two blanks could then each be surfaced. GE never stopped producing great ideas.

  To the council’s ultimatum, he had no satisfactory answer. “We cannot know that a 200-inch mirror can be made until it is an accomplished fact. It is therefore impracticable to give you a ‘guaranteed price for a satisfactory 200-inch,’ and if no other solution will be satisfactory to the Council we will have to shut up shop.”

  Hale wrote to a friend, “The way they swallow money without blinking is not good for one’s nerves! We have just reached the end of this rope and must try another…. The financial depression has complicated the situation—we are fortunate in not completely losing the support of our backers. But perhaps you will partially realize why I have not known what minnit’s gwine to be the nex’, as Uncle Remus puts it.”

  Even if it had emerged from the furnace flawless, the last disk came too late. “It is evident that you have accomplished an important technical achievement,” Hale wrote Ellis. “And that if further tests, including optical ones, prove the disk of suitable quality, it can be used for some purpose, such as solar work, where a glass mirror would not serve…. if we had not already spent the huge sum of $639,000, and if the estimates for the larger mirrors were not so far beyond our means, we should certainly wish to proceed at once with a larger disk.” Under the circumstances, “We cannot be responsible for any further expense incurred.”

  15

  New Light

  The Corning Glass Works was a proper company. George McCauley had been ready to make mirror blanks since 1929, the company was eager for new work, and there were rumors that GE was having difficulties producing satisfactory disks for the Caltech telescope project—but no one from Corning formally approached Hale about the two-hundred-inch-telescope project while GE was working on the fused quartz.

  Yet even good business manners couldn’t keep the old boys of the University Club from talking. The Reverend Anson Phelps Stokes put in a word to his friends at the Rockefeller Foundation: “I believe that my friend, Mr. Houghton, our former Ambassador to England, is the head of the C
orning Glass works.” At a meeting of the National Association of Science in the spring of 1931, Robert Millikan got into a conversation with Arthur Day, a vice president of Corning and a respected scientist with the Geophysical Laboratory of the Carnegie Institution of Washington. Millikan told Day about the problems GE was having with the fabrication of a sixty-inch disk. “If we could get a good Pyrex disk from you for $20,000,” Millikan said, “It would be a mere drop in the bucket in our expense account and would serve to guarantee us against the consequences of a second failure by General Electric.” Day realized that Millikan had done his homework: Twenty thousand dollars was exactly the figure Corning had quoted the University of Michigan for a sixty-inch disk.

  Day had followed the big telescope project for years. He was widely respected as an expert on glass and would have served on the mirror committee if he hadn’t had an earlier falling-out with GE. Day believed that they had stolen his own process for making fused quartz. When he made his initial agreement with GE, Hale had asked Day’s opinion of the possibility of a Pyrex disk for a large telescope. “The plan is entirely beyond any experience available from here or elsewhere,” Day answered. “Courage is an essential asset to be reckoned with.”

  Now, after three years of unsuccessful experiments at GE, Hale wrote Day again, reporting the dismal progress at West Lynn and the recent good news about the mirror-support-system experiments on the one-hundred-inch telescope. Theodore Dunham, of the Mount Wilson Observatory, had gone over Pease’s test figures and found the differences with the new mounts “little short of revolutionary.” It was clear that a Pyrex disk would have a temperature coefficient well within the range needed for the two-hundred-inch telescope. Hale asked Day straight out: Could Corning produce a disk for the telescope?

  Day answered that the Corning Glass Works officially and its laboratory staff individually had been interested from the beginning in the project to cast a two-hundred-inch mirror disk. “I think now, as I have always thought, that they could successfully make it out of glass of Pyrex type.”

  The problems at GE were open gossip at meetings of the National Academy of Sciences and the AAS. Day may also have heard about or guessed at the mostly unspoken differences in attitude between GE and the Observatory Council. The attitude at Corning, he assured Hale, would be different: “We should not overlook the fact that the making of one great disk in Corning would be exclusively a task for the Research and Development Department and not a manufacturing problem.”

  One thing held Corning back: “You have no wish or use for more than one disk and you could hardly wish to tax your resources unnecessarily with the development costs of two different varieties of these.” Corning didn’t mind secret flirtation, but it remained a proper front door suitor, unwilling to get into extra contractual affairs. Day was making it clear that Corning wouldn’t begin work on a disk until the Observatory Council broke off its relationship with GE.

  In fall 1931, with his nerves in one of their increasingly rare periods of remission, George Hale took a train east to New York. “We mean to succeed,” he told Arnett at the Rockefeller Foundation, “especially as the importance of the 200-inch has become far greater than originally appeared. This is the result of our recent work, as viewed by Einstein and [James] Jeans, who count on the 200-inch to give the key to their most vital problems.”

  Ellis made a final desperate try to keep the work at GE. He told Swope that GE’s original quotes had been misunderstood because Porter, who visited West Lynn as Hale’s representative, was hard of hearing. When Hale reported that the Observatory Council was considering at least two other methods of making mirrors, Ellis wrote, “I cannot believe they are seriously considering the grave uncertainties of any other methods, in view of the progress we have made and the money they have already sunk in this project.” Despite his efforts, even Ellis seemed to know that the battle was lost. The GE publicity department in Schenectady was about to unleash a new salvo. Ellis urged that we “use our powder in celebrating the first 60-inch and leave the 200-inch until history has been made.”

  Ellis carried his campaign around Hale, going out to Little Compton, Rhode Island, where Max Mason was on vacation for Labor Day. He told Mason that Hale’s demands were impossible. What would your best price for each disk be? Mason asked. Ellis said GE simply couldn’t commit to a firm price. The next day, Hale and Mason met Ellis at the University Club in New York. Together the three went out to Gerard Swope’s country house near Ossining. Swope had prepared for the meeting by discussing the matter with C. E. Eveleth, chief engineer of GE’s Schenectady works. Swope’s best estimate was that the cost of a quartz disk would be an additional $1 million, perhaps more. Given the experimental nature of the work, GE could not guarantee that it would be able to produce the disks at that price. Walter Adams later estimated that if they extrapolated the rate of expenditure at GE during the spring and summer of 1931, the total cost for the mirrors would be closer to $2.5 million, almost half the budget for the telescope. Everyone at the meeting knew it was a formal closure: a final polite lunch before the inevitable divorce.

  Hale continued his round of courtesy calls by going up to West Lynn to talk personally with Elihu Thomson. The two men had known one another for more than a quarter century. They were titans in their fields, Thomson the inventor who could produce solutions for any problem, Hale the grand old man of solar physics and the builder of machines to see to the edge of the universe. For years, at the annual meetings of the AAS or the National Academy of Sciences, they had talked of pooling their skills, and most outside observers would have bet on the success of a collaboration. Neither man was accustomed to failure. Thomson had a reputation for doing what the scientists said couldn’t be done. He had taken on much tougher challenges than this one, and had succeeded. If only Caltech would commit the funds, he no doubt told Hale, GE would produce the mirrors.

  For the record, they later exchanged stiffly formal letters, congratulating and thanking each other for the contributions to science and technology that the experiments with fused quartz had provided.

  Five days after he met with Swope, Hale held a meeting at the University Club in New York with Mason and three representatives from the Corning Corporation: O. A. Gage, the head of the Optical Glass Division; J. C. Hostetter, the director for special projects in the division; and George McCauley, the chief scientist. McCauley summarized Corning’s experiments on telescope mirror disks. The variety of Pyrex that seemed the most promising had a coefficient of expansion three times that of quartz but one-third that of plate glass. The only potential problem McCauley foresaw was in annealing the glass—not that it couldn’t be done, but that no one had ever tried to anneal large masses of glass. Hale was impressed with the no-nonsense technical confidence of the Corning people, especially McCauley.

  A week later, on October 14, the group met again, this time joined by Dr. Day and by General Carty of AT&T, still one of Hale’s most trusted advisers. The Corning group was ready with samples of Pyrex glass, showing the kinds of surface that would emerge if different powdered materials were used to coat the refractory bricks of a mold before the glass was poured. Corning also presented a budget and a timetable—exactly what GE had never been able to deliver. Hostetter, McCauley’s boss in the research division at Corning, laid out the budget. The minimum cost for three 60-inch disks, one 60-by-80-inch oval disk, a 120-inch disk, and the two-hundred-inch disk was $150,000. Unanticipated difficulties or repeated castings could push the figure as high as $300,000. The use of surplus equipment from the GE effort in West Lynn, or a disk design that would require less glass, such as a hole in the center or ribbed backs, would reduce the cost. He estimated that Corning could finish the mirrors in a little more than a year and a half from the time they got the go-ahead. If a second run were necessary for each mirror, the time might stretch to as long as thirty-one months.

  The division of responsibility was as clear as the budget: Corning would expect Hale and his group to
work out the optimum design for the disks, taking into account “best glass practice.” Corning would take care of annealing and would investigate shipping possibilities.

  The cool decisiveness of the Corning people impressed Hale and his colleagues. And McCauley liked Hale’s idea that the each disk should be a model for the next. As a start he proposed that Corning would mold a 26-inch solid disk, then a 30-inch disk with a ribbed back like the designs that had been discussed at Caltech but never tried on a mirror. Once the 30-inch disk was ground and polished into a mirror, Corning would proceed to a series of Pyrex disks, a 60-inch, a 120-inch, and the two-hundred-inch—each based on the experiences with the previous ones. If the two-hundred-inch could be successfully cast, there would be further orders for additional 60-inch mirrors and other special-purpose mirrors. Hale agreed that if the project were to go ahead, the final price to the Observatory Council would be computed on a “cost plus” basis. The “plus” was a 10 percent profit that Corning would bill only after the disks were figured into mirrors and accepted by the Observatory Council.

  Hale raised one other topic at the meeting. If the Observatory Council were to go ahead with an order for Pyrex disks, he said, they wanted no publicity of any kind. The failure of the quartz project was an embarrassment. More than $600,000, one-tenth of the entire telescope budget, had been expended on experiments that had not produced a single satisfactory mirror. At a time when eight million Americans were unemployed and the median income for an American family was $1,231 per year, $600,000 was a big expenditure to justify. Hale, with the reticence of a scientist, had resented the stream of press releases from GE. The research necessary to build the telescope, like the observations of an astronomer, were raw data, scientific work in progress, not fodder for the publicity mill.

 

‹ Prev