I wrote an editorial for
Aviation Week & Space Technology,
a major aerospace publication, concerning Glenn’s mission. The piece was published in the September 21, 1998, issue. I closed it with these comments: “It bodes very poorly for any team when management needlessly accepts risk and then silently hopes for the best. It’s little things like this that ultimately pave the road to another
Challenger
…”
Five years later, in 2003, another commission would investigate the
Columbia
tragedy. Its conclusions would hauntingly mirror those of the
Challenger
Roger’s Commission—cultural issues within NASA had led to
Columbia
’s loss. No one should have been surprised. The lessons of
Challenger
had been forgotten long before
Columbia
was dust falling through the Texas sky. Watching Mr. Glenn strap into the shuttle was proof of that.
*There were exceptions. Charlie Walker flew three missions and many of the Spacelab PSes flew multiple times too.
*Again, there were exceptions. Most astronauts felt the European Spacelab and Canadian astronauts, as well as McDonnell Douglas’s Charlie Walker and a handful of other part-timers, were valuable additions to crews.
Chapter 25
The Golden Age
If ever there was a Golden Age for the space shuttle program, that period was 1984 to
Challenger
. In those two years there were a total of fifteen successful shuttle missions, ten of those coming in the final twelve months. The shuttle would never again achieve that flight rate. In April 1985,
Discovery
and
Challenger
were launched only seventeen days apart, another STS record. (The seventeen-day record marks the interval between successful launches.
Challenger
’s final mission was launched only sixteen days after a
Columbia
mission.) The missions were coming so fast that shuttles were simultaneously being readied for launch on pads 39-A and -B. KSC was looking like a spaceport out of science fiction.
The history recorded in this Golden Age was remarkable. It included the world’s first tetherless spacewalks by jet pack–wearing astronauts, the first on-orbit repair of a satellite by spacewalkers, and the first retrievals and return to earth of malfunctioning satellites. With its fifty-foot-long robot arm and spacewalking astronauts, the shuttle repeatedly demonstrated its unique ability to put man to work in space in ways never before possible. It was also during this period that the orbiters
Discovery
and
Atlantis
joined
Columbia
and
Challenger
to complete the four-shuttle fleet. And that fleet showed its muscle: Twenty-three satellites, totaling 142 tons of payload, were deployed from shuttle cargo bays. Just as NASA had promised, the shuttle was doing it all…launching commercial satellites, DOD satellites, and science satellites.
On the surface things looked glorious for NASA. But there was a problem: Getting to the twenty-plus missions per year that would give the shuttle a cost-competitive advantage over other launch systems was proving to be a much more formidable task than expected. The shuttle was a voracious consumer of man-hours. After every landing there were thousands of components that needed to be inspected, tested, drained, pressurized, or otherwise serviced. There were 28,000 heat tiles and thermal blankets on the vehicle. Each one had to be inspected. Mission-specific software had to be developed and validated. Payloads had to be installed and checked out. Severely hampering every turnaround was the lack of spare parts. Just-landed orbiters were being cannibalized of their main engines and other components to get the next shuttle ready. The necessary requirement to meticulously document all work was another drag on vehicle turnarounds: Just tightening a screw generated multiple pieces of paperwork. The joke within the astronaut corps was a space shuttle could not be launched until the stacked paper detailing the turnaround work equaled the height of the shuttle stack…two hundred feet.
At just ten missions per year the shuttle was driving the system to its knees. The message was the same everywhere: “I need more people. I need more equipment. I need more spare parts.” But NASA didn’t have the money to buy these things. While commercial customers offset a portion of the expense, the cash flow was nowhere close to making the shuttle the pay-as-you-go enterprise promised years earlier to Congress. Significant taxpayer money was needed to underwrite the program, and those funds were fixed in the budget. The launch rate had to be doubled with the funds available. The end result was that more was being demanded of the existing manpower and equipment to achieve a higher flight rate. Everybody had a story about how this was overwhelming the various NASA teams. I recall being with an MCC controller when his boss brought in more work for him. The controller objected, “I haven’t had a day off in six weeks. My wife and kids don’t know who I am.” The supervisor was sympathetic but had no other option. “We’re all in the same boat. I don’t have anybody else. You’ve got to do it.” I could see it in both of their faces. They were exhausted, totally burned out. And they weren’t the exception. In many areas NASA only had a first string. There was no “bench” to call on for substitutes. One of our STS-41D prelaunch hangar tests of
Discovery
had been botched for that reason. The first string had been supporting the pad checkout of the shuttle being readied for the next launch, so the contractor had scraped together a team for us from God-only-knew-where. One of the technicians had apparently been called from home because he arrived in the cockpit with the smell of alcohol on his breath. It was an outrageous violation and Hank Hartsfield confronted the man’s supervisor about it. He apologized for the intoxicated worker as well as for the entire test debacle, adding, “I don’t have enough people to cover everything.”
The story was no different for the engineers at the SRB Thiokol factory in Utah. The pressure to keep flying was hammering them even while they were struggling with a major anomaly. The O-ring problem first seen on STS-2 had not gone away. In fact, it had gotten worse. Beginning with STS-41B, launched in February 1984, and up to
Challenger,
only three missions did not have O-ring problems. The other fifteen flights of this period returned SRBs with eroded O-rings. Astonishingly, in nine of these fifteen flights, the engineers had recorded “blow-by,” in which heat had not only eroded the primary O-rings but, for very brief moments, had gotten past those rings. On STS-51C, the blow-by had been exceptionally significant. That mission had launched in January 1985, after the stack had waited on the pad through a bitterly cold night. Engineers suspected that cold had reduced the flexibility of the rubberized O-rings, which, in turn, had allowed a more significant primary O-ring leak, resulting in a more significant blow-by. But in all cases none of the observed erosion equaled what had been recorded on STS-2’s damaged O-ring, and that mission had been fine. In effect the STS-2 experience had become the yardstick against which all following O-ring damage was being measured. If the damage was less (and it always was), then it was okay to continue flights. In what would later be defined as “normalization of deviance” in
The Challenger Launch Decision
by Diane Vaughan, the NASA and contractor team responsible for the SRBs had gotten away with flying a flawed design for so long they had lost sight of its deadly significance. The O-ring deviance had been normalized into their judgment processes.
There were a handful of individuals who resisted this normalization of deviance phenomenon. Thiokol engineer Roger Boisjoly was one. In a July 31, 1985, memo to a company vice president, Boisjoly expressed his concern about continuing shuttle flights with the SRB O-ring anomaly. He concluded the memo with this prophetic sentence: “It is my honest and very real fear that if we do not take immediate action to dedicate a team to solve the problem with the field joint [a reference to the O-ring] having the number one priority, then we stand in jeopardy of losing a flight along with all the launch pad facilities.” Boisjoly feared a catastrophic failure at booster ignition that would not only destroy the shuttle and kill her crew, but would also destroy the launchpad.
Another engineer, Arnold Thompson, wrote to a Thiokol project engineer on August 22, 1985: “The O-ring seal problem has lately become acute.”
An October 1, 1985, interoffice Thiokol memo contained this plea: “HELP! The seal task force is constantly being delayed by every possible means.” In his last paragraph, the memo’s author, R. V. Ebeling, obliquely highlights the major problem of the operational STS…not enough people. “The allegiance to the O-ring investigation task force is very limited to a group of engineers numbering 8–10. Our assigned people in manufacturing and quality have the desire, but are encumbered with other significant work.” He finished his memo with the warning, “This is a red flag.”
Another indication of the crushing workload being borne by the Thiokol engineers is found in an October 4, 1985, activity report by Roger Boisjoly. “I for one resent working at full capacity all week long and then being required to support activity on the weekend…” The operational shuttle program was devouring people.
Astronauts remained ignorant of the O-ring bullet aimed at our hearts. It was never on the agenda of any Monday meeting. None of the memos being circulated at Thiokol made it to our desks. But there were other things happening in the Golden Age of which we were aware—terrifying near misses.
On April 19, 1985, as
Discovery
landed from STS-51D at KSC, the brake on the inboard right-side wheel locked on, resulting in severe brake damage and the blowout of the tire. Unlike large aircraft, which have engine trust-reversers to aid in stopping the machine, the shuttle is completely dependent on brakes…and it lands 100 miles per hour
faster
than airplanes of comparable size. (A deployable drag chute was added in 1992.) When a shuttle touches down, it is a hundred tons of rocket, including several tons of extremely dangerous hypergolic fuel, hurling down the runway at 225 miles per hour. While the shuttle runways at KSC and Edwards AFB, at 3 miles in length, are sufficiently long for stopping, they are only 300 feet wide. A perfectly landed shuttle is only 150 feet from an edge, an eye blink in time at those speeds. It was a minor miracle that
Discovery
didn’t experience directional control problems as a result of the blown tire and careen off the runway.
STS-51F experienced the second engine-start pad abort of the shuttle program. While not really a near miss, pad aborts have the potential to become dangerous. Afterward, I watched that crew put on their Right Stuff, no-big-deal faces for the press, just as we had done following our 41D pad abort. Astronauts are great actors.
STS-51F also became the first shuttle mission to perform an ascent abort when
Challenger
’s center SSME shut down nearly three minutes early. It was later determined that the malfunction was due to two faulty engine temperature sensors. There had been nothing wrong with the engine. With only two SSMEs, the crew was forced into an Abort to Orbit (ATO). Fortunately, this was the safest of aborts. The shuttle had been high enough and fast enough at the time of the engine failure to limp into a safe orbit on its two remaining engines. Had the engine failure occurred earlier, the crew would have faced a much more risky 15,000-mile-per-hour, thirty-minute TAL to a landing at Zaragoza, Spain.
Having experienced both an engine-start abort and a powered-flight abort, the 51F crew had gone through ten lifetimes of heartbeats. After they returned, astronauts joked that a cocked, loaded gun pointed between the eyes of any of them would not have elicited the slightest fear response. The mission had desiccated their adrenal glands.
STS-61C (Congressman Nelson’s flight), the last mission prior to the
Challenger
disaster, experienced a pair of bizarre and dangerous malfunctions even before it was launched. During a January 6, 1986, countdown attempt, a temperature probe inside one of
Columbia
’s propellant pipes broke off and was swept into a valve that controlled fluid flow to an SSME. Unknown to anybody, the valve was jammed in the prelaunch open position. Engineers in the LCC noted the temperature sensor was not responding, but erroneously assumed it was due to an electronic malfunction. It had not occurred to anybody that the probe might have actually broken free and was floating around in
Columbia
’s guts. The countdown continued using a backup temperature sensor. The mission was ultimately scrubbed for other reasons and the valve jam was discovered in the countdown reset. Had
Columbia
launched, there was a good chance the jammed valve could have caused a turbo-pump to overspeed and disintegrate during the engine shutdown sequence at MECO. The resulting shower of hot steel inside the engine compartment would probably have trashed the vehicle hydraulic system, dooming the crew on reentry.
During the same 61C countdown, a malfunction of a different valve (this time on the launchpad side of the plumbing) caused the drain back of a large amount of liquid oxygen from the gas tank. For a variety of technical reasons, the LCC had remained ignorant of the lost propellant. The shuttle very nearly lifted off without enough gas to reach its intended orbit. The crew’s first indication of a problem would have come when all three SSMEs experienced a low propellant level shutdown somewhere over the Atlantic. How high and fast they were at that moment would have determined whether the crew lived (TAL, AOA, or ATO abort) or died (contingency abort). Again, the day was saved when the launch was scrubbed for unrelated reasons and the drain-back problem was discovered in the turnaround.
These near misses should have been warning flags to NASA management that the shuttle was far from being an operational system. They were indicative of the types of problems that occur in the early test phase of any complex aerospace machine. Every military TFNG had seen it happen in new aircraft systems they had flown. In fact, we were used to having urgent warnings appear on our ready-room B-boards concerning newly discovered failure modes on aircraft types that had been seasoned in decades of operations. It is the nature of high-performance flying. The machines are extremely complex and operate at the edge of their performance envelopes. And the space shuttle was about as high-performance as flying got. There were certainly more surprises awaiting us in its operations. In fact, if the shuttle program should survive for a thousand flights, I am certain engineers will still be having occasional moments of “Holy shit! I never expected to see
that
happen.”