Washington

Getting it right? NASA hopes not to confuse metric and imperial for its Mars Polar Lander Credit: NASA/JPL

Confusion between imperial and metric units was to blame for the loss of the US space agency NASA's Mars Climate Orbiter, according to project managers at the Jet Propulsion Laboratory.

The spacecraft was lost on 23 September, when a targeting error sent it into, rather than safely above, the Martian atmosphere as it was about to enter orbit (see Nature 401, 415; 1999). The spacecraft was either destroyed or skipped off the atmosphere into an unknown orbit around the Sun.

A preliminary analysis has revealed that the builder of the spacecraft, Lockheed Martin Astronautics of Denver, had been supplying information about routine propulsion manoeuvres in imperial units. But NASA space navigators assumed throughout the nine-month voyage from Earth to Mars that the acceleration data were in metric units, resulting in a large targeting error when the orbiter reached its destination.

The spacecraft's operators performed a major manoeuvre to correct its trajectory a week before it reached Mars, and thought they were on course at that point. But subsequent, smaller propulsion manoeuvres were required every 15 hours or so to counteract the energy built up in onboard ‘reaction wheels’ used to orientate the spacecraft.

Errors in these routine smaller manoeuvres are believed to have resulted in the large deviation when the spacecraft reached Mars. Three separate review teams are currently examining the mishap in detail, with reports due in mid-November.

The first priority for NASA is to ensure that its Mars Polar Lander, also built by Lockheed Martin and due to touch down on the planet's surface on 3 December, does not have the same problem.

A trajectory correction planned for the lander early this month has been postponed until project engineers have a better grasp of the situation, says Noel Hinners, a vice-president at Lockheed Martin, which monitors the spacecraft from its operations centre near Denver. “We don't want to do anything until we understand this one,” he says.

Unlike the Mars Climate Orbiter, the lander does not use reaction wheels to control its altitude during the interplanetary cruise, according to Hinners, so the same problem is unlikely to occur. But having made one costly mistake, Lockheed Martin is carefully scrutinizing the two missions it currently operates for NASA—the Mars Polar Lander and the Stardust comet sample return—to see if “there are any other things that could have slipped through”.

The company has long experience of controlling NASA spacecraft, dating back nearly a decade to the Magellan Venus mission. But a recent external review of its Space and Strategic Missiles Sector, of which the astronautics division is a part, found systematic problems with quality control, with overemphasis on cost savings, and with management of subcontractors and parts suppliers.

The confusion between imperial and metric that doomed the spacecraft originated with a Lockheed Martin subcontractor, says Hinners. The supplier of the spacecraft's propulsion system provided accompanying data for its system in imperial units, and Lockheed Martin simply incorporated those data and passed them through thesystem.

”We should have caught [the mistake] then and there,” he says. The propulsion system for Lockheed Martin's Mars Global Surveyor, which is now operating successfully in Martian orbit, was built by another subcontractor who used metric units.

The embarrassing mistake also underlines the lack of uniformity in the US space sector. While virtually all scientists use metric units, many US engineers, both inside and outside the space programme, use imperial units, converting them when necessary. If the United States had converted to the metric system when the subject was debated more than 20 years ago, muses Hinners, “it might have saved this mission”.