The day the room went silent: NASA design engineer remembers Columbia explosion

In 2003, Columbia disintegrated over Texas while returning from a 16-day trip to space. Later analysis found that a small bit of foam insulation that broke off an external fuel tank during launch had torn a hole in the orbiter's wing and led to the disaster. Gene Grush recounts the events from NASA's perspective.

We were the design engineers, and we knew the space shuttle better than anyone. We were just there in case something out of the ordinary occurred. On Feb. 1, 2003, that’s exactly what happened.

I was the lead engineer over many other specialized engineers that worked the propulsion and power systems on the Orbiter at the time. My official title was Energy Systems Division Chief Engineer for Orbiter.

STS-107 was supposed to be just one of many re-entries that I had supported. I liked entries best; they seemed less stressful than ascents. Ascent involved complex propulsion systems, after all, and we had already lost one vehicle during this phase -- Challenger, in 1986. Entry was a glide-and-land operation with a winged vehicle that we had done many times.

Even so, we were always concerned about what could go wrong.

More On This...


I reported three hours before entry to the same building as Mission Control but a different room. It was crowded, and since the room was shaped like a U, the sides couldn’t communicate except through the Mission Evaluation Room (MER) communication system. I had no seat -- I roamed the room watching the subsystems I was responsible for over my engineers’ shoulders.

For the first two hours, entry was as uneventful as ever. The excitement usually starts with the re-entry burn, an hour before touchdown. This slows the Orbiter just enough so that on the other side of the Earth, it will be low enough to use the friction of the atmosphere to complete the re-entry. There’s no turning back after this point.

After watching this burn, it was my normal practice to go to the Auxiliary Power Unit (APU) and Hydraulics console. We turn on the APUs to supply high pressure hydraulic fluid to all the aero-surfaces for re-entry control.

I stood at the console with my wife, who also worked Orbiter, when we noticed we were losing hydraulic sensors on the left wheel well area. I remembered that potential foam hit during ascent, and asked my wife where it was. “The left side,” she said worriedly.

She immediately left to see if there was a simple explanation for the sensor data we were seeing -- maybe a common signal conditioner failing. There wasn’t.

A sinking feeling enveloped my body, but we were still seeing data and everything should be fine. Still, the waiting got longer. And then we lost communication with the Orbiter. My wife had already left to talk to MER management and to contact her task lead and others.

The silence in the room was frightening, broken only by the occasional question -- “Are they okay?”

When the truth finally sank in, professionalism took over. The MER went into lock-down mode, as did Mission Control. Follow standard procedure. Protect the data. I had seen this during the Challenger accident years earlier. Within hours the Shuttle Program and Orbiter Project were organizing into teams to figure out what went wrong.

After calling my division to let them know what happened, I led a team to map all the vehicle data.  The Orbiter has thousands of measurements that are continuously sending data to the ground every second.  What was this data telling us? My wife co-led the timeline team, reconstructing the sequence of events.

Our daughter would not see her parents until late at night. Our niece was visiting Johnson Space Center that day with her school class from New Orleans. We had planned to meet up with her and her class after the Orbiter landed. She went home without seeing us that night.

Limited instructions were given, but we all knew what to do. Work the data, understand the sequence of events. Find out what went wrong.

I had the luxury of a whole division at my disposal, about 90 people. My wife’s timeline team gathered data from all the experts in the room and across the country, chronologically documenting it. Others begin analyzing the data almost immediately to look for clues; later test teams would try to recreate the events from this data and analyses. No one got much sleep.

What is “data mapping”? What do they expect? Somehow 24 hours raced by and I was running a 100-plus temperature -- some bug. Poor timing. I needed to direct the team.

I decided I needed a graphical sequence of the data to show what sensors were lost when, and where the sensors were on the vehicle. The team proceeded to outperform my vision: Our chronological sequence was on the front pages of key reports and magazines within a week of the accident.

Because I had a frontline view of the data, I was also assigned to help lay out the most probable scenario that resulted in the loss of Columbia. By timing the end of the data transmissions, forensic analysis of the recovered parts, and various vehicle measurements eventually made a compelling case for a breach of the left wing on the leading edge at panels 7 & 8.

This is only one person’s experience, of course, and many people made many contributions. We did it because we were professionals, but we also did it because the families of the astronauts had a right to known what happened to their love ones.

We were just there in case something out of the ordinary occurred.

Gene R. Grush and Joyce M. Seriale-Grush are retired NASA engineers. Gene was acting propulsion and power division chief at retirement and Joyce was orbiter chief engineer at the end of the shuttle program.