Failure intelligence, not failure trivia
F Failurepedia

Aerospace and Space Exploration

Mars Climate Orbiter

NASA's Mars Climate Orbiter launched in December 1998 to study the Martian atmosphere and relay data for a companion lander. On September 23, 1999, it approached Mars far lower than planned and was destroyed, because Lockheed Martin's ground navigation software output thruster force in pound-force seconds while JPL's flight software expected newton-seconds. The unconverted unit mismatch pushed the spacecraft's estimated position dozens of kilometers off, and no one caught it before arrival.

Failed launch Failed initiative High
Company
NASA
Started
1998-12-11
Ended
1999-09-23
Total mission cost (spacecraft, launch, and operations)
$327.6M
Collapse speed
Sudden
Preventability
High
Lesson transfer
Universal
Last reviewed
2026-08-06

Narrative

The story

The ambition

Mars Climate Orbiter was one half of NASA's Mars Surveyor '98 program, launched December 11, 1998, from Cape Canaveral on a Delta II rocket. Its job was to orbit Mars and study the planet's atmosphere and climate in more detail than any prior mission, while also serving as a communications relay for its sister spacecraft, Mars Polar Lander, which was due to touch down a few months later. It was a comparatively low-cost mission, part of NASA's "faster, better, cheaper" push to fly more planetary missions on tighter budgets after the billion-dollar Mars Observer had been lost in 1993.

The rise

The spacecraft's nine-month cruise to Mars went smoothly. Navigation teams at NASA's Jet Propulsion Laboratory in Pasadena and at Lockheed Martin Astronautics in Denver, which had built the spacecraft, tracked its trajectory using periodic thruster firings to keep it on course, exchanging trajectory and maneuver data throughout the flight. As the orbiter closed in on Mars in September 1999, controllers prepared for the critical orbit-insertion burn that would place it around the planet.

The cracks

The trajectory problem was building unnoticed for months. Lockheed Martin's ground software calculated the small trajectory corrections produced by the spacecraft's thruster firings and delivered that data in pound-force seconds, an English unit of impulse. JPL's navigation software, which used the data to model the spacecraft's course, was written to expect the data in newton-seconds, the metric unit, a difference of a factor of about 4.45. Nobody converted the numbers, and nobody caught the mismatch through the mission's design or test process. The error accumulated gradually, each uncorrected thruster firing nudging the navigation team's model of the spacecraft's true position further from reality.

The collapse

On September 23, 1999, Mars Climate Orbiter fired its main engine to enter Mars orbit. Because of the accumulated navigation error, the spacecraft approached Mars far lower than planned, at roughly 57 kilometers altitude rather than the intended 226 kilometers, well under the approximately 80-to-85-kilometer altitude NASA judged survivable. Flight controllers lost contact as the spacecraft passed behind Mars and never regained it. JPL formed a mishap investigation board, and within about a week the team had identified the pound-force-versus- newton unit mismatch as the direct cause. NASA concluded the orbiter had either burned up in the Martian atmosphere or, in an alternate scenario the board considered, skipped back out into an orbit around the sun.

The aftermath

NASA's leadership was explicit that the failure was not simply an arithmetic mistake. NASA Administrator Edward Weiler said the problem was not the unit error itself but "the failure of NASA's systems engineering, and the checks and balances in our processes, to detect the error." The investigation also found that at least two navigators had flagged unexplained trajectory discrepancies during the cruise but that the concerns did not trigger a formal review in time. NASA did not single out Lockheed Martin for blame, framing the loss instead as a process failure spanning both organizations. The agency's next Mars lander, Mars Polar Lander, which Climate Orbiter would have supported as a relay, was lost on arrival about ten weeks later in an unrelated failure, compounding a difficult year for the Mars Surveyor program. NASA subsequently tightened its verification procedures for unit consistency and systems-engineering review across projects.

The lessons

Mars Climate Orbiter is a textbook case of an interface failure: two competent teams each built software that worked correctly on its own terms, and the mission was lost in the gap between them. The specific bug, English units feeding a metric-only system, was trivial to describe after the fact and would have been caught by a single explicit unit check at the interface. The deeper lesson NASA drew was about process, not arithmetic. Early warning signs existed in the trajectory discrepancies navigators noticed during cruise, but the reporting process for flagging those concerns was not followed and no review caught the error before arrival. A system with real safety margin needs an independent cross-check at every interface where data changes hands between organizations, not just correct code within each side.

Causal timeline

Failure Anatomy

  1. 1998-12-11

    Launch on Mars Surveyor '98

    Mars Climate Orbiter launched December 11, 1998, from Cape Canaveral to study the Martian atmosphere and relay data for the companion Mars Polar Lander mission. [1]

  2. 1998-12-1999-09

    Ground and flight software disagree on units

    Throughout the cruise, Lockheed Martin's ground software delivered thruster-impulse data in pound-force seconds while JPL's navigation software treated the numbers as newton-seconds, silently corrupting the trajectory model. [2]

    Technical failurePoor execution
  3. 1999-09

    Navigators flag discrepancies, without resolution

    At least two navigators noticed the spacecraft's tracked position was off from expectations during the cruise, but the concern was not escalated through a formal review before arrival. [3]

    Information failure
  4. 1999-09-23

    Orbit-insertion burn at the wrong altitude

    On September 23, 1999, the spacecraft fired its engine to enter Mars orbit based on the corrupted trajectory data, putting it on a path roughly 57 kilometers above the surface instead of the planned 226 kilometers, below the survivable minimum. [4]

    Technical failure
  5. 1999-09-23

    Spacecraft lost, cause identified within days

    Controllers lost the signal as the orbiter passed behind Mars and never regained it. JPL's mishap investigation board identified the pound-force-versus-newton mismatch as the direct cause within about a week. [5] [6]

    Technical failure

Structured analysis

What Went Wrong

Root causes

Unit mismatch between ground and flight software. Lockheed Martin's ground navigation software calculated thruster impulse in pound-force seconds while JPL's flight navigation software expected newton-seconds, and the values were never converted, corrupting the spacecraft's estimated trajectory. [2]

No systems-engineering check caught the mismatch. NASA's own post-mortem concluded the root failure was not the unit error itself but the absence of systems-engineering checks and balances that should have caught it before arrival. [7]

Contributing factors

Early trajectory discrepancies were not escalated. At least two navigators noticed the spacecraft's tracked position diverging from expectations during the cruise to Mars, but the discrepancy was not escalated through a formal review in time to catch the underlying cause. [3]

Interface requirements were not enforced. The mission's design and test process did not include an explicit, enforced check that data crossing the Lockheed Martin-to-JPL interface used consistent units. [2]

Immediate trigger

Orbit-insertion burn using corrupted trajectory data. On September 23, 1999, the spacecraft executed its Mars orbit-insertion burn based on a trajectory model that was wrong because of the unconverted unit mismatch, sending it in far too low. [4]

Visible symptoms

Signal loss on Mars arrival. Flight controllers never reacquired the spacecraft's signal after it passed behind Mars for orbit insertion, the first direct sign something had gone wrong. [5]

Warning signs

Unexplained trajectory discrepancies during cruise. Navigators tracking the spacecraft's course during its nine-month cruise noticed its actual position diverging from the modeled trajectory, a symptom later traced to the unit error, but it was not resolved before arrival. [3]

Affected groups

Taxpayers

Keep reading

  • Mars Polar Lander

    The Mars Climate Orbiter it would have supported as a relay was lost about ten weeks later in an unrelated failure.

Evidence

Claims & sources

Every numbered marker in the analysis links to the claim it rests on, and each claim to its sources.

  1. [1]

    Mars Climate Orbiter launched December 11, 1998, from Cape Canaveral aboard a Delta II rocket, as part of the Mars Surveyor '98 program, to study the Martian climate and atmosphere and relay data for the companion Mars Polar Lander mission.

    High Fact Mars Climate Orbiter
  2. [2]

    The spacecraft was lost because Lockheed Martin's ground navigation software calculated thruster impulse in pound-force seconds, an English unit, while JPL's flight navigation software expected the values in newton-seconds, the metric unit, and the mismatch was never corrected.

    High Fact Mars Climate Orbiter Team Finds Likely Cause of Loss Mars Climate Orbiter How NASA Lost a Spacecraft Due to a Metric Math Mistake
  3. [3]

    At least two navigators noticed the spacecraft's tracked position diverging from expectations during its cruise to Mars, but the discrepancy was not resolved through a formal review before arrival.

    Moderate Reported explanation How NASA Lost a Spacecraft Due to a Metric Math Mistake
  4. [4]

    On September 23, 1999, the spacecraft's orbit-insertion burn, based on corrupted trajectory data, put it on a path roughly 57 to 60 kilometers above Mars instead of the planned 150 to 226 kilometers, below the approximately 80-to-85-kilometer altitude NASA judged survivable.

    High Fact NASA's Mars Climate Orbiter Believed to Be Lost Mars Climate Orbiter
  5. [5]

    Flight controllers lost contact with Mars Climate Orbiter as it passed behind Mars for orbit insertion on September 23, 1999, and never reacquired its signal.

    High Fact NASA's Mars Climate Orbiter Believed to Be Lost
  6. [6]

    JPL's mishap investigation board identified the pound-force-versus-newton unit mismatch as the likely direct cause of the loss within about a week of the failure.

    High Fact Mars Climate Orbiter Team Finds Likely Cause of Loss
  7. [7]

    NASA Administrator Edward Weiler stated that the underlying problem was not the unit error itself but the failure of NASA's systems engineering and its checks and balances to detect the error.

    High Fact Mars Climate Orbiter Team Finds Likely Cause of Loss

Sources