A fuel sensor that has given NASA headaches in the past acted up Friday morning as space shuttle Atlantis’ crew prepared for liftoff.

The agency continued with launch preparations as engineers tried to figure out if the shuttle could fly with the glitch. Atlantis’ six astronauts, dressed in orange flight suits, arrived at the launch pad and strapped into the shuttle for the scheduled 11:40 a.m. EDT launch.

The sensor glitch was another problem the agency didn’t need as it rushes to get the shuttle and its crucial addition to the international space station into orbit following several delays.

NASA managers have two options: scrub for the day and try Saturday or fly with only three of the four hydrogen fuel sensors working properly. Although no decision had been made, the space agency was leaning toward postponing for a day, said NASA press secretary Dean Acosta.

In order to fly Friday, NASA would need to waive a rule requiring all four sensors work properly, said NASA spokesman Bruce Buckingham.

The fuel gauges are designed to prevent the main engines from running too long or not long enough during the climb to space. If two sensors fail, a main engine would shut down. An engine shutdown at the wrong time could prove catastrophic, forcing the astronauts to attempt a risky emergency landing overseas, or leading to a ruptured engine.

Workers finished pumping more than 500,000 gallons of supercold liquid hydrogen and liquid oxygen shortly before 6 a.m. EDT, more than an hour later than scheduled because the agency had to replace a nitrogen valve on the launch pad.

Earlier this year, the launch of Discovery was delayed by almost two months so four hydrogen fuel tank sensors could be replaced after one was found to be faulty. A similar problem briefly delayed last summer’s launch of Discovery on the first shuttle flight since the Columbia disaster in 2003.

NASA managers Thursday evening cleared another problem that had delayed the takeoff: a short in a motor in one of the shuttle’s three electricity-generating fuel cells.

Atlantis was supposed to launch Aug. 27, but was delayed first by a lightning strike at the launch pad, then by the approach of Tropical Storm Ernesto. It was held up again Wednesday morning when the fuel cell problem arose right before the shuttle’s tank was filled.

But the wait for this 11-day construction mission goes back far longer for Atlantis and its six astronaut crew.

One of the two girders that Atlantis is hauling up to the international space station has been waiting at Kennedy Space Center for nearly seven years. Atlantis’ mission is to bring up two girders and solar panels — weighing 17 1/2 tons — to the orbital outpost.

Originally, the mission was slated for May 2003, but the February 2003 Columbia accident put the mission on hold for more than three years. Atlantis, which has flown 26 times, hasn’t launched since October 2002.

Atlantis’ astronauts will restart construction on the half-built international space station for the first time since the Columbia disaster.

NASA has only five-minute launch windows on Friday and Saturday — and Saturday was a last-minute addition after talks with Russian space officials — before it has to scrap launch plans for at least two-and-a-half weeks.

The Russians are launching a three-person Soyuz capsule to the space station on Sept. 18, and Atlantis has to leave the station before the Soyuz arrives to prevent a cosmic traffic jam.

NASA managers had to convince themselves that the baffling electrical problem cleared Thursday wouldn’t be a safety risk — something they did over the objections of NASA’s safety managers and the makers of the fuel cells. Shuttle engineers deduced — but couldn’t prove — that the problem was likely from thinning wires that caused a coolant pump motor to give errant results.

The 30-year-old motor last flew in 1999 on a mission that had a fuel cell problem, but it was likely unrelated to the problem that cropped up days ago, said Steve Poulos, shuttle orbiter projects manager.

Engineers calculated that the two-week job of changing fuel cells would not only delay a launch but could cause more problems than letting the fuel cell stay as is, Poulos said.

A failure wouldn’t be a safety problem and would only require a shortened mission, said shuttle program manager Wayne Hale.

If the launch isn’t completed by Saturday, NASA could open up launching opportunities in late September and early October by relaxing a requirement that launches take place in daylight.

The rule was implemented for the first few flights after the Columbia accident so NASA can check if foam comes off the external fuel tank and could damage the shuttle’s heat shield as it did Columbia’s.

On the Net:

NASA at www.nasa.gov