A slight misalignment in the way a vent port on the shuttle Endeavour's external tank was built into the structure is the leading candidate for what caused gaseous hydrogen leaks that derailed two launch attempts June 13 and 17, the shuttle program manager said Friday.

The use of a different type of seal where the vent line attaches to the side of the external tank may resolve the problem, he said. The alternative seal design should provide a tighter fit that is less susceptible to the temperature-induced mechanical shrinkage and motion that can put uneven stress on the interface and lead to leaks.

To find out, engineers at the Kennedy Space Center are making plans for a fueling test at launch pad 39A. Endeavour's tank will be loaded with super-cold hydrogen rocket fuel and sensors will measure leakage where the vent line connects to the side of the tank. The leaks typically show up when the mechanism is subjected to cryogenic temperatures.

Shuttle Program Manager John Shannon told CBS News Friday that if the tanking test goes well, Endeavour should be able to make its July 11 target for launch on a space station assembly mission.

"I'm pretty confident," he said in a telephone interview.

A gaseous hydrogen vent line attached to shuttle Endeavour's external fuel tank. Leaks at this interface derailed to launch attempts.

(Credit: NASA)

NASA initially attempted to launch Endeavour June 13. But the night before, as the shuttle's tank was nearing its full load, sensors detected a significant gaseous hydrogen leak at the ground umbilical carrier plate where the vent line attaches to the tank with a quick-disconnect fitting.

Liquid hydrogen, at minus 423 degrees Fahrenheit, constantly "boils off" as the tank is being filled. The vent line is used to carry the potentially dangerous vapor away from the shuttle to maintain the proper internal pressure. At launch, an explosive bolt detonates and the vent line falls away from the tank.

While some leakage at the quick-disconnect fitting is acceptable, hydrogen concentrations higher than 40,000 parts per million are grounds for calling off a countdown.

That limit was exceeded during both of Endeavour's fuelings and after the second scrub, the flight was put on hold to make way for Thursday's launch of NASA's Lunar Reconnaissance Orbiter from the Cape Canaveral Air Force Station.

Endeavour's launch window closes this weekend, due to temperature constraints related to the International Space Station's orbit. Assuming NASA can, in fact, fix the hydrogen vent line problem in time, the next launch window for the shuttle will open on July 11 at 7:39 p.m. EDT.

But the agency will only have four days to get the shuttle off the ground or the flight will slip to July 27 because of a critical Russian Progress space station resupply mission scheduled for launch July 24.

The Progress can "loiter" in orbit for five days, but it must dock by July 29. And that means Endeavour must take off by July 14 to complete its 16-day space station assembly mission in time to undock before the Progress arrives.

Shannon said three engineering teams have been set up to resolve the vent line seal issue. One team will begin precise measurements of the vent line interface this weekend before taking the mechanism apart next week. Those measurements are needed to confirm the root cause hypothesis.

Engineers at Lockheed Martin's external tank plant in Michoud, La., are making plans to replace the current rigid Teflon seal with the alternative flexible two-part seal. They also are evaluating washer-like shims on the umbilical plate's mounting hardware to further counteract the clocking misalignment.

Finally, engineers at the Kennedy Space Center will be making plans for a tanking test. Shannon said it would take about 10 days to finalize those plans. The actual fueling test would take place shortly thereafter.

"The hardware is still, they call it 'quarantined,' it hasn't been touched since we scrubbed," Shannon said. "The vehicle and the pad have been secured. I think the underlying root cause, they at least have a plausible reason why we had a leak twice.

"The way the umbilical line that carries the hydrogen away from the tank (is attached) to the tank is there's a plate that's bolted on with a pyrotechnic bolt. And there's a receiving plate that's on the external tank and those two are lined up, there are two little pieces of metal that go down to these hinge pins that keep it from moving side to side. And there's a Teflon seal ... on the inside of that flange coming out of the tank. The line going into it pressure fits in."

When Endeavour's tank was delivered from the assembly plant, "they were doing measurements and that flange on the ET side is cocked counter-clockwise .65 degrees," Shannon said.

"What has happened, we are pretty sure, is that when you put that external plate with the line on it onto the ET flange, there's a pyrotechnic bolt that holds those two together. It's above that round pipe. That whole system can rotate about that pyro bolt. If the two plates are in perfect alignment, it's not going to rotate, it'll just move slightly up and down. But since it is cocked a little bit ... it pulls the entire structure to the right (when the hardware contracts at cryogenic temperatures) and that allows a leak on the left side."

The clocking problem was the leading suspect after the June 13 launch scrub, he said. Shims were used to provide a firmer connection, but the seal still leaked. Engineers believe the two-part seal that will be used for the upcoming tanking test will provide the strength needed to resist the temperature-induced asymmetrical loading believed to be responsible for the leaks. Shims will be used as well.

The shuttle Discovery also was grounded by a vent line leak in March. But in that case, the misalignment was not as great and after the seal was replaced a subsequent fueling went well.

Shannon said engineers noted 18 to 20 gaseous hydrogen leaks at the ground umbilical carrier plate interface during previous fuelings but in all but two of those cases, cycling the vent valve caused enough vibration to help the seal seat itself.

The so-called two-part seal has been used in two previous shuttle fuelings, but NASA ultimately returned to the current Teflon seal design because it tended to leak less. But the leaks experienced by the two-part seal were within specification and did not require any corrective action.

If higher-than-allowable leakage is seen during the fueling test using the alternative seal design, Shannon said the team may have to consider more extensive work to remove, realign and re-attach the vent line flange on the external tank. In that case, Endeavour likely would be moved to a different tank and launch would face a more significant delay.

Shannon said no other tanks in production at the assembly plant have misalignment issues like Endeavour's, which he described as 'way out of family.'"