In the aftermath of Columbia, NASA engineers learned something they already knew: the foam insulation on the shuttle's external tank flakes away on every flight and poses a threat to the orbiter's heat shield. The phenomenon was not viewed as a major threat, however, and seven astronauts lost their lives.

The first and most obvious fix was to remove the bipod ramp foam that caused Columbia's destruction. The foam was in place to prevent ice from building up before takeoff on fittings where two struts connecting the shuttle's nose to the tank attach. NASA solved that problem by installing heaters on the fittings and removing the foam altogether.

Going into Discovery's flight last year, engineers knew foam on another area of the tank - the so-called protuberance air-load, or PAL, ramps - posed another potential risk. The foam is applied by hand, on top of robotically sprayed insulation, and then sculpted to form long wind deflectors. NASA kicked off a program to redesign or eliminate the PAL ramps but decided to launch Discovery last summer with the old design.

Again, the foam surprised then. During launch last July 26, a one-pound chunk of foam broke away from the hydrogen PAL ramp. While it did not hit anything, it represented a problem that had to be fixed before flights could resume.

Then came Hurricane Katrina, which devastated New Orleans and damaged Lockheed Martin's Michoud Assembly Facility, where the huge tanks are built. Struggling to help the contractor work force and complete testing and analysis, Hale decided late last year to remove the PAL ramps before the shuttle flew again.

"We knew the PAL ramps were a hazard and we had a plan that over the course of about two years we were going to do a very orderly, systematic and thorough job and remove the PAL ramps from the external tank," Hale said. "We accelerated that plan. That was the most hazardous piece of foam we had on the outside of the tank. We decided to fix it and we accelerated that work."

He made that decision on the assumption that computer modeling and wind tunnel data would prove the external brackets, pressurization lines and cable tray could stand up to the expected aerodynamic buffeting in the absence of the protective ramps.

"There were over 100 different piece parts on that tank, mostly brackets, structural items such as the cable tray, the pressurization lines that run on the outside of the tank, that were affected aerodynamically ... by the removal of the PAL ramp," Hale said.

In the end, after months of testing and analysis, the results showed the tank was, in fact, structurally strong enough.

"We have made the largest aerodynamic change to this vehicle that has been made since we started flying 25 years ago, removal of what we call the PAL ramp," Hale said. "We had to verify that the components, the metal components that were protected from the aerodynamic flow by this ramp were still capable of withstanding those loads. We put together an intense campaign of wind tunnel tests, computational fluid dynamic runs and other calculations to provide updated force environments for the structural analysts.

"The structural analysts have done thousands of calculations in the last two-and-a-half months and were able to report that in all areas they can show that we meet the required factor of safety and we are safe to fly structurally without those two large pieces of foam on the outside of the tank."

In fact, the structural engineering community "demonstrated conclusively that we met what we call our 1.4 factor of safety," Hale said. "So in the worst case load on the weakest part, we have 140 percent of the strength that you need to ensure that your structure holds together."

Sensors have been mounted in the cable tray to measure the actual stresses on the external fittings during launch.

The external tank is covered by some 4,000 pounds of lightweight foam insulation. It's there to prevent ice from forming in the humid Florida weather when the tank is loaded with supercold liquid oxygen and hydrogen. It also helps minimize how much propellant boils off in the tank before launch.

About 75 percent of the foam is sprayed on robotically in a tightly controlled process. NASA's flight history shows that foam tends to stay in place. But about 25 percent is manually applied, including the 35 pounds or so that made up the now-removed PAL ramps and the still-in-place ice-frost ramps. It is that manually applied foam, sprayed on top of existing foam, that appears to be the most dangerous.

The most significant concentration of manually applied foam left on the tank after removal of the bipod and PAL ramps is the foam insulating the 37 brackets used to hold the pressurization lines and cable tray in place.

At first blush, the IFR foam would not appear overly dangerous. In flights where photography of the tank is available, the largest piece of missing IFR foam weighed an estimated .09 pounds. But analysis indicates pieces weighing up to .2 pounds could come off and if one did, at the worst possible time and from the worst possible place, catastrophic impact damage could result.

Complicating the picture is analysis of a tank that was dissected last year in the wake of Discovery's first return-to-flight mission. Engineers found underlying cracks in the insulation where foam was applied on top of pre-existing insulation. That means there is a possible failure mechanism in place with the ice-frost ramps on Discovery's tank.

"We have data that shows we have potentially cracks underneath large foam, or foam that's put on top of other foam," said Bill Gerstenmaier, NASA's associate administrator for space flight. "And we have flight history that doesn't show that we lose a lot of foam. ... So the dilemma is how can we not rule out that at some point in the future we're not going to have some larger foam loss with this underlying problem? And therein lies the debate.

"We can't figure out the theory that can explain to us why we haven't had larger foam loss with this underlying crack. Obviously, there's something we think that's protecting us in the physics of the situation but we don't know what that is."

And so, while engineers believe .2 pounds is the upper limit on IFR debris, they cannot prove larger pieces are not possible. As a result, NASA kicked off a program to redesign the ice-frost ramps and briefly considered modifying Discovery's tank before flight.

But during a wind tunnel test, a large chunk of the redesigned ramp blew off. The old design, subjected to similar worst-case conditions in the wind tunnel, came through with only minor shedding. Griffin, Hale and other senior managers approved shipping Discovery's tank to Florida with the old design.

O'Connor and Scolese, however, opposed the move then and registered a more formal complaint when they voted no-go for launch during a flight readiness review at the Kennedy Space Center earlier this month. The IFR foam remains classified as "probable/catastrophic."

"One of the things I think that was wrong before Columbia is we didn't pay enough attention to some of the problems we had, in particular foam problems," Hale said. "We pay a lot of attention to those problems today and we're not making any decisions on an ad hoc or a schedule-driven basis. We're taking a very hard engineering look at these things. I've said it before and I'll say it again, if we could go back 25 or 30 years and change the design of the shuttle, the number one thing I would say we needed to do was not put our fragile heat shield in an area where debris can come off during the first three minutes of flight and cause us catastrophic damage.

"But it's the vehicle we've got and it's the only vehicle we're going to have for the next several years. So if we intend to be a space-faring nation and put Americans into space on an American-flagged vehicle, it's going to be the space shuttle. Our job is to make that as safe as we practically can. It's not going to be perfectly safe."

Hale agrees NASA should redesign the ice-frost ramps. He even agrees with the "probable/catastrophic" classification. But he believes NASA should test the PAL-free tank in flight before making another major change to the insulation.

"We think we could lose up to two tenths of a pound of foam," he said. "Off Columbia, we lost a pound and a half, 1.6 pounds. Off the PAL ramp off STS-114, we lost a pound. We've eliminated both of those. The next largest piece we can lose, the biggest we've ever seen is .09, less than a tenth of a pound. We think statistically we might lose two tenths of a pound even though we've never seen that. So you're looking at smaller and smaller pieces of foam that come off less and less frequently. We're on a continuous improvement path.

"The real question is on this particular flight, why would we not stand down and fix all the ice frost ramps? There are a couple of good reasons. One is we don't have a good design in hand today. But secondly, there is a principle in flight test that says when you have made a major aerodynamic change and you have studied it in the wind tunnel and you've studied it with computer simulations and you think it's good to go fly with, you go fly with that one change. Because all those wind tunnels and all that analysis and all those computer models are never quite as good as real life. You really need to go fly it and see whether it works like you think it was going to and get that data before you proceed to change anything else.

"It's that flight test principle that we're resting on here today," Hale said. "It is nice to say you ought to stand down and fix all your problems. If we adopted that philosophy we would never fly. We think we have identified the most critical problem, we've fixed that, we need to go flight test to that. We have additional sensors this time, we have additional cameras this time that will bring us back information on how well that worked, and once we feel comfortable that we have accommodated the PAL ramp successfully, we're going to move onto the ice-frost ramp redesign."

Engineers are hopeful they can eliminate most, if not all, of the IFR foam by replacing the aluminum brackets with a different material, one less susceptible to thermal effects. Titanium is one candidate. For tanks already in the pipeline, different ramp designs might improve safety.

Gerstenmaier said he hopes to have a new IFR design ready to fly by the end of the year. But that depends on the results of more analysis and wind tunnel testing.

Despite the controversy generated by the objections raised by O'Connor and Scolese, many current and former space agency officials agreed with the decision to press ahead.

"I trust the judgment of shuttle program leaders Bill Gerstenmaier and Wayne Hale," legendary Apollo flight director Gene Kranz wrote in an op-ed piece carried by Florida Today. "I brought them into the business of mission control, trained them and watched them grow to shoulder great responsibilities. I have trusted these men for more than a decade and found their risk judgment superior. I trust that they made the proper risk judgment when they committed to the Discovery launch."

Time will tell.

PREVIEW REPORT PART 4 --->