Space: The Trouble with Hubble

The failure to properly shape the Hubble Space Telescope's mirrors offers a sharp view of something that is often quite fuzzy—the inner workings of government. The Hubble trouble is a chance to judge a government-funded activity by standards of success that are grounded in the experience of the real world.

The Hubble telescope had for years been hailed as the instrument that would stand at the pinnacle of astronomy, offering the clearest and sharpest views of the faintest and most distant objects in the universe. To achieve such goals, at a cost of $1.5 billion, its design featured the most exquisite precision ever achieved in polishing large mirrors, along with a dazzling array of the most advanced electronic instruments.

But the Hubble failure did not involve the high-tech parts of the system, in which the risk and difficulty were greatest. It involved a straightforward but crucial matter: the overall shape of the main mirrors. These are in a configuration known as Cassegrainian, which astronomers have used for more than a century.

In this arrangement, light from a distant galaxy falls on a gently-curved "primary mirror," which in the Hubble is 94.5 inches across. This mirror's shape resembles that of a large and shallow saucer with a hole in the middle. The light reflects from this primary mirror and falls on the secondary mirror, 12 inches across in the Hubble, which bulges outward like a cake in its baking pan. Then, reflecting in turn from this mirror, the light passes again down the telescope's length and through the hole in the primary mirror, coming to a sharp focus. This arrangement makes it possible to point and operate the telescope with ease.

The problem is that one of these mirrors was formed in the wrong shape. It is very much as if your optometrist were to give you the wrong eyeglass prescription. Such an error, of course, should have been found in ground testing, and the faulty mirror sent back to be reshaped. But the proper tests weren't performed.

Why not? The answer is not clear. There appears to be a good deal of technical similarity between the Hubble telescope and the nation's large military reconnaissance satellites. One looks up while the others look down, but both types of spacecraft have been built by Lockheed Missiles and Space Co. and have featured high-power telescope elements built by the Perkin-Elmer Corp. (now a division of Hughes). Aviation Week and Space Technology, the industry weekly, reports that the Air Force offered NASA the chance to test the Hubble mirrors using its spy-satellite facilities, but NASA turned it down on the grounds that the people working on Hubble would have had to get top-secret clearances.

To reproduce such test facilities would have cost $100 million or more. But within a budget of $1.5 billion, that might have been the proverbial horseshoe for want of which the battle was lost. And some telescope experts say a simpler, far less expensive test would have detected the flaw. Eastman Kodak Co., which bid unsuccessfully for the project, proposed a final-assembly test that would have cost about $10 million.

What is indisputable, however, is that other organizations have not had such problems. In size, shape, and arrangement, the Hubble mirrors closely resemble those of the 100-inch telescope on Mount Wilson in California—which went into operation, successfully, in 1917. They are much smaller than those of the 200-inch Mount Palomar telescope, which began operating in 1948. The Mount Wilson instrument was built at such an early date that its elements had to be carried up the mountain by mules. But that was not too early, nor was Mount Palomar too large, to preclude successful testing prior to construction.

Moreover, the past quarter century has seen an enormous upsurge in telescope building. New instruments have featured mirrors nearly as large as Palomar's. And no astronomer has had to complain that any of these instruments didn't work properly because the engineers, in effect, had worked from the wrong blueprint.

It is true that the Hubble mirrors are far more precisely shaped than these earlier ones and feature automatic equipment for exact focusing. But these niceties are not the source of the problem. The knowledge necessary to design proper Cassegrainian mirrors is presented in such standard references as the Amateur Telescope Making series, which can be found in any good university library.

The loss of Hubble may justifiably be seen as merely the tip of NASA's iceberg. The agency has received a great deal of criticism: for a space shuttle fleet that costs too much and underperforms, for increases in the time and expenditures required for major projects. But it has been hard for critics to present well-defined critiques rather than mere complaints. After all, no other organization is flying a shuttle fleet or launching the largest of spacecraft. This lack of alternatives to NASA precludes standards by which the agency's performance might be judged.

The Hubble problem is different. It involves a straightforward technical issue that other outfits have successfully dealt with, but which NASA blew. In this respect it is reminiscent of the Pentagon's $1,000 toilet seats. The Pentagon, like NASA, has long operated on a scale far exceeding that of corporations or of any other organizations that could provide standards for comparison. And like NASA, it too has faced frequent complaints about costly programs and equipment that underperforms.

Defense Department officials, like their counterparts in NASA, have been able to maintain that they face special requirements that demand high budgets. Without standards for comparison, critics of the Pentagon had often been reduced to mere grumbling. Then came the gold-plated toilet seats, along with similar disclosures of Donald Trump-like prices for ordinary hammers and other items. Suddenly critics had concrete examples on which to hang their charges.

Now it is NASA's turn. The agency's heart's desire, at least for now, is the space station—which William Broad of the New York Times declares will cost $120 billion to build and operate. A recent NASA study found that the station would require 3,800 hours of maintenance each year, compared to the original estimate of 130. Beyond this, NASA's eyes are clearly focused on expeditions to the moon and Mars, at the cost of the savings and loan bailout—$500 billion. NASA's supporters say we must do these things in the name of exploration, in the spirit of Columbus and Magellan and Captain Cook.

To which a suitable riposte might now be: At least Captain Cook had a telescope that worked.

Contributing Editor T.A. Heppenheimer is an associate fellow of the American Institute of Aeronautics and Astronautics.