Shuttle Atlantis and five astronauts now are tentatively scheduled to launch July 12 on a mission to deliver the $164 million airlock, which will serve as a staging area for spacewalks.

Delivered to the station in April, the new crane is required to install the airlock because the shuttles shorter robot arm is not long enough to do the job.

A passing problem with the shoulder joint of the station arm, however, is prompting NASA managers to consider delaying the high-stakes airlock installation mission until September.

During the last half of May and early June, the joint failed to operate properly when back-up control systems were being used to operate it. Those problems, however, mysteriously disappeared last week and the arm since has been operating without major troubles.

In fact, Helms and fellow flight engineer Jim Voss used back-up control systems Thursday to carry out a dry run of the airlock installation job. And both the arm and its shoulder joint worked pretty much as advertised.

"It appears like the robot arm is now back in good working condition," Helms said. "We have a lot of reason to hope that these were transient problems."

Engineers, however, so far have been unable to determine the cause of the problem or duplicate it during tests. Consequently, managers fear the problem could crop up again, leaving the airlock stranded at the end of the crane while astronauts are trying to mount it to the station.

A less troublesome problem caused the arm to strike the side of the 17-story station during the dry run of the airlock installation job. And engineers now are chalking that up to a cradle position the arm had been maneuvered into earlier this month.

With the shoulder of the arm attached to an exterior anchor on the U.S. Destiny lab, the tip of the crane was latched to a pin-like grapple fixture on the side of a nearby shuttle docking port. The so-called "joint maintenance position" would enable spacewalking astronauts to perform what amounts to shoulder surgery if that became necessary.

Stowing the crane like that apparently created thermal conditions that caused the tip of the arm to hang up on the fixture during the initial start-up of the dry run test. The tip then sprang loose with such force that it rebounded and banged into the grapple fixture.

The tip "held on for a little bit longer than it should have, and then the built-up forces allowed it to release," Voss told reporters Friday. "It just backed off a little bit, came back in and contacted the grapple fixture, and then bounced off again."

At the time, Helms told ground controllers she heard one distinct "bump." Voss later said the "clang" was similar to the sound heard when a window shutter on the Destiny lab is opened too rapidly.

"We heard it hit," Voss told ground controllers in a space-to-ground radio debriefing after the test. "The clang sound on the outside of the station (was) not as loud as when you open the window shutter too fast, but still loud enough to get our attention."

A photographic survey was carried out and no dents or dings were apparent.

"We looked at it afterward with some cameras, and there was no damage that could be seen," Voss said. "And it worked just fine while we operated it later in the day."

Helms, meanwhile, indicated that the airlock installation job could be carried out even if similar problems were to crop up.

"If there are problems along the way, there is no problem with stopping, evaluating the situation, seeing if we can proceed or back out of it," she said. "I think we have a lot of options if we run into hiccups along the way."

NASA project managers will decide next week whether to delay the airlock installation mission, a move that might be made to give engineers more time to sort out the situation with the arm.