Sometime between Sunday and Monday, the short, unhappy life of Phobos-Grunt will end in a fiery plunge to Earth. Launched on November 8, the spacecraft was supposed to fly to the Martian moon Phobos, capture a soil sample, and return it to Earth for analysis. Instead, the probe wound up stranded in Earth orbit when engines failed to ignite in a pair of controlled burns that would have put it on a trajectory to the Red Planet.
Attempts by the Russian Space Agency Roscosmos to communicate with the spacecraft were only passingly successful, despite the assistance of the European Space Agency. Without the ability to maneuver the vehicle, ground control teams could only watch as the orbit degraded. Drag from the solar panels has turned the spacecraft in orbit so that as of January 1, it was moving backward, adding credence to one theory that the problem was triggered by a failure in the orientation systems.
Now, the $165 million spacecraft is poised to fall to Earth along with its 7.5 tons of hydrazine fuel, breaking up into around 20 fragments large enough to survive the trip through the Earth’s atmosphere.
It's a deeply unfortunate turn of events for the astronomy community, which had hoped to answer the question of whether Phobos is a true Martian moon or merely a wandering asteroid captured by the planet's gravity. It's a setback for China’s space program, whose Yinghuo-1 Mars orbiter was piggybacking on the mission. It's a little bit of heartbreak for the Planetary Society whose Living Interplanetary Flight Experiment was designed to investigate whether biological organisms can survive interplanetary transit. More than anything, it's a blow to the Russian space agency, the latest in a string of troubled outings.
The scuttled mission had to leave everybody on the team bitterly disappointed, and more than a little uneasy about being saddled with responsibility for the issue. But to blame the fault on shadowy saboteurs, as Roscosmos agency director Vladimir Popovkin hinted in recent comments, squanders a valuable opportunity to learn.
Failure analysis is the time that engineers put on their deerstalker hats and turn into detectives, exhaustively studying every aspect of a project to try to determine what went wrong. In the process of designing and testing a cold encoder for the Curiosity Mars rover, for example, the Jet Propulsion Laboratory engineering team found the device failing lifetime tests again and again. Each time, they performed in-depth analysis and each time they learned a little bit more. Ultimately, they elected not to fly the cold encoder on this particular mission, but they continue to develop the technology, buoyed in part by what they learned throughout their efforts. Sure, success beats failure, but both can be valuable.
Much engineering involves taking on unique design challenges. Not every project is going to work out, but acknowledging failure and then moving on to analysis is the first step in extracting something positive from the experience. In the aftermath of the Challenger disaster, for example, the investigation not only revealed a design flaw involving o-rings in the solid rocket booster, but it also uncovered systemic program faults that needed to be corrected for the sake of the agency's future.
If nothing else, engineering teaches us to take an unflinching look at a situation, accept what's there, and use that information to take the project to the next level. In some ways, every success is built on the foundation of the failures—and failure analyses—that have come before. Let's hope Roscosmos uses this opportunity to better understand systemic causes of the failure so that the efforts of their very talented scientists and engineers are not wasted.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.