Frames from the newly enhanced video just prior to (top) and after (below) a piece of foam from the external fuel tank appears to strike Columbia's wing. Photo: CAIB.

The Columbia Accident Investigation Board today released the clearest video yet showing a tumbling piece of foam insulation slamming into the shuttle's left wing during launch Jan. 16.

Investigators say the enhanced video, along with ongoing mathematical modeling, indicates the foam struck the wing at some 529 mph, imparting up to a ton of force across an area of the leading edge measuring roughly six by 12 inches.

Investigators believe the foam impact likely cracked or breached one of the reinforced carbon carbon panels making up the leading edge of the left wing or damaged a so-called T-seal between two adjacent panels. Whatever the exact mechanism, investigators believe Columbia began its re-entry Feb. 1 with a breach at or near RCC panel 8 or perhaps near the T-seal between panels 8 and 9. Super-heated air burned its way into the wing through this presumed breach, leading to the shuttle's eventual destruction.

CAIB and NASA investigators are gearing up for a crucial series of tests at the Southwest Research Institute in San Antonio, Texas, early next month to help pin down whether the foam impact was, in fact, a "root cause" of the disaster.

Five initial test runs already have been carried out shooting foam "bullets" at heat shield tiles on a landing gear door taken from the prototype shuttle Enterprise. The tests were set up early on in the investigation, before engineers knew the impact had actually occurred at the leading edge.

In a teleconference with reporters today, CAIB member Scott Hubbard said only minimal tile damage resulted from the initial impact tests using foam bullets, weighing between 1.2 and 2.5 pounds, impacting at angles between 5 and 13 degrees. But the tests have helped verify the predictions of two computer models and confirm the actual impact happened at or very near the leading edge.

Early next month, engineers will begin firing foam at a high-fidelity mockup of a shuttle leading edge, complete with RCC panels taken from the shuttle Discovery that have flown more than two dozen times. The leading edge simulator will be heavily instrumented with more than 100 channels of data charting the stresses and strains imparted by the impacts.

The foam will be fired from a 30-foot-long nitrogen-gas canon with a rectangular bore measuring 5.5 inches by 11.5 inches. Foam bullets similar in size to the actual debris that hit Columbia will be fired at the simulator at velocities of up to 775 feet per second, or 528.5 mph. Six high-speed cameras, some capable of recording 7,000 frames per second, will photograph the impacts in exquisite ultra slow-motion detail, permitting precise determination of impact velocities and angles.

"All the experts looking at all the data have begun to home in on a sweet spot," Hubbard said of work to determine the size of the foam debris that struck Columbia. "The current best guess, and this may change a little bit over the next week or so, is an impact projectile of about 1,240 cubic inches traveling at about 775 feet per second."

One wild card is the contribution of the foam's rotational velocity. Up until now, engineers have calculated the force imparted by the impact based on the foam's straight-line velocity. But Hubbard said the debris was tumbling wildly and that regardless of the direction of the tumble, the rotational velocity must be factored in.

"A major element has been to include the rotational velocity," he said. "This is something that had personally been gnawing at me in looking at the video, of how this piece was rotating. It seemed to me to be a source of additional energy. ... Some preliminary calculations show that we may need to either adjust the velocity or the angle to compensate for this."

Some outside observers worried the initial test results, showing only minimal damage to the landing gear door tiles, might indicate the foam would have little effect on the RCC panels, leaving NASA without a clear-cut root cause for the disaster.

But Hubbard said today there is little or no data on how RCC panels respond to impacts. And unlike the tiles, which are supported across their full length and width, RCC panels are only supported at their edges, by the T-seal. The central, unsupported area is just a third of an inch thick.

"The difficulty is in modeling this curved surface," he said of the U-shaped RCC panels. "There was some initial ... analysis that was done on an impact of a piece of foam of the 2-pound variety traveling at 700 feet per second against a flat plate (of carbon composite material). What that seems to indicate - and I really underscore 'seems' - is that it should break the panel.

"But I don't want to go any further than that because that's a flat panel and not a curved surface. As we've learned from doing this first set of tests, getting the tumbling, the angle and all of that just right is a tricky business. So I don't think at this point we know exactly what we might find. That's why we're doing the tests."

The impact angle chosen for the tests is especially critical. That angle is measured relative to the flat bottom of the wing, not relative to the tangent of the curve at the point of impact as common sense might dictate. Relative to the belly of the orbiter, impact angles for debris striking the lower side of the leading edge RCC panels could range from 10 to 20 degrees or so.

Velocity also is critical. As every high school physics student learns, an object's kinetic energy is one half its mass multiplied by the square of its velocity. While the mass is low in this case, the velocity is high and that is the term being squared.

"Very light things, when you accelerate them to very high velocities, carry an enormous amount of force and the types of forces we're dealing with here, with the foam size and velocities we're talking about, even at these relatively shallow angles, are something close to a ton of total force, upwards of 2,000 pounds of total force delivered in relatively small area of about 5 or 6 inches by a foot," Hubbard said.

Trying to extrapolate the results of the initial tile impacts to the RCC panels is not possible because "it's a little bit apples and oranges," Hubbard said. While engineers understand how tiles respond to impacts, no such database exists for carbon composites.

"The tiles have a certain crush force that we now understand fairly well based on all the earlier tests," Hubbard said. "We don't have that same level of information yet for aged RCC. One case (tile), you have a glass-like material that is very sensitive to the angle of the impact and that angle, I feel almost certain, has something to do with how the edge of the foam digs into it. If the leading edge were at 20 degrees or so and made out of tiles, I think we would see a substantial amount of damage. The fact is, that angle that we saw, the angle that exists on the bottom of the orbiter, is only 5 or 10 degrees. That's one story.

"A completely different story is the RCC panel, which is only supported on the ends and is maybe a third of an inch thick or so and has quite different material properties. It will probably show some angular dependence. The smallest angle of intersection that we've measured is about 10 degrees and it goes up to well over 20 degrees.

"At that range of angles, you transfer a lot of force and that amount of force could be, we think, enough to break it. But we won't know for sure until we do the tests."

Investigators have not yet decided exactly where they will aim their foam bullets at the high-fidelity leading edge simulator. Because each strike will affect the simulator in some fashion, causing an unknown amount of damage, only a limited number of firings will be possible.

The current schedule calls for shooting at panels 5 through 7 beginning the week of June 2. Panels 8 through 10 will be hit starting the week of June 23rd.

"At this point I'm fairly well convinced we're going to hit panel 6 pretty much in the middle of the panel," Hubbard said. "I'm looking at two or three options for the panel 8-9 tests. One is down at the bottom of the panel. The other though is ... the T-seal area (between panels 8 and 9). And I haven't made, nor have we agreed with NASA yet, exactly the best place to put that impact and that's one of the things we're going to be looking at in the next two weeks or so."