A program executive was a project’s in-house representative at headquarters. If you were a mission project manager at JPL, you didn’t want constant calls from congressional affairs, the NASA chief financial officer, or the office of the chief engineer. Enter Joan, whose job it was to run interference for the project and act as its advocate—and enforcer—at headquarters. It fell to her to keep projects in line with agency expectations. She was to Europa management what Curt was to its science team.
Her first Europa meeting at JPL was in 2010. Joan hadn’t known what to expect. Engineers are prickly already, and the lab could sometimes instill employee paranoia when it came to the priorities of NASA. But across the agency—at every field center and partner office, from the redwood forest to the Gulf Stream waters—there was one catholic conviction: everyone at headquarters is an idiot.477 A speed bump. In the way. Not at All Welcome Here. Before she joined headquarters, Joan had spent twenty years equally devout in that belief. And now she was on the other side. Everyone had only just taken their seats in the conference room, Tom Gavin presiding as per custom from the head of the table, when she walked in.
Tom saw her, stood up, introduced her to everyone, and then introduced himself to her.
Can I get you some coffee? he asked.478
Joan was stunned. The engineers were stunned. The coffeemaker was stunned. Anywhere else, any other meeting at any other center, and she would have been told where the coffee machine was, and just take your time. But here was Tom—in Gavin Tower, no less—and he didn’t just get her a cup of coffee—he found a mug, washed it, and made a fresh pot. And this was really happening and the jaws of his staff were slack, their eyes wide. Tom Gavin was—well, he was Tom Gavin! And he was treating this . . . headquarters creature . . . so . . . so regally. WHO WAS SHE?
Joan would come to understand that Gavin wanted emphatically to establish her as worthy of extraordinary respect by the team. It made her job going forward—one where she would sometimes be the bad guy—a lot easier. It also made his retirement easier, knowing things would be right with headquarters once he punched out for the day one final time.
Tom retired, again, in 2012. He was seventy-three, and he just didn’t have the energy to run a flight project any longer. Gavin told Charles Elachi, the lab director, that it was time to find a real eight-day-a-week flight project manager.479 During Tom’s transition out, Keyur Patel—who Karla Clark believed wanted the job but would be bad at it—stepped in briefly as manager of the Europa Habitability Mission. He didn’t last long. Headquarters was happy with Keyur’s selection by JPL management, but the choice did not sit well at the Applied Physics Laboratory. From Tom Magner’s vantage, to suggest that Patel had thinly veiled contempt for APL would have been to suggest the presence of a veil, and Magner felt like his full-time job had become saving the alliance. Ultimately, Elachi offered Keyur the top spot over the Deep Space Network, and Keyur took it, to the relief of many on the project leadership team.
Asked for his recommendation, Tom Gavin endorsed Barry Goldstein for the job of project manager. Barry had done the Phoenix lander—a hard, major league mission—was a manager of enviable ability, was the obvious choice. He even had familiarity with Europa, having sat on the technical review board that examined the orbiter, lander, and Clipper concepts, and like everyone else on the board, was just besotted by the multiple flyby concept. Of course Barry wanted to run the project!
One person who was not asked for an opinion on the matter was Bob, who, he felt, as project scientist, was supposed to have at least a say in who his project manager would be. Indeed, he had been told explicitly that he would be consulted when the time came to choose a project manager. When he went to lab management with this complaint, their response was: Whatever.480
Since his days as Cassini project scientist, Bob had tried to assert and protect the authority of the role in lab missions. His job had one portfolio, and the project manager had another, but any management decision affecting science had to go through him. That was the whole point of the project scientist position, and it was the only way to protect the scientific integrity of the mission. Karla got that, and Tom really got that, went so far, in fact, as to take the organizational chart—the hierarchy of who reported to whom—and put his name and Bob’s on the same line, with a comma between the two.481 But Barry . . . there was never any ambiguity in Barry’s mind about his role vis-à-vis Bob.482 There weren’t any commas on the new org chart.483
And Barry wanted to change some things on Europa Clipper before they became problems. The spacecraft was to use a power system called the advanced Stirling radioisotope generator, which was being developed at the NASA John H. Glenn Research Center in Cleveland.484 Though the Stirlings promised greater efficiency than traditional thermoelectric generators while using less plutonium (which was ever in short supply), they were still in study and as a result presented ridiculous levels of technical risk to the mission. Barry told the team it could either get solar panels to work—a relatively new concept in outer planets exploration, enabled by breakthroughs in highly efficient solar cell technology—or they could use traditional radioactive power sources, but, listen: Quit trying to make Stirlings happen. They won’t fly on this spacecraft.
Then there were Clipper’s articulated antennae—a high-risk feature for low-priority science. If it moves, it can break. (Ask the Galileo team.) You bolt down something, you’ve eliminated a list of lurking mechanical misfortunes. So once again: find a way to do it with a fixed antenna, or it’s out.
The most pressing problem, perhaps, was planetary protection. The idea had been to just remain relaxed about the whole thing. Let the spacecraft marinate in microbes during development, and, at the very end, submit the spacecraft for “dry heat microbial reduction,” i.e., stick it in an oven and bake at two hundred thirty degrees Fahrenheit for fifty hours.485 Politely speaking, Barry did not like that idea. It had worked in the seventies for the twin landers Viking, but any time you exposed a sophisticated, sensitive spacecraft to extreme temperatures, you introduced the risk of damage, obvious or imperceptible.486, 487 When he had managed the lander Phoenix, its sample collection arm—a relatively tiny part of the spacecraft—had to be baked, and those three thousand minutes of terror were enough for one project manager’s lifetime.488 To do it on the entire spacecraft? No, thank you.
These were meaningful changes Barry was insisting upon, and he knew he was asking a lot. But he brought implementation experience to the table, had spent his entire career under pressure, delivering spacecraft on a schedule, and that mattered, went a long way with engineers. It was one thing to make demands, but quite another if your team knew that you had done it yourself, that you had lived through what you were asking them to do, and that you were doing it even now.
BRIAN COOKE ONLY knew he didn’t want to drive the train. When he was in the fifth grade, he decided to be an engineer when he grew up. His understanding until then was that an engineer wore the hat, leaned out the locomotive window, and blew the horn. Somewhere in there was this business with a shovel and coal. But no, it was explained to him: an engineer takes things that are sitting around and makes better things out of them. Well, he heard that and he was in! Brian eventually attended Virginia Polytech Institute and State University and in 1995 earned an undergraduate degree in multidisciplinary engineering.
He had at the time a friend who had been working an internship at JPL and needed to return to Pasadena. He didn’t want to make the drive alone, so he asked every girl he knew if she might want to join him on the transcontinental trek, VA to CA, and none said yes. Out of options, he asked Brian to ride shotgun, and Brian said yes, but on the condition that he (i.e., Brian) could go rollerblading on Venice Beach. When they arrived in town, however, it was raining—purposefully—vigorously—indefatigably—and the rollerblades sat unused in the corner of his hotel room.
His friend finally took pity on Brian and offered to bring him on a tour of the laboratory. By
the time it was over, Brian had four job interviews lined up. (He had that kind of personality.) But because he had come to Pasadena strictly to rollerblade, he had packed only Tevas, t-shirts, and shorts for the trip, and showed up for said interviews wearing exactly that. Still, he was hired straightaway. (JPL was that kind of place.)
Cooke was soon developing command sequences to send to Cassini for execution at Saturn. Three years later, he was hired to work on GALEX—the Galaxy Evolution Explorer, a small orbital telescope being built on an absurdly small budget: one hundred fifty million.489 (Its price tag customarily would have been in the large percentages of a billion, but this was the Faster-Better-Cheaper era. You just made it work for less.) GALEX launched, was a success, and filled a critical gap in the observational abilities of astrophysicists, but what its engineers really bragged about over beer was that one of the images taken by the telescope became a stock photo in the space screensaver for Microsoft Windows.
Brian next worked on the Dawn spacecraft, which would fly to Vesta and then to Ceres. It would be the first spacecraft to visit a dwarf planet (Ceres), and the first to enter orbit around two separate objects. Brian was the project verification and validation engineer on that one. He made sure it did what it was supposed to do, to spec—with no surprises come liftoff and arrival—and was even tapped to serve as launch director. Before it flew from Earth, it was he who said, Dawn is go for launch, which, I mean, was pretty cool.490 Twelve years now into his career, he had a good sense of what the lab was doing right, what it was doing wrong, and he had a burning desire to go make his own mistakes. It was the first time he was really able to say, “If I were in charge, I would do things a little differently,” while having experience enough to mean it.
He first came to know Tom Gavin while working as project system engineer on the Kepler space telescope. As Brian was the lead technical authority, part of his job involved assessing engineering issues encountered by the project and reporting regularly to a council composed of the head of safety and mission assurance; the head of astronomy; the head of engineering and science; and the head of the council, the associate director of flight projects: Tom Gavin.
Brian called it the four-headed dragon. You would walk to Building 301, go up to the fourth floor, room 427—it was one of those conference rooms with the boardroom layout—and you knew exactly what to expect. The four heads were well known for their even temperament: they were always angry. And there was Tom Gavin at the end of the table, twenty-five feet away, glaring at you, the other three seated adjacent to him, scowling, all totally unimpressed by anything about you. You’re presenting opposite them, and it feels almost ritualistic, an offering. The dragon fed on technical risks and issues. Your spacecraft-in-development ran into a potential problem—say, a transient signal was discovered and, under very specific conditions, could cause some unintended behavior—and you briefed the council on it, and they evaluated your solution. Of the four members of the panel, Tom was the worst. He would get spun up, and his . . . enthusiasm . . . would spin up the other three, and they would start to dig in—hard. You never wanted to go in unprepared. You identified a risk item, and you explained the solution in excruciating detail, and you also brought fifteen pages of backup. And they would go through every single page in your supporting documentation. You were walking them through schematics, through circuit diagrams for your exact fix, what work was left to do, the details of said work, schedules, milestones. Between the four heads, they had seen everything, success and failure, and knew when you were tap-dancing up there. You couldn’t just wing it, and you couldn’t deliver a “good enough” solution, because they knew what would and would not work. At the end of your presentation, they either gave you a thumbs-up or a thumbs-down, gladiator style. Their collective intuition, Brian came to see, bordered on clairvoyant and was one of the secret sauces of the laboratory that made it work so well.
ON FEBRUARY 10, 2010, while waiting in the recovery room after his wife had given birth to their daughter, Brian pulled out his cell phone. He was standing at the window looking out from the hospital, and there was snow on the San Gabriel Mountains, and for whatever reason, he thought of Tom Gavin and called him. Cooke had heard that the Europa team was restructuring, and he wanted to put his name out there if an opening appeared for a technical lead. Tom was noncommittal. He made Brian sell himself for a few months before finally bringing him on board, and after Karla Clark and Rob Lock, the lead system engineer, were out and the Jupiter Europa Orbiter died with the Decadal, Gavin had Brian restart the study team and get moving on the Europa Habitability Mission concept. From the start, the new spacecraft went through a series of redesigns, each in search of what engineers called “elegance.”491 Elegance had no definition, exactly. You knew it when you saw it. Spacecraft feng shui. It wasn’t aesthetics, necessarily—pleasing to behold in the traditional sense: sleek lines and sweeping curves—the starship Enterprise. Science vessels tended often to look ungainly and hodgepodge; parts stuck together. This was because, once in space, they had no need for aerodynamics, never actually confronting air. Rather, it was engineering elegance: the ability to meet objectives with grace and style. You had to maintain mass efficiency while distributing the science instruments in such a way that they would not interfere with one another. This was all done in computer software. During pre-project planning, you worked only with a straw man payload; you knew it carried a magnetometer, but you didn’t know which magnetometer. Each would be made to order by some other institution, which meant quirks and necessary accommodations. Beyond that, you had to consider the construction of the thing, and potential problems down the road. For example: an early Europa Habitability Mission concept had radiation-sensitive avionics embedded in a vault that was itself tucked inside the propulsion system and surrounded by fuel and oxidizer tanks. The tanks provided additional radiation shielding. It was brilliant! Unless you had a problem with avionics late in the spacecraft build—how would you get to them?
Each spacecraft concept started with a cylindrical body. The vessel at launch would be affixed atop a rocket pushing it into space from the bottom; a tall cylinder was thus as mass efficient as you could get. Various parts of the spacecraft would be moved around, repositioned, creatively mounted, and tested. Even in the earliest stages of design, the software produced models of adequate fidelity for engineers to conduct structural analysis, with answers accurate to eighty percent or higher—easily enough to make large-scale strategic and architectural decisions. The devil, as ever, was in the details.
When the multiple flyby concept study moved from APL to JPL, Tom had insisted that the spacecraft be made modular. Rather than construct a monolithic vessel—the way you might build a house—he wanted it designed as a series of mechanically separable elements, built independently and in parallel, each tested to one hundred percent comprehensiveness and merged late in the development process with high confidence that the system would work. This was how airliners were built. You would need, then, only limited testing of the fully assembled spacecraft. A further advantage was that if you ran into problems with any one element of the spacecraft, all work didn’t stop.
Once Barry came on and the Stirling generators were removed, the team initially went with multimission radioisotope thermoelectric generators, but engineers latched quickly on to using solar panels instead. For something the size of Europa Clipper, solar simplified a lot of things. Just avoiding the nuclear launch approval process made it worth studying. With solar, no one would chain themselves to the fences of Cape Canaveral at launch. Only one other outer planets mission had ever attempted to use solar power, however—the spacecraft Juno—and Juno wasn’t yet orbiting Jupiter, so it was a bit early yet to call it a success. Furthermore, unlike the solar cells of Juno, whose orbit avoided the worst of the Jovian radiation belt, Europa Clipper’s would have to fly stridently through the belt and survive.492 And if the radiation didn’t kill the cells, the temperature might: a Europa spacecraft would follow
an equatorial orbit, unlike Juno, which meant that half the time the largest planet of the solar system would separate the sun and the spacecraft. Not only would you be running on batteries during that time, but while on the far side of Jupiter, temperatures would plunge to cryogenic levels.493 This would need to be a hearty spacecraft indeed.
Moreover, solar panels introduced complications to the science payload. There were questions about their “magnetic cleanliness.” Solar arrays were notorious sources of electromagnetic noise because of their size, composition, and the currents running through them, which could impair the integrity of the magnetometer.494, 495 Electrostatic discharge could interfere with the plasma instrument. Radiofrequency noise could affect the ice-penetrating radar. And the solar panels would need to be big—very big. The core Europa Clipper spacecraft would be about the size of a small car. The solar arrays would make the structure the length of a basketball court.
The Mission Page 35