NASA originally hoped to launch Discovery on Nov. 1, but the flight was delayed to Nov. 5 by bad weather and technical problems.

The ship was fueled for launch Nov. 5, but the countdown was called off when sensors detected a gaseous hydrogen leak in a quick-disconnect fitting used to attach a vent line to the side of the tank. That problem was quickly resolved, but engineers also discovered cracks near the tops of two adjacent structural rib-like "stringers" near a flange in the external tank that supports the upper liquid oxygen tank. More cracks were found later, after Discovery was returned to NASA's Vehicle Assembly Building for more detailed inspections.

Tracing the manufacturing history of the tank, engineers discovered that most of the original 108 stringers in the "intertank" section that separates the tank's oxygen and hydrogen sections were made using an aluminum-lithium alloy from a lot that was more brittle than usual and more susceptible to temperature-induced fractures when the tank is loaded with super-cold propellants.

That weakness, engineers concluded, along with subtle manufacturing issues, were the most likely causes of the observed cracks.

In the most invasive repair of an external tank ever attempted, engineers installed structural stiffeners, called radius blocks, to better secure the top few inches of 105 stringers to prevent cracks from forming when the tips of the stringers are pulled inward due to the contraction of the liquid oxygen tank.

The goal was twofold: to ensure the tank's overall structural integrity and to prevent cracks that could cause foam insulation to pop off during the climb out of the dense lower atmosphere when debris impacts pose the greatest threat to Discovery's fragile heat shied.

The addition of the radius blocks more evenly distributes stresses in the stringers and reduces the strains on the top few fasteners.

The tank was designed to be structurally "fail safe" even if three adjacent stringers developed cracks. Data from a Dec. 18 fueling test, elaborate laboratory tests and an exhaustive engineering analysis showed the tank would remain structurally sound even if more than three stringers developed cracks greater than four inches in length as long as they were separated by undamaged stringers.

The tank's calculated factor of safety is not precisely known, but it is greater than one in all cases, meaning that even with extensive modifications and stringers that are more brittle than usual, the structure's strength is sufficient for flight.

Testing also showed cracks greater than four inches long were most likely to occur during fueling, when the hardware is "shocked" by extreme low temperatures and that any such fractures likely would cause visible cracks in the tank's insulation. With the radius block modifications, engineers do not expect such cracks to form. But during Discovery's countdown, cameras will be focused on the liquid oxygen flange area to look for any signs of insulation damage. If any cracks are seen, the countdown will be called off.

Bill Gerstenmaier, director of space operations at NASA headquarters, praised the NASA and contractor engineers and technicians who did the troubleshooting, analysis and repairs of Discovery's tank. Many members of the repair team were called in from Lockheed Martin's Michoud Assembly Facility near New Orleans where the giant tank was built. With only three more shuttle flights planned, the production line has been shut down.

"They did a phenomenal job of doing this testing and pulling the work together," Gerstenmaier said. "I couldn't be prouder of what they've done. ... A lot of these folks, some of them in the tank world, were laid off, they were already in other jobs and we called them back to do some of this work. ... There was no question about their dedication. They really want to see this vehicle fly."

CONTINUE TO PART 3 -->