More than three years after Columbia disintegrated over Texas, the shortcomings of the space shuttle's external fuel tank have made the nation's space program the hostage of an exasperating piece of hardware on the threshold of obsolescence.
Fourteen times since Columbia went down Feb. 1, 2003, NASA has postponed shuttle launches, and in all but one case engineers cited problems with the external tank as the main reason. Hurricane damage to NASA facilities accounted for the other.
NASA flew Discovery last year after more than two years of redesign and modifications, only to ground the fleet immediately after the tank — a 154-foot orange torpedo that holds half a million gallons of liquid oxygen and liquid hydrogen — shed unacceptably large pieces of foam insulation during liftoff.
Since then, the tank's vagaries have caused shuttle planners to postpone the next mission five times. Engineers are currently forging ahead on two redesigns of the insulation and tests of swapped-out tank components in hopes of making a July 1 launch date, and John Chapman, manager of NASA's external tank project, said in a telephone interview that "things are so far looking good."
So far. The loss of insulating foam is the external tank's basic flaw, known since the first shuttle flight in 1981 but largely ignored until a 1.67-pound piece of it punched a hole in Columbia's thermal tiles during launch, causing it to burn up on reentry.
Since then, NASA engineers have chipped away at the problem, but they acknowledge that it can never be truly fixed as long as America's signature spaceship is mounted alongside its volatile fuel supply, where foam can hit it.
"We have a unique design," N. Wayne Hale, shuttle program manager, said in a telephone interview. "Maybe I wish we hadn't done it this way, but it is what it is." The shuttle "will be an experimental vehicle until the last time we fly it," said Hale, who is joining NASA Administrator Michael D. Griffin to brief reporters today on NASA's efforts to ready Discovery for a July launch.
Dealing with the tank has exacted a price. Between 2003 and the end of 2005, NASA officials estimated spending $1.3 billion on post-Columbia "return to flight" expenses, many of which involved the external tank.
The consequences for the space program as a whole, however, are much grimmer: Between the Columbia tragedy and Sept. 30, 2005, NASA spent about $10 billion on the shuttle -- 26 percent of its total budget — and flew just once.
This year, NASA has $4.8 billion for the shuttle and plans to fly twice. It hopes to fly 17 times in all before retiring the orbiter in 2010 and replacing it with a new spaceship. NASA, supposedly the most forward-looking agency in the federal government, is using the biggest single chunk of its budget to fund a dinosaur.
But the agency has no choice: "Conventional wisdom before Columbia was that foam can't be a problem because there's not enough mass," Hale said. "But the people who thought about it didn't think hard enough. Think, instead, about straws driven into a board by hurricanes."
NASA troubleshooters have solved individual tanks' problems but have failed to find a comprehensive solution. Discovery's foam loss last year was unforeseen and unrelated to the casualty that doomed Columbia.
"Any problem can be solved, given the resources and time," said engineering consultant Charles C. Daniel, a member of the Return to Flight Task Group that oversaw NASA's efforts to revamp the shuttle program after Columbia. But, he added, "it may be a problem for which the cost of the solution does not justify the time it will take."
The cost is beginning to bite. Late last year, Griffin told Congress the shuttle program faced a $3 billion to $5 billion shortfall between 2006 and 2010, and this year he has begun to sweat the rest of NASA's budget to fill the gap.
Griffin has incurred the wrath of many in Congress and the scientific community by trimming the growth of NASA's unmanned exploration budget by $3.1 billion between 2007 and 2010. This has proved hard to justify, for the limping shuttle program does not show well alongside NASA missions that in the past 2 1/2 years have landed rovers on Mars, put a spacecraft in orbit around Saturn, crash-landed a projectile in the face of one comet and brought home dust from the tail of another.
"They're cutting everything in long-lead science and engineering," said Charles M. Oman, director of the Massachusetts Institute of Technology's Man Vehicle Laboratory. "We've got a large number of research applications, but the funding is so down we're only able to offer assistantships to half a dozen people. We offer 50 most years."
Still, Oman acknowledged that Griffin "is in a box." To reinvigorate the nation's human space flight program, President Bush in 2004 introduced the "Vision for Space Exploration," designed to build a new infrastructure capable of returning to the moon and eventually traveling to Mars.
The plan's cornerstone is a next-generation spaceship to replace the shuttle, but while its development proceeds, Bush's plan also requires NASA to fulfill treaty obligations by completing the international space station.
For this, NASA must have the shuttle, because only the shuttle is big enough to lift the station's remaining components into orbit. And to have the shuttle, NASA must deal with the external fuel tank.
And do it in time to finish the station by 2010. When Bush announced his plan, some experts questioned the wisdom of retiring the shuttle program so quickly, but that criticism has waned, because "every month the shuttle is delayed creates more financial strains," Oman said in a phone interview.
The original shuttle design, conceived in the 1970s, called for foam only on those tank sections that held liquid hydrogen. Without insulation, the hydrogen, cooled to minus-423 degrees Fahrenheit, would boil off even as engineers attempted to fill the tank. The liquid oxygen part of the tank, at minus-298 degrees Fahrenheit, would have no insulation.
That scheme, however, would have allowed Cape Canaveral's humid air to form ice on the tank's skin while it sat on the launch pad, and once planners had produced the final side-mounted orbiter design, ice was forbidden. Shuttle engineers warned that "an ice cube dropped four inches could crack the coating" on the thermal tile that protected the orbiter's underbelly against the heat of reentry, said Myron A. Pessin, NASA's former external tank chief engineer and a tank expert since the dawn of the shuttle program. To control ice, the entire tank had to be blanketed with foam.
And it worked, except that the tank shed foam, scarring or gouging the orbiter's heat shielding on every mission. The Columbia Accident Investigation Board documented an average of 143 debris "hits" to the orbiter on every launch, but with no consequences, Pessin said, engineers came to ignore foam loss as a "maintenance issue."
NASA's shortcomings in this regard were well vetted in the aftermath of Columbia. What is not as thoroughly understood are the reasons why problems have continued.
Hale noted that foam is a "relatively new product" with mechanical properties that are "not very well studied." NASA, after neglecting the foam for two decades, is doing basic testing about how it behaves during the stresses of a launch.
Also, Chapman noted, the external tank, unlike the orbiter, is expendable. It burns up or falls into the ocean after separating from the orbiter, so "we can't look at it after it flies." Until engineers installed new sets of sensors on the tank after Columbia, "we could only assess its performance based on ground testing, computer modeling and photographs," Chapman said.
But Daniel, one of seven "dissenters" in the task group who criticized NASA's performance in the aftermath of Columbia, said engineers have also been too optimistic.
"NASA did not understand the difficulty of the problem, and they've had a very bad run of luck," Daniel said. "They were gathering data at the same time they were trying to fly. If your solution happens to be right, you can cut the time, but unanticipated nuances led them down blind alleys."