NASA managers had hoped to launch Atlantis on mission STS-117 on March 15, but the Feb. 26 hail storm derailed the agency's plans and ultimately delayed launch three months.

"I'll never forget the day of the hail storm itself and then the first time I saw the external tank in the VAB," said Launch Director Mike Leinbach. "I was really wondering if we were going to be able to fix this tank or not."

John Chapman, external tank program manager at NASA's Marshall Space Flight Center in Huntsville, Ala., said the hail storm caused more than 4,200 individual dings, dents, pits and gouges to the foam insulation.

Of that total, between 1,400 and 1,500 were tightly clustered at the very tip of the tank near its carbon composite nose cap. Rather than make individual repairs in that area, the foam was sanded down to eliminate the pits and then filled in with sprayed-on insulation. That fresh insulation was then milled to the proper slope and to an acceptable thickness by a cleverly engineered "pencil sharpener" device that rotated about the top of the tank.

Of the remaining damage sites, 449 were clustered together on the side of the oxygen tank and repaired with a second large-area spray. Another 1,038 pits and gouges were fixed using a pourable foam known as "PDL" and 889 dents were fixed by simply sanding them down using a "sand-and-blend" technique. Another 412 sites were so minor no repairs were required.

To make sure the required repairs could stand up to the aerodynamic and thermal rigors of launch, engineers re-assessed the flight performance of past PDL repairs, simulated damage sites using steel balls to impart hail-like crush forces and then subjected repairs to a hot-gas wind tunnel at Marshall.

"We were able to place simulated repairs in close proximity to each other to study the effect of multiple repairs, both side by side and downstream from each other," Chapman said. "In fact, we even developed a drop test capability to be able to drop steel balls that would represent the amount of kinetic energy the hail stones would have ... and then put that into the hot-gas facility and see how it performs."

The goal, Hale said, was assurance no dangerous debris would fall off the tank during ascent.

"We have gone through an extensive set of tests and analyses with the repairs that have been done in these test facilities to ensure they won't release debris and all our testing today indicates that will not happen," he said.

During the first 110 seconds of ascent, atmospheric friction raises the tank's temperature to some 650 degrees in some places and heating continues throughout the climb to space. Along with thermal concerns, foam debris falling off the tank could threaten a Columbia-type impact to Atlantis' heat shield.

"As you accelerate this vehicle supersonically in the lower atmosphere, you build up quite a bit of heat, several hundred degrees," Hale said. "But you increase in temperature all the way up and even in second stage, when you are almost at orbital altitudes, you continue to put heat in.

"The point is, we have to protect both the thermal environment, in other words you cannot allow those temperatures to get to the aluminum-lithium substrate, which would weaken the metal, nor can you allow a significant amount of foam to come off," he said.

"The foam up in the forward part of the tank, because it does get that warm, does have an ablation process, so you lose at the molecular level small pieces of foam, but when we're talking about debris that can cause damage, again, the goal is not to have any of that. That's the kind of testing that was done at the hot-gas facility at Marshall to ensure these repairs don't liberate any significant amount of foam, even at the elevated temperatures."

No previous PDL repairs have ever failed in flight. While there is some slight additional risk flying the repaired tank, Hale said, it is not considered significant.

The hail storm, Chapman said, "left the external tank team with a tremendous amount of work to do. This team has been essentially working 24/7 since the storm, doing engineering analysis, testing and repair of the tank. In my estimation, they have done a fantastic job. This has truly been unique. We've had hail damage before, but never to this magnitude."

Said Leinbach: "The team has come through with flying colors and gotten a completely flight worthy tank put together. ... We are essentially done with all the repairs on the tank now."

As with all post-Columbia shuttle flights, NASA will utilize an upgraded tracking camera network at the Kennedy Space Center and Cape Canaveral Air Force Station to make sure any foam that might fall off is seen as soon as possible.

Columbia was destroyed during re-entry Feb. 1, 2003, because of damage to the leading edge of the ship's left wing that was caused by the impact of foam debris during launch 16 days earlier. The foam in question has been eliminated and a variety of other changes have been made to minimize potentially dangerous debris shedding. The two most recent tanks performed well and NASA managers believe they have done all that can reasonably be done to improve safety.

But the tank still features foam buildups known as ice-frost ramps and the threat of impact damage remains high. The ability to quickly spot such damage, giving engineers time to assess the consequences and possible repair options, is critical.

At launch complex 39A, some 38 16mm cameras are mounted on the launch pad itself with three short-range camera sites around the pad perimeter featuring two 35mm cameras and one high definition TV camera each. Another 11 medium-range camera sites are positioned around the pad between one and six miles away, each one equipped with a 35mm camera and all but one equipped with an HDTV camera. Another 11 long-range camera sites are located between four and 40 miles of the pad. All long-range sites include 35mm cameras, two have 70mm cameras and 10 are equipped with HDTV.

From liftoff through the first 30 seconds of flight, objects an inch wide or larger can be seen. Between 30 seconds and one minute, the resolution drops to objects three inches in diameter or larger and from one minute to 90 seconds, it drops to objects eight inches or larger. Between 90 seconds and booster separation two minutes after liftoff, ground-based tracking cameras can detect objects 15 inches across and pinpoint an impact site to within five feet.

Finally, a radar system is in place featuring one ground-based C-band and two ship-based Doppler X-band instruments to look for debris coming off the external tank.

For Atlantis' flight, eight cameras mounted on the shuttle, its tank and twin boosters will provide close-up views of the external tank and the orbiter's belly during ascent.

Another camera mounted high up on the external tank looking down on the underside of the space shuttle will beam back live television views throughout the eight-and-a-half-minute climb to orbit.

As with Discovery's flight in December, Atlantis is equipped with four other cameras, two near the top of each booster and two mounted near the back end of the powerful rockets. Each booster also carries a camera focused on a region of the tank known for losing small, popcorn-like pieces of foam.

Imagery from the six booster cams will be available after the spent rockets are recovered and towed back to Port Canaveral a few days after launch.

In addition, a digital camera mounted in a cavity where a propellant line enters the belly of the orbiter will photograph the tank as it separates in space.

As if all that wasn't enough, an X-band marine radar seven-tenths of a mile from the pad will be on the lookout for vultures and other large birds. During Discovery's launch in 2005, a large vulture struck the external tank a few seconds after liftoff, rammed by the shuttle at some 70 mph. If any large birds are seen prior to Atlantis' launch, the countdown can be halted briefly if necessary.

Continue to Part 4 -->