24 February 2003
Analysis hints at shuttle’s last seconds

By James Oberg,

Image: Flight Director Leroy Cain Seen On February 1
Entry flight director Leroy Cain holds his hand over his face as workers in Mission Control wait for a signal from the Columbia crew minutes before the scheduled landing in this image from Feb. 1.

Feb. 24 — New analysis of the garbled last 32 seconds of radio signals from the space shuttle Columbia has raised the possibility that the crew survived up to a minute after the spaceship began tumbling out of control and breaking up. This reconstruction of the tragic end of the mission on Feb. 1 contrasts sharply with most preliminary assessments that the craft disintegrated suddenly and totally.

IN THIS VIEW, the shuttle lost its struggle to keep its nose pointed ahead, and began a flat spin to the left. The airstream, though ferocious, was not powerful enough to tear the vehicle apart immediately. The analysis indicates that Columbia could have turned through at least one full tumble in about 20 seconds while the cabin remained intact and pressurized. Following the breakup of the vehicle, the cabin fell for tens of seconds before it was crushed by the heat and deceleration.

For now, this scenario is only an analysis based on the assembly of still-incomplete pieces of the Columbia puzzle. The scenario has not been confirmed by senior officials at NASA or by members of the board investigating the Columbia tragedy. But it is shared by a growing number of space experts, inside and outside NASA, who have discussed their views on condition of anonymity.

The image that emerges is of the shuttle turning end over end at least once before the fuselage breaks apart. During the tumble, large pieces of the wings, tail and engine nozzles would have been torn off. But the crew cabin would survive for additional tens of seconds until crushing deceleration finally tore it apart.


The most persuasive evidence for this scenario comes from the 32 seconds of corrupted data that followed the last readable telemetry and voice signals from the shuttle. The signals were unreadable in real time because of massive “data dropouts” — and thus they did not appear on Mission Control’s flight control screens. But the bitstream was recorded at a ground station in White Sands, N.M., and it has slowly been yielding its secrets to mathematical analysis.

At the point when Mission Control’s readable data stopped, Columbia was approaching the Dallas area at an altitude of 206,000 feet and a speed of 12,500 mph. With its nose pitched up by 40 degrees, the shuttle was in a steep left bank as part of a series of “S-turn” maneuvers, aimed at bleeding off orbital energy and slowing down for its planned landing at the Kennedy Space Center in Florida.

Although the shuttle was going Mach 16, the air was so thin that the effective “dynamic pressure” on its structure was the equivalent of a sea-level wind speed of 170 mph, or a Force 5 hurricane. If a space shuttle were sitting on the Florida runway in such a storm, major damage would be expected — but not instantaneous disintegration.

The re-entry heating reached as high as 3,000 degrees Fahrenheit, but this was in the shock wave of squeezed air that was piling up a few feet in front of the vehicle. This high temperature was not caused by friction of the air moving across the shuttle’s skin itself — behind the “shock front,” the air moved across the skin at only a few hundred miles per hour. This hot air conveyed the tremendous heat of the re-entry shock wave into the shuttle’s protective tiles. If those tiles were damaged, as investigators suspect, the heat would have entered directly into the shuttle’s metal structure.

According to several sources, the deciphered data show that for several seconds after Houston Mission Control saw a loss of signal, the flight continued much as it had up to that point, except that additional steering rockets in the shuttle’s tail had turned on. This was apparently the autopilot’s attempt to counter growing drag on the left wing, which was pulling the shuttle’s nose to the left.

The origin of that drag remains unexplained, although most analysts agree it was almost certainly the result of damage to the wing, perhaps caused by debris that broke off the shuttle’s external fuel tank shortly after liftoff 16 days earlier.


After a few seconds of garbled data, the communication pathway — leading from the shuttle to a satellite in space and then down to the White Sands ground station — apparently went totally dead for about 20 seconds. Not even a garbled carrier signal came through.

But then, a burst of three or four seconds of corrupted data was received, followed by unbroken silence.

Some space engineers have interpreted this sequence as evidence that the shuttle, succumbing to the growing leftward torque, had turned away from pointing its antenna toward the relay satellite. The later burst of data could have resulted from the antenna momentarily turning again to face the satellite. The engineers surmised that a computer program designed to select different antennas could not cope with the fast turning.

Data from those final few seconds show a spaceship that was mortally wounded, but still working. Power was still being generated by the fuel cells under the payload bay, and signals were being received all the way from the back end of the craft.

Although the crew cabin was still pressurized and the four primary control computers were still functioning, other systems were in terminal distress. The three redundant hydraulic pressure generators — needed to control the shuttle’s aerosurfaces — were still functioning. But hydraulic pressure on the left side of the shuttle was zero in all three lines. The thruster system in the tail was reporting massive leakage of propellant.
These failures would have filled the cabin with the noise of alarms. Indicator lights would have been ablaze. The crew on the flight deck would have responded to these alarms in accordance with their training, if they were able. Those seated on the middeck would have prepared for emergency bailout once they got low enough in the atmosphere.

Based on the still-fragmentary readings, some engineers believe the shuttle was in a flat left spin. Others have suggested that the left wing was totally torn off, or was bent up against the side, causing the vehicle to roll left. In either scenario, the vehicle would have turned its back end into the wind. Parts of the tail (including, apparently, the drag chute package) would have torn off first, along with the bell-shaped rocket nozzles for the main engines and the orbital maneuvering engines. Damage there would explain the propellant leak alarms, which would have been followed by fiery detonation of the mixing chemicals.

Interpretation of the videotapes of the disintegration over Dallas remains unclear. Some smaller pieces are seen coming off a main body, followed by flashes of light that could indicate the detonation of propellant in the leaking tanks. A much larger scatter of large and small objects then can be seen. Heavier objects — three in particular, possibly the main engine blocks — forge ahead. Lighter tumbling objects, likely wing segments, slow and fall more quickly.


The path of the crew cabin can only be guessed, once it tore loose from the rest of the fuselage and electrical power ceased. Buffeted and braked by air drag, it would have been heated by the surrounding shock-induced plasma. Falling deeper into the atmosphere, G-forces would have built up to the point that the heat-weakened aluminum frame collapsed in on itself. Some pieces broke loose and were carried away by the aerodynamic forces.

A more precise analysis depends on the scatter of impact points of the cabin and its contents — information that is still being analyzed by the accident team.

This analysis follows in the footseteps of the investigation into the 1986 Challenger tragedy. In Challenger’s case, the initial impressions were that the crew had perished instantly when the shuttle came apart, a minute after its launch from Kennedy Space Center. Only months later did it become clear that the crew cabin had separated cleanly and had risen to an altitude of 65,000 feet before falling back to impact the ocean with a force of 200 G’s.

Equipment recovered from the wreck showed that at least some of the crew had survived the initial breakup and had activated their safety equipment. Medical specialists later concluded that they soon lost consciousness but were not killed until the impact with the ocean, two minutes after the explosion.

Columbia’s crew had better survival gear, including pressure suits and personal parachutes. Assuming they were conscious of the emergency, they would have closed their visors when cabin pressure was lost. Their suits would have automatically pressurized. It would then be only a question of hoping that the cabin held together until it fell low enough — below 40,000 feet or so — for them to blow the escape hatch and jump free. This would not have seemed an entirely hopeless situation, until the cabin’s own structure began to fail.

No one can know what Columbia’s seven astronauts were actually experiencing and doing in the final seconds of their flight, but the engineers who discussed the possible scenarios were deeply shaken by the implications. The overwhelming consensus is that the lack of knowledge is probably the merciful way it should be.

James Oberg, space analyst for NBC News, spent 22 years at the Johnson Space Center as a Mission Control operator and an orbital designer.


Global Network