When Challenger exploded on 28 January 1986, it crushed many dreams of America’s space programme. The Shuttle was expected to open new frontiers and deliver people into orbit with unrivalled frequency and reliability…but herein was one of its fatal flaws. Fifteen missions were planned in 1986, followed by 24 in 1987, and it was long feared that this monstrous launch rate would push the Shuttle’s workforce beyond its limits and drive the fleet to its knees. The Rogers Commission would harshly criticise not only the technical causes of Challenger, but also the managerial cancers that enabled it to happen. Schedule pressure was one such cancer and in few other areas was its oppressive breath more acutely felt than in NASA’s plan to launch two Shuttles, within five days, in May 1986.
Twenty-six years ago, this week, Challenger and her sister Atlantis should have been on Pads 39B and 39A, primed and ready to send a pair of spacecraft to Jupiter. One of them, Ulysses, would employ the giant planet’s gravity to slingshot it out of the ecliptic plane and explore the Sun’s polar regions, whilst the second, Galileo, was destined to enter orbit around Jupiter itself. With only a few days available in the ‘launch window’, the Ulysses flight (Mission 61F) would fly on 15 May and the Galileo flight (Mission 61G) would follow on the 20th. Both would last four days, meaning that one would blast off less than 24 hours after the other had landed…and both would highlight an appalling game of Russian roulette that NASA played with its astronauts and a pair of multi-billion-dollar national assets.
“As with any flight,” said 61F commander Rick Hauck in a NASA oral history, “if everything goes well, it’s not risky. It’s when things start to go wrong that you wonder how close you are to the edge.” The two missions became known as the ‘Death Star’ flights, because their risk encompassed far more than the obvious dangers associated with a Shuttle launch. Ulysses and Galileo both carried plutonium-fed nuclear generators and both would be propelled from Earth orbit by a liquid-fuelled rocket, the Centaur-G Prime, built by General Dynamics. Nine metres long and four wide, the Centaur was a thin-skinned ‘balloon’ tank; it depended upon full pressurisation to support its weight. “If it were not pressurised,” explained Hauck, “but suspended, and you pushed on it with your finger, the tank walls would give and you’d be flexing the metal!” Notwithstanding the hazards, the Centaur provided more oomph to push payloads toward interplanetary targets than solid rockets could accomplish.
Historically, NASA’s safety rule of thumb dictated that no single failure should endanger the vehicle or its crew…but much of the Centaur’s pressure regulation hardware, worryingly, was non-redundant, with no backup capability. That was not all. The booster carried so much propellant – more than 16,500 kg of volatile liquid oxygen and hydrogen – that engineers and astronauts feared it might ‘slosh’ around in the tanks and impair the Shuttle’s controllability. If an emergency occurred, minutes after liftoff, necessitating a launch abort, the astronauts would have to dump the Centaur’s propellants whilst flying to a contingency landing site, either at Cape Canaveral or in Africa. To do that in a safe manner, the Shuttle would be fitted with redundant parallel dump valves, helium systems to control them and software to execute the abort. The valves were located on opposite sides of the orbiter’s aft fuselage and could remove the Centaur’s entire load within 250 seconds…but their close proximity to the main engines and orbital manoeuvring pods raised the spectre of possible fuel leaks or explosions.
In support of the Centaur, both Challenger and Atlantis would have undergone extensive modifications. As well as the dump valves and associated plumbing, they would have carried S-band data transmitters and flight deck control panels to monitor the booster’s status. A Centaur Integrated Support Structure at the rear of the payload bay would have held the booster and primed it for deployment. In mid-February 1986, Atlantis was to be rolled out to Pad 39B, carrying a ‘real’ Centaur and a mockup of Galileo, for several weeks of fuelling tests. She would then have returned to the Vehicle Assembly Building for the ‘real’ Galileo to be loaded and then rolled to Pad 39A in early April. By the middle of the month, Challenger would join her on adjacent Pad 39B with Ulysses.
As NASA’s newest orbiter, Atlantis had received many Centaur modifications during her construction and, ironically, Challenger would have undergone her own modifications in the aftermath of her tragic Mission 51L. Atlantis’ cargo – the Galileo spacecraft – was several times heavier than Ulysses and in January 1986, NASA accepted a recommendation to fly with her ‘Phase II’ main engines running at an untried 109-percent rated thrust.
The crews for the two missions were named in May 1985. Rick Hauck had already served as the astronaut office’s Centaur representative and it was thus unsurprising that he would command the first mission. He was joined on Mission 61F by Roy Bridges, Mike Lounge and Dave Hilmers. The 61G crew, meanwhile, comprised Dave Walker, Ron Grabe, John Fabian and James ‘Ox’ van Hoften. By September, Fabian had resigned and was replaced by fellow astronaut Norm Thagard. One of the reasons for Fabian’s departure was his conviction that NASA prized commercial respectability above operational flight safety. He spent enough time with the 61G crew to see a technician clambering onto the Centaur with an untethered wrench in his back pocket and another smoothing out a weld, then accidentally scarring the booster’s thin skin with a tool. In Fabian’s mind, it was bad enough that the Shuttle was carrying a volatile booster with limited redundancy, without adding new worries about poor quality control oversight and a lax attitude towards safety.
In fact, the Centaur dictated virtually every detail of the two missions. According to the 61F crew activity plan, released by NASA on 14 January 1986, Hauck’s men would launch at 4:10 pm EST on 15 May, with everything aboard Challenger kept to a minimum: from the number of crew members to the number of crew provisions…to the orbital altitude itself. Both flights would enter relatively low orbits of 170 km, simply because they needed the main engine performance simply to get the heavy Centaur into space. Additional weight savings were achieved by the absence of secondary experiments, an otherwise empty payload bay and even the food-preparation galley in the middeck was removed.
The presence of the Centaur also decreed the amount of time the astronauts had to deploy their precious cargo. Periodically, it was required to dump its boiled-off gaseous hydrogen in order to keep tank pressures within mandated limits. After too much time, it would have ‘bled’ so much hydrogen that the remainder would have been insufficient to perform the necessary engine firing to begin the voyage to Jupiter. Consequently, for both 61F and 61G mission planners scheduled three deployment opportunities, with the first occurring just seven hours after launch.
As the commanders of the two missions, Rick Hauck and Dave Walker spent much of 1985 questioning their own judgement about how many failure modes and problems they could live with. A few days before the loss of Challenger, they were working an issue pertaining to the redundancy of the helium actuators for the Centaur dump valves. “It was clear,” said Hauck, “that the [senior managers] were willing to compromise on the margins in the propulsive force being provided by the pressurised helium. We were very concerned about it. We went to a review board to argue this was not a good idea to compromise on this feature. The board turned down the request.” Severely defeated, and doubtless angry, Hauck returned to the astronaut office and gathered his crew together. Safety had been compromised, he told them, and if they wanted to resign from the mission, he would support them.
Of course, none of them did so. Years later, Mike Lounge rationalised their thinking. In the days before Challenger, NASA was still riding on the coattails of past glories, including Project Apollo, and the astronauts considered the Shuttle bulletproof. “We assumed we could solve all these problems,” Lounge told the NASA oral historian. As if to underline the point, the 61F crew were in a flight procedures meeting on the very morning that Challenger exploded, reviewing the techniques they might need to vent the Centaur’s propellants in an abort. “Until Challenger, we just thought the things would always work.”
Assuming an on-time start to Mission 61F, Challenger would have landed on the afternoon of 19 May and Atlantis would have blasted off 23 hours later, at 4:21 pm on the 20th. It would have marked the most rapid turnaround between launches ever seen in the Shuttle programme. Could it have been done? It is a question which brings positive and negative replies. Some experts told the Rogers Commission that a lack of spare parts would have brought the Shuttle to its knees by this point, whilst others were adamant that launching two Shuttles within a week of each other was achievable. “I’m convinced to this day we would have made the launch window,” said Marty Winkler of General Dynamics, “but it was a sprint to the finish. It was like the racehorse that overtakes you at the end.”
Yet the achievability of the twin missions reveals a great deal about prevailing attitudes of the time (“bulletproof,” to borrow from Mike Lounge’s words) and an incessant pressure to get Ulysses and Galileo away on time. Any delay beyond May 1986 would ground both missions until the next Jupiter launch window opened in June 1987…and this would have incurred a cost penalty in excess of $50 million. As early as January 1984, Centaur technicians at General Dynamics had been given free 28-month calendars…which ended in May 1986. Clearly, the company’s deadline to be ready for launch was as immovable and inflexible as a rock.
The Challenger disaster marked the end of the road for the Shuttle-Centaur partnership. In early February 1986, the Kennedy Space Center’s safety office refused to approve the booster’s advanced processing, citing “insufficient verification of hazard controls”. Cost overruns to the tune of $100 million also conspired against the Centaur and in June 1986 the project was cancelled by NASA Administrator Jim Fletcher. When Galileo and Ulysses eventually flew, in separate Jupiter launch windows, in October 1989 and October 1990, they did so atop less powerful, but safer Inertial Upper Stage boosters. One other part of the story endured beyond Challenger…for several of the astronauts remained together to fly two of the most historic missions of the late 1980s. Hauck, Lounge and Hilmers formed part of the STS-26 crew, which returned the Shuttle to normal flight operations, whilst Walker, Grabe and Thagard flew STS-30, which launched the Magellan Venus orbiter…another spacecraft originally baselined for the unlucky Centaur.