During wind tunnel tests earlier this week, NASA subjected a full-scale mockup of a shuttle external tank section to aerodynamic forces greater than a real tank would experience during launch. In one series of tests, unmodified foam insulation used to prevent ice buildups around external fittings suffered only minor damage while a redesigned "ice/frost ramp" suffered major foam loss.

The redesign is being considered in a bid to remove as much insulation as possible from the tank to minimize the threat of debris shedding during launch. The old ice/frost ramps, which insulate the fittings used to hold two pressurization lines and an electrical cable tray in place, featured long, sloping ramps to smooth the flow of air.

But those very ramps, it was believed, were susceptible to cracks and failure, providing a potential source of debris that could strike a shuttle during launch. The redesigned ramps are much more blunt. But in initial wind tunnel tests at the Arnold Engineering Development Center in Tullahoma, Tenn., the old design fared better than the redesign. During two sets of test runs this week, the new design suffered major foam loss.

"That's exactly what testing is all about," said one official. "This is where you want to find a problem, not on launch day."

More tests are planned and other design changes are under consideration, along with the option of flying the ramps "as is." But a decision on how to proceed is needed by the end of the month to give engineers time to make any required changes before NASA's planned July launch of the shuttle Discovery on the second post-Columbia mission.

In other developments, shuttle program manager Wayne Hale has told project engineers to begin planning for a shuttle fueling test around June 1 to verify performance of new low-level fuel sensors in the tank scheduled for use by shuttle Discovery in July. The test has not yet been formally approved and is opposed by some - including, sources say, William Gerstenmaier, NASA's chief of space operations - because of concern about putting unnecessary thermal stress on the tank's foam insulation.

But Hale strongly favors the test in the wake of recent, unprecedented work to replace four engine cutoff - ECO - sensors, which are part of a critical backup system used to make sure the shuttle's main engines shut down on time. Problems with ECO sensors before Discovery's launch last July on the first post-Columbia mission were never fully resolved and a tanking test would give engineers confidence the new sensors will work properly on launch day. A tanking test also would provide an opportunity to monitor the performance of a relief valve used to maintain proper tank pressurization during flight. A valve in a tank last year cycled more often than usual.

Discovery remains targeted for launch at 3:48 p.m. EDT on July 1 (a detailed flight plan and the STS-121 Quick-Look page have been updated).

To make the July 1 launch target, however, the shuttle team must overcome several challenging hurdles, including completion of wind tunnel testing and analysis to confirm recent changes to the external tank's foam insulation are safe.

During Discovery's flight last July, a large piece of foam insulation broke away from an air deflector called a protuberance air-load - PAL - ramp running down the outside of the hydrogen section. The PAL ramp was in place to smooth the flow of air across two externally mounted pressurization lines and an electrical cable tray as the shuttle rockets out of the dense lower atmosphere.

Shuttle managers decided late last year to simply remove the ramps based on computer modeling that indicated the pressurization lines, cable tray and attachment fittings - and the foam ice/frost ramp insulation used to prevent ice formation on the brackets - were tough enough to withstand the expected aerodynamic buffeting.

Discovery's tank, without PAL ramps, was shipped to the Kennedy Space Center in March on the assumption wind tunnel testing would confirm the computer modeling. Tests using a scale model of the shuttle and others using a larger scale section of the tank have been completed at NASA's Glenn and Ames research centers. Those tests were focused on learning more about the acoustic environment the tank is subjected to during launch as well as the aerodynamic "loads" to make sure the tank and its external lines and fittings can stand up to ascent forces.

Those data have not yet been fully analyzed. Despite initial concern about somewhat ambiguous results, engineers in recent days have expressed optimism the end result will support the decision to remove the PAL ramps.

NASA currently is carrying out wind tunnel tests at the Arnold Engineering Development Center using a full-scale model of a tank segment that can be equipped with ice/frost ramps of various shapes. A turntable allows engineers to properly orient the mockup to ensure the wind strikes it at the proper angle to reflect the forces acting at any given moment during ascent. The testing is designed to collect data on all aspects of the PAL ramp-free tank design.

The ice/frost ramps were partially shielded from aerodynamic buffeting by the now-removed PAL ramps and questions were raised about whether the old ramp design could stand up to launch forces. In one set of runs at Tullahoma, the original design suffered only minor damage, sources said. But in two sets of tests using the redesigned ice/frost ramps, large chunks of foam blew off. It's not yet clear what caused the failures and more tests are planned.

In the meantime, the launch team at Kennedy is pressing ahead for July. A team from the Michoud Assembly Facility near New Orleans, where the external tanks are built, is refoaming the bottom of the tank following the ECO sensor swap out work. The tank is scheduled to be attached to a pair of already assembled solid-fuel booster rockets on April 24 and if all goes well, Discovery will be attached to the tank around May 12. Roll out to the launch pad is targeted for May 19.

The proposed fueling test would not prevent a July 1 launch, engineers say, because of contingency time built into the schedule following a March 15 decision to delay launch from May 10 to July 1. But the concern about subjecting the tank to unnecessary thermal stresses by loading it with super-cold rocket fuel remains.

Engineers now believe the expansion and contraction of the tank during multiple fueling cycles contributes to the formation of tiny cracks and other defects that can lead to foam loss during flight. Opponents of the fueling test believe the health of the ECO sensors can be confirmed through electrical testing without the need for a fueling cycle.

This issue raises thorny questions about how NASA will conduct future launch campaigns. In the past, the agency routinely fueled space shuttles for flight even when forecasters predicted bad weather. In some cases, the forecasters were wrong and the shuttle was able to fly. But in others, they were right and NASA was forced to drain the tank and try again another day.

If thermal stress can, in fact, lead to cracks and other issues with the foam insulation, as now seems probable, NASA may be forced to rethink its long-standing policy of fueling a shuttle when bad weather is expected. And that could have a bearing on how long it takes to get Discovery off the ground in July given the frequency of afternoon thunderstorms along the Florida Space Coast.

Weather aside, the fragility of the foam also could force NASA managers to consider putting limits on how many times a tank can be loaded with liquid oxygen and hydrogen. But at this point, that is speculative and no such discussions have been held or are currently planned.

Other issues still on the table include the possible swap out of main engine No. 2 because of a computer controller issue; completion of a complex debris verification review, showing the tank won't shed large pieces of foam or allow dangerous ice buildups around external fittings; and completion of a design review of the tank modifications in general.

Another wild card is the potential threat posed by so-called "tin whiskers" that can form on circuit card guide rails in various electrical components. While the precise causes are not known, small whisker-like extrusions can form in tin and other metals under certain stress conditions. The concern is that a whisker could break off, fall on a circuit board and cause a short.

Whiskers recently were found during inspection of an ascent thrust vector control box, used to control the system that moves the shuttle's main engines for steering. Discovery has numerous electrical boxes with a common design that are susceptible to such whiskering. Sources say engineers believe Discovery can safely fly as is based on redundancy in the systems involved. But a final decision on how to proceed has not yet been made.