The Loss of Mars Climate Orbiter

The MCO spacecraft was launched on Dec 11, 1998, and headed to Mars for a multi-year mission to support its partner, the Mars Polar Lander, and to study the atmosphere and clouds of Mars. It spent months flying through space, finally reaching Mars in September, 1999. On Thursday, September 23, 1999, the spacecraft fired its rockets to alter its trajectory so that it would enter orbit around Mars. It passed behind the planet as it was firing ... and was never heard from again! cue spooky music

What happened?

After several weeks of investigation, it was determined that the fundamental cause of the mishap was a mistake in a computer program used to control the spacecraft's thruster rockets. Engineers at Lockheed Martin wrote a computer program called "SM_FORCES", which calculated the thrust produced by the rockets for a given sort of firing. The output was supposed to be in metric units called "Newton-seconds." But the actual code produced thrust in English units called "pound-seconds".

       1 Newton-second  =  4.45 pound-seconds

The rockets were used a number of times during the long trip between the planets; each time, they caused the spacecraft to deviate a bit more from its proper path. The navigation team, at the Jet Propulsion Laboratory, didn't notice the error. By the time the MCO reached Mars, its actual position was significantly different from its expected position: the navigation team thought that it was about 110 km above the surface of Mars at closest approach ... whereas it was really only about 57 km above the surface. The aerodynamic forces of this improperly low altitude destroyed or disabled the spacecraft.

There were a number of reasons that no one caught the error.

the "SM_FORCES" software, and the entire rocket system, were designed and built by one group of engineers at one institution (Lockheed Martin) ...
but the software was used by a team of navigators at a different institution (JPL)
the navigation team was understaffed (and busy with three spacecraft at the time)
although the navigation team did notice that there was some sort of problem with the rockets and their effect on the spacecraft's trajectory as early as April, 1999, it was considered a low-level problem, and never brought to the attention of the entire project management, or resolved

For more information, see http://mars.jpl.nasa.gov/msp98/news/mco991110.html