Boeing technicians work with one of the Control Moment Gyroscopes before launch at Kennedy Space Center. Photo: NASA

The 800-pound gyros are critical to station operation, keeping the outpost properly oriented without the use of precious rocket fuel. The lab complex can operate with just two working gyros and as such, two more failures would have to occur before an on-board crew would have to shut the system down and instead rely on Russian rocket thrusters to provide orientation control.

That's the worst-case scenario and lead flight director Paul Hill said the station is in no danger of losing gyro control.

"Losing a CMG is a big deal, it's a major component," Hill said at an afternoon briefing. "But from a risk perspective right now, we're in good shape. We're single-fault tolerant whether we're docked or not docked. So for the long haul on station, we can still lose one more CMG and still hold attitude on the station side and minimize the amount of propellant we use. So we're in good shape.

Even so, losing a second CMG would "be a gut check," Hill said, "because then after one more failure we can't hold attitude with the CMGs and we're going to be burning a lot more Russian propellant until we get the CMGs fixed."

The CMGs are built by L-3 Communications of Teterboro, N.J. NASA has a spare CMG available, but it cannot be launched until early next year. That's because a CMG package - the gyro and necessary sub-assemblies - weighs some 1,100 pounds at launch and must be mounted on a special carrier beam in the shuttle's cargo bay. The next two shuttle flights, in August and October, will carry up huge sections of the station's solar array truss and don't have room for a CMG. As a result, the station may have to get by with three CMGs until early next year.

"For now, we're still a failure away from having zero fault tolerance, so we're two failures away from really having a technical problem that we need to jump through hoops for," Hill said. "But this is a major component that's failed and we're going to do the best we can to get the next CMG ready to fly and into an orbiter and get it changed out. But it could be six to nine months before we get this thing into orbit and get it changed."

The four CMGs are mounted inside a truss that extends upward from the Unity module's zenith port. The Z1 truss, attached to the space station during shuttle mission STS-92 in October 2000, also carries the station's main solar arrays.

Engineers first noticed problems with CMG-1 Friday evening. This morning, telemetry indicated current spikes and higher temperatures when suddenly, just before 11 a.m., one of two bearing assemblies apparently failed. Outgoing station astronaut Carl Walz, floating in the Unity module, reported hearing "a pretty loud, audible noise, kind of a growling noise, from inside the node" in the direction of Z1.

Station flight director Rick LaBrode said ground controllers quickly disconnected the gyro from its drive motor. An unpowered wheel normally takes 16 to 20 hours to freely spin to a stop. This morning, CMG-1 spun down in one hour, grinding and growling all the way.

Hill said a quick look at the Z1 truss using a camera mounted on the shuttle's robot arm revealed no obvious signs of "uncontained fragmentation," as the saying goes. But additional inspections likely will be carried out later in the mission to make sure.

In the meantime, flight planners are assessing the potential impact of CMG-1's failure on upcoming work to install robot arm attachment platform to a small rail car already mounted on the unfinished solar array truss.

The flight plan originally called for the mobile base system platform to be hauled out of Endeavour's cargo bay Sunday, near the end of a spacewalk by Franklin Chang-Diaz and Philippe Perrin. The MBS was to be positioned within a few feet of the rail car overnight and then mounted in place on Monday.

As it turns out, the shadow of the shuttle's nose will fall on the area where the MBS was to be parked. If left in that position overnight, the MBS could be damaged by low temperatures. Engineers already were planning to test an unusual station orientation on this flight, one in which the CMGs would be used to tilt the complex 10 degrees to one side to improve solar power generation. That 10-degree roll bias would have put the MBS in direct sunlight, eliminating the thermal issue.

But engineers are not sure they can safely implement the 10-degree tilt with just three CMGs. Troubleshooters currently are analyzing the procedure in light of today's failure and studying whether the robot arm can simply park the MBS somewhere else overnight and achieve the same thermal results. The issue is somewhat complicated, however, because the temperature of the MBS must be within 50 degrees of the temperature of the rail car for a precise attachment.

But LaBrode said he is confident the engineers will figure out a workable solution and that the MBS will be installed on the rail car Monday as planned, after Sunday's spacewalk by Perrin and Chang-Diaz. The spacewalk, the first of three planned for Endeavour's mission, is scheduled to begin around 11:08 a.m.