Ongoing analysis of a "treasure trove" of data from a recorder recovered in the wreckage of the shuttle Columbia shows a deadly plume of super-heated air first began eating its way into the ship's left wing just five minutes after the orbiter fell into the discernible atmosphere. The sudden temperature increase, in a cavity behind the U-shaped panels making up the leading edge of the left wing, came a full three minutes earlier than previous telemetry indicated the start of unusual heating.

Entry interface occurred at 8:44:09 a.m. and a sensor mounted on a brace behind leading edge panel No. 9 began responding to a very sharp temperature increase at 8:49:09 a.m., nearly two minutes before the shuttle entered the region of maximum heating (this is earlier than officials reported Sunday evening based on quick-look data analysis of the OEX data). In any case, exactly three minutes after that initial temperature rise, the sensor failed and dropped off line after detecting a temperature of 450 degrees Fahrenheit. Eight seconds later, at 8:52:17 a.m., a sensor mounted on a brake line in the left main landing gear wheel well recorded the start of an unusual temperature rise.

Illustration of orbiter wing marks some of the RCC panels. Measurements from the OEX recorder shows temperature spikes behind panels 9 and 10. Credit: NASA

Harold Gehman, chairman of the Columbia Accident Investigation Board, said today the timing of these events suggests Columbia began its re-entry with a fatal flaw that allowed hot gas to first enter the cavity behind the leading edge panels and then to burn its way into the interior of the wing. The hot gas presumably entered the wheel well through a vent on its forward face, triggering the brake line temperature increase.

A few seconds later, beginning at 8:52:49 a.m., a sensor mounted on the forward interior edge of the wing, just behind the failed sensor in the leading edge cavity, began registering a rapid increase as hot gas shot through the interior. Seven seconds after that, data from sensors mounted toward the back of the wing began dropping off line as the plume burned its way through a wire bundle routed around the outside of the landing gear wheel well.

"Now interestingly enough, the first temperature rise we see inside the wheel well occurs at 52:17," Gehman said. "Just about the time the sensor outside the wing spar fails, the temperature inside the wing spar starts to go up. So this is interesting to us."

The sensor mounted on the inner surface of the wing spar roughly behind leading edge panel No. 9 "starts to rise 520 seconds after EI, which is something like eight minutes and 40 seconds, which would make it something like 52:49," Gehman said. "The temperature sensor inside the wing starts to rise almost coincidentally with the time the temperature sensor outside the spar goes off line, which leads you to believe there was a whole lot of heat outside the spar which finally ate its way into the wing.

"Remembering the melting temperature of aluminum is something like 900 degrees, it got really hot outside the spar, ate its way through and the temperature sensor outside the spar goes off line," he said. "It probably got destroyed."

Enhanced video of Columbia's launching shows a suitcase-size piece of foam debris from the shuttle's external fuel tank pulled away and slammed into the left wing about 82 seconds after liftoff. The two-foot-wide impact footprint is centered on reinforced carbon carbon - RCC - leading edge panel No. 6. The impact could have affected two adjacent RCC panels as well as protective tiles atop so-called "carrier panels" marking the interface between the lower edge of the RCC panels and permanent heat-shield tiles on the lower surface of the wing.

Gehman said the temperature data recovered from Columbia's orbiter payload experiments recorder, or OEX recorder, does not rule out or confirm an initial breach at or near RCC panel 6. He said engineers will be making calculations over the next few days to determine what sort of breach would be required near RCC 6 to produce a temperature increase like the one seen on the sensor behind RCC 9. At the shuttle's extreme altitude, the dynamic pressure was low and there was not much air to transmit heat. Whether the aluminum structure of the wing spar could have conducted that much heat, or whether a breach occurred outboard of RCC 6, is not yet known.

But it now seems all but certain Columbia began its entry with a fatal flaw already in place, an entry point for the super-heated air that ultimately led to its destruction.

"At the altitude and the speed the orbiter was going at the time the new temperature rise was seen, the aerodynamic pressures are extraordinarily low," Gehman said. "So you could conclude the orbiter began it's entry with a pre-existing fault. In other words, it wasn't air pressure that knocked this thing off, it wasn't some kind of a weakness in which some part of the orbiter was ripped off when it entered the atmosphere, this looks to us like it probably had a pre-existing condition."

But Gehman urged reporters not to jump to any conclusions about the possible location of the initial breach.

"Where the temperature sensor is does not suggest where the breach is," he said. "We are still pulling all the data together to attempt to indicate where the breach is. But this certainly leads us away from things like tile and landing gear doors and things like that."

The OEX recorder is capable of recording data from some 721 sensors scattered across the shuttle. About 150 sensors were not operational during Columbia's entry because of past problems or wiring issues. Gehman said engineers hope to recover data from about 570 sensors when all is said and done.

"It's a treasure trove of data," Gehman said.

Along with showing when the first signs of heating occurred in the left wing, the OEX data shows unusual heating effects on the left orbital maneuvering system rocket pod just to the left of the shuttle's vertical stabilizer.

"Temperature sensors on the skin of the orbiter high up in the vicinity of the OMS pod, the maneuvering engine covers high up on the port side, they start to do all kinds of strange things," Gehman said. "Some of them go up, some are going down. Some go down and then rise later. Some rise and then cool off. We don't know what to make of that except the heating pattern and the flow pattern over the top of the left wing, which then spreads over the side of the fuselage, is obviously disturbed in some way."

Prior to the discovery of the OEX recorder, realtime telemetry from the shuttle showed an initial loss of contact at 8:59:32 a.m. Thirty seconds later, a final two-second burst of telemetry was received that showed Columbia was in the process of yawing out of control to the left. Gehman said today the OEX recorder ran for an additional eight seconds. That data has not yet been recovered, but it should shed additional light on how much of the ship was still intact at that point and when final breakup began.

At the other end of the scale, investigators are working to extract data about the forces acting on Columbia during launch. Gehman said engineers are especially interested in seeing what sorts of forces were acting on the orbiter around the time the external tank debris hit the left wing. Columbia encountered high wind shear during launch, prompting more extreme booster steering movements than normal. Whether those forces contributed to stresses that, when aggravated by the debris strike, could have led to an RCC failure or a carrier panel problem is not yet known. But Gehman said OEX data from ascent will be factored into upcoming tests in which foam will be fired at leading edge components to determine what sort of damage they might have suffered.

Gehman said engineers will add the OEX data to NASA's current entry timeline (revision 15) and issue updates as warranted. Whenever those updates are made available, they will be added to our integrated timeline