December 1998: NASA launched toward Mars a 638-kilogram probe named Mars Climate Orbiter, tasked with studying the climate of the Red Planet. Nine months later, the craft approached its target to enter orbit. It would never reappear on the other side of the planet.
The spacecraft weighed 638 kilograms. A featherweight for a mission that would nevertheless become one of the most studied cases in the history of space engineering.
What caused the mission to fail was not a mechanical malfunction, nor a classic trajectory error. Mars Climate Orbiter, one of the probes sent by the American agency toward the red planet, failed to achieve orbit, burning up in the Martian atmosphere or crashing onto the surface. The reason? An unrecognized divergence in units of measurement that remained invisible for months of flight.
- Mars Climate Orbiter crashed on Mars in 1999 due to a divergence of units between two engineering teams.
- Jet Propulsion Laboratory used the metric system, while Lockheed Martin calculated in Imperial units, creating an error factor of 4.45.
- A warning was detected but ignored, and trajectory corrections accumulated until the probe deviated entirely from its orbit.
Two Teams, Two Languages, the Same Figure
The investigation revealed that one group of engineers used the metric system, officially adopted by NASA, while a second group carried out all calculations using the Anglo-Saxon system—feet, inches, and pounds. On one side, the Jet Propulsion Laboratory, which steered navigation. On the other, Lockheed Martin, the probe’s builder, responsible for supplying the thrust data for the engines. Lockheed Martin’s ground software, a module called SM_FORCES, produced angular momentum data in pound-force seconds, while JPL’s navigation system expected newton-seconds.
Yet the mission’s technical specification left no room for doubt. The interface software’s requirements demanded metric units, but Lockheed Martin’s software violated this specification throughout the nine-and-a-half-month cruise, with no process in the program detecting it.
A factor of 4.45 separated the two units. The JPL team had designed its navigation software to compute thruster firings in newton-seconds, the metric unit, while Lockheed Martin had written its software to output force data in pound-force seconds, the imperial unit, knowing that one pound-force second is equivalent to about 4.45 newton-seconds. Each trajectory correction, seemingly tiny, was therefore underestimated by the same proportion.
An Alert Ignored for Months
The most troubling aspect isn’t that the error slipped by at first. It is that it was detected, then dismissed.
The navigation team had noticed that the trajectory data did not match the forecasts and had flagged the problem, but concerns were pushed to the back burner, in part because documentation procedures had not been followed. The alarm did go off. No one had the means to hear it clearly, due to a lack of clear traceability between the two teams. Each corrective maneuver, taken in isolation, seemed harmless.
The desaturation maneuvers of the probe’s angular momentum occurred 10 to 14 times more often than the navigation team had anticipated. This unexpected frequency multiplied the impact of the unit error, yielding an ever-widening trajectory deviation with every correction. The damage had been done long before the approach to Mars.
A final check took place, too late to alter the outcome. Just before orbital insertion, the JPL navigation team reworked its calculations and detected a problem; a corrective maneuver was contemplated, but the team ultimately decided not to execute it, deeming themselves, according to the investigation report, unprepared for such an abnormal scenario. The probe continued its approach without correction.
What NASA Changed Afterwards
The agency did not delay in reacting. A commission of inquiry confirmed that the main cause of the loss of the craft was the absence of converting Anglo-Saxon units into metric units in a segment of the navigation software, a program used by ground engineers and not aboard the probe. The report did not stop at this technical finding alone.
The commission issued an initial report proposing various measures to prevent this type of incident, applicable with immediate effect. Practically, this meant strengthened cross-checks between teams working on different software, an increased requirement for documentary consistency on the units used at each stage of a project, and heightened vigilance for weak signals raised by field teams. The lesson extends beyond the simple feet-versus-meters issue. Following this episode, NASA revised its procedures; it lost Mars Climate Orbiter but gained a higher level of design rigor for its subsequent missions.
This episode is today taught in engineering curricula and cited in academic case studies focused on failures of complex systems. The reason lies less in the unit conversion itself than in what it reveals: two impeccable teams, each producing accurate numbers in their own reference frame, can bring down an entire project simply because no one checked that those references aligned at the interface between the two software systems. The world’s most advanced space technology can fail over a detail that a middle-school student can master in a physics class.
Sources: spacedaily.com | loctopusjournal.fr
