Skip to the timeline
Let's go to Mars

Two spacecraft lost at Mars in ten weeks

Climate Orbiter flew to 57 km altitude instead of 226 because two teams used different units. Ten weeks later Polar Lander probably shut its engines off 40 metres up.

Two technicians in white caps and face masks stand at a spacecraft wrapped in gold and white thermal foil, with exposed instrument decks, cabling and a white flexible duct. The nearer one, in pale blue coveralls, leans over it from behind; the other looks on from the lower left. A large dish antenna sheathed in silvery covering fills the right of the frame, and the chamber walls behind are dark red.
Mars Climate Orbiter during acoustic tests that simulated launch conditions, 27 May 1998. The spacecraft flew as built; the units error was in the ground software on Earth that modelled its thruster firings. · Image: NASA

Mars Climate Orbiter lit its main engine at 09:00:46 UTC on 23 September 1999 and went behind the planet 49 seconds earlier than predicted. It never came out the other side.

The cause was identified six days later. Lockheed Martin's ground software wrote thruster impulse values into the angular momentum desaturation file in pound-force-seconds; the JPL navigation software's interface specification required newton-seconds. Nobody converted, and from April 1999, when the files first entered orbit determination, the firings they described were modelled as 4.45 times weaker than they were. Navigators had noticed that their Doppler-only solutions kept putting the spacecraft closer to Mars than the combined solutions did, and said so, but they raised it by e-mail rather than filing an Incident, Surprise, Anomaly report. The final trajectory correction aimed at a first periapsis of 226 kilometres. An hour before the burn, better tracking put it at 110; the board's reconstruction afterwards put it at 57. The minimum the orbiter could survive was 80.

Mars Polar Lander entered the atmosphere on 3 December 1999, bound for 76 degrees south, and was never heard from. To save money the mission carried no telemetry during entry, descent and landing. The JPL review board under John Casani concluded, from ground tests, that the most probable cause lay in the Hall-effect touchdown sensors in the three landing legs, which emit a spurious pulse when the legs snap into place at about 1,500 metres. The flight software, already sampling the sensors, latched that pulse; when engine shutdown was enabled at 40 metres it saw "landed" and commanded the descent engines off. The lander was coming down at about 13 metres per second; with the thrust gone it would have hit the ground at about 22, against a design touchdown speed of 2.4. The board called this probable, not proven.

Two Deep Space 2 penetrators, Scott and Amundsen, rode down with the lander and were never heard from either. No cause has been established.

The Mars Surveyor 2001 lander was cancelled and put into storage. Thomas Young's independent assessment in March 2000 found the Mars '98 project underfunded by at least 30 percent against what it had been asked to deliver.