The unraveling of the rendezvous began Dec. 20 at 5 p.m. when a major bipropellant burn that would put the spacecraft on track for an orbit insertion was aborted after sensors detected steering thruster shuttering that exceeded limits programmed into its onboard computers. The spacecraft defaulted to a safe mode, waiting for further instructions, and communications between NEAR and the Mission Operations Center stopped.

At that point, the NEAR team did not know the extent of the spacecraft's problems. They didn't know that it was tumbling or that it had lost 30 kilograms of fuel, which leaves the spacecraft with only about one-third the fuel it would have had if the Dec. 20 rendezvous burn had been successful.

NASA's Deep Space Network (DSN), which had been tracking the spacecraft, searched for it for the next 24 hours. At 7:30 p.m. on Dec. 21, they picked up a weak signal slightly off course from where they expected NEAR to be, and hope began to build. At about 8 p.m., verification was received that the signal was indeed coming from NEAR, and broad grins spread through the Mission Operations Center, located at APL.

The team anxiously waited for an opportunity to send a command to the spacecraft, telling it to remain pointed to Earth to receive further communications. Three hours later they got their chance when NEAR made a preprogrammed 360-degree sweep looking for a signal from Earth. Only a 10-minute window of opportunity existed for the DSN to locate a signal, but they found it and Mission Operations Center staff immediately started uploading crucial commands.

The day-plus period of communication loss had frayed nerves and flattened optimism. "It was the longest 27 hours of my life," Tom Coughlin, APL's Project Manager for the mission, was later to declare on a special Maryland Public Television broadcast on NEAR.

When contact with the spacecraft was regained, the team went into an intense scramble to determine exactly what went wrong and to examine alternative options that had been worked out ahead of time with the hope that they wouldn't be needed. The goal was to find a way to still rendezvous on Jan. 10. But time ran out before enough analysis could be completed to ensure a safe new burn.

They were now faced with a new challenge: get as much as you can from a "flyby" of the asteroid. New programs were written that would allow the spacecraft to take images of Eros and collect valuable data as it flew past. On Dec. 23, the asteroid came barreling toward NEAR just as the last of the commands were uploaded.

For two hours the spacecraft's instruments worked to glean as much data as they could before Eros moved out of range. The result was enough information to provide preliminary shape, mass, and composition estimates. Images taken as Eros approached then flew past showed no evidence of moons, 100 meters or bigger, orbiting the asteroid.

Looking back on the rendezvous abort, Bob Farquhar, APL's NEAR Mission Manager, says, "What saved the mission was our resilient mission design. We had worked out contingency plans ahead of time and had plenty of margin. If it wasn't for that we wouldn't have recovered."

A small hydrazine engine burn on Jan. 20 was successful, and NEAR and Eros are now traveling in nearly identical orbits around the sun and will meet again in mid-February 2000. For the next year, the spacecraft will travel in cruise mode with periodic instrument checks and hydrazine engine burns to keep it on track. A year from now it will close in on Eros for a second time and if all goes well, begin its yearlong mission, a year late, but with long-term science objectives intact.