There are institutional differences as well. Unlike NASA, JAXA is not an independent government agency. Its funding comes from the Ministry of Education, Culture, Sports, Science and Technology, but its employees are not civil servants, and there are fewer safeguards against a "revolving door" tradition of employees moving back and forth between government and industry. The relationships between industry contractors and government overseers can be puzzling to Americans. I learned this first-hand, when Kibo's pressurized module was about to be shipped from the Mitsubishi factory in Nagoya, where it was built, to the NASDA space center in Tsukuba in 2001.
Japanese quality inspections conducted during the final phases of manufacturing sometimes have the tone of a religious ritual. Top company executives are on hand along with the engineers and technicians, and attendance by outsiders is strictly controlled. The Mitsubishi representatives suggested that I watch the inspection on a TV monitor in a remote office. I was stunned—had I traveled from Houston to Nagoya just to watch a ceremony on TV? This was supposed to be a real test, not a formality, and it was the last chance to raise concerns about the module before Kibo's hatch was closed and ownership was transferred from the manufacturer to the Japanese space agency. When I pointed out that NASA was required to verify that every bolt and fixture inside the module fit the standard astronaut tools on the space station, I was reluctantly allowed to participate as "Test Subject 2."
"Test Subject 1," I was surprised to learn, was an employee of Mitsubishi, not NASDA. His job was to go through the entire module, methodically fitting tools to every fixture—all under the watchful eye of his boss' boss' boss. Imagine the pressure he felt. One misfit would ruin the ceremony.
The inspection began with each person taking a place around a circle, each place marked with footprints. The hierarchy was crystal-clear. Non-participants stood behind the circle. A reader called out the first fixture and the tool to be used. One guide showed us the fixture as another guide pulled the proper tool out of the box and carefully handed it to Test Subject 1. He fit the tool on the fixture, then handed it to me so I could do the same. We both stated that the tool fit, and the concordance was duly marked by the recorder.
Over the course of many hours, Test Subject 1 and I tried every single bolt, nut, and fixture in the module. The vast majority were absolutely fine, better than most space hardware I have inspected. But every once in a while, Test Subject 1 would fit a tool and hand it to me, and I would have trouble making it fit. The offending fixture was duly noted and marked for repair. In the end, everyone agreed the test was a resounding success.
When I returned to Tsukuba, I asked about the policy of allowing contractors to do their own inspections, and was assured that space agency personnel would conduct a complete inspection once the module arrived at the research center. Of course, repairs would be more troublesome at that point than when Kibo was still at the factory, but to do otherwise could insult the contractor. As for me, at least I had made things less tense. The Japanese could consider the flaws that I found just overzealous pickiness by the NASA guy. And, even though repairing them really was required, it could be construed as a favor, not a reflection of deficient manufacturing.
None of this is meant as a criticism of Japanese space engineering. In fact, JAXA and its predecessor agencies have a solid record of successes, going back to the 1970s, in launching spacecraft for astronomy, solar physics, and Earth observation. But the switch from building satellites to building hardware for human spaceflight, with its more stringent safety requirements and need for tight coordination, is difficult for any nation, including the United States and Russia, which, nearly 50 years after their first spaceflights, still make the occasional mistake. And Japan has proven no exception.
My own shuttle mission, STS-72, offered a small example. While the Space Flyer Unit was a remarkable satellite and ultimately a success, it did have glitches. The first hint of trouble came when its solar arrays jammed while retracting. We had to jettison them, which took valuable time; the satellite could last only about an hour on battery power after the arrays were shut down. Without power, the onboard heaters would fail, causing the satellite's propellant lines to freeze and rupture. We couldn't bring a propellant-spilling satellite into the shuttle's bay; we'd have to abandon it.
Koichi Wakata had the job of using the shuttle's robot arm to bring the spacecraft into the cargo bay. All was going well until the last step, when the satellite wouldn't engage the locking latches. We could get two or three to lock, but not the required four. As he tried over and over, Koichi was the picture of cool concentration—no panic in that guy. But time was running out. Finally, with just five minutes of battery life remaining, he turned to us and said, "Here goes a slam dunk; be ready with those latches!" He revved up the arm and brought the satellite down like Karl Malone going to the basket. Sure enough, four latch lights lit, and we returned the satellite safely to Earth.
What had gone wrong? We never found out for sure. Most likely, a combination of weightlessness and exposure to the hot-cold cycles of space had slightly warped the satellite frame. We were lucky the warping was within the margins of what the latches could accommodate. It was another lesson in the complexities of space hardware, which—especially on large projects—often has to fit parts fabricated in different places at different times.