This narrative is based primarily on three sources: AIAA Paper No. 70-1260, by Glynn S. Lunney (one of the three flight directors on Apollo 13), "Discussion of Several Problem Areas During the Apollo 13 Operation," presented at the AIAA 7th Annual Meeting and Technical Display, Houston, Texas, Oct. 19-22, 1970; NASA, "Report of Apollo 13 Review Board," June 15, 1970; and "Apollo 13 Technical Air to Ground Voice Transcription," April 1970. See also the published House and Senate hearings cited in notes 75 and 79, Chapter 11.
Apollo 13, the third manned lunar landing and exploration mission, had been tentatively scheduled in July 1969 for launch in March 1970, but by the end of the year the launch date had been shifted to April. In August 1969 crew assignments for Apollo 13 were announced: James A. Lovell commanded the prime crew, which included Thomas K. Mattingly II as command module pilot and Fred W. Haise as lunar module pilot. Their backups were John Young, John Swigert, and Charles Duke. The target for the mission was the Fra Mauro Formation, a site of major interest to scientists, specifically a spot just north of the crater Fra Mauro, some 550 kilometers (340 miles) west-southwest of the center of the moon's near side.
On March 24, 1970, during the countdown demonstration test for Apollo 13, KSC test engineers encountered a problem with an oxygen tank in the service module. The spacecraft carried two such tanks, each holding 320 pounds (145 kilograms) of supercritical oxygen. They provided the oxygen for the command module atmosphere and (along with two tanks of hydrogen) three fuel cells, which were the spacecraft's primary source of electrical power. Besides power, the chemical reaction in the cells produced water, which not only supplied the crew's drinking water but was circulated through cooling plates to remove heat from certain critical electronic components. The tanks were designed to operate at pressures of 865 to 935 pounds per square inch (psi) (6,000 to 6,450 kilopascals) and temperatures between -340 deg. F and +80 deg. F (-207 deg. C to +27 deg. C). Inside each spherical tank were a quantity gauge, a thermostatically controlled heating element, and two stirring fans driven by electric motors. The fans were occasionally operated to homogenize the fluid in the tank; it tended to stratify, leading to erroneous quantity readings. All wiring inside the tank was insulated with Teflon, a fluorocarbon plastic that is ordinarily noncombustible. Each tank was fitted with a relief valve designed to open when the pressure rose above 1,000 psi (6,900 kilopascals); the tanks themselves would rupture at pressures above 2,200 psi (15,169 kilopascals) . Both tanks were mounted on a shelf in the service module between the fuel cells and the hydrogen tanks.
The countdown demonstration test called for the tanks to be filled, tested, and then partially emptied by applying pressure to the vent line, thus forcing oxygen out through the fill line. Number one tank behaved normally in this test, but number two released only 8 percent of its contents, not 50 percent as required. Test engineers decided to proceed with the rest of the test and investigate the problem later. The next day, after KSC engineers had discussed the problem with colleagues at MSC, North American Rockwell (builders of the service module), and Beech Aircraft (manufacturers of the oxygen tanks), they tried emptying the tank again, with no success. Further talks led to the conclusion that the tank probably contained a loose-fitting fill tube, which could allow pressure to escape without emptying the tank.
When normal procedures again failed to empty the tank, engineers decided to use its internal heaters to boil off the contents and applied direct-current power at 65 volts to the heaters. This was successful but slow, requiring eight hours of heating. It was then decided that if the tank could then be filled normally it would not cause a problem in flight. A third test gave the same result as the second, requiring heating to empty the tank.
In view of the difficulty of replacing the oxygen shelf - a job that would take at least 45 hours - and the possibility that other components might be damaged in the process and the launch delayed for a month, NASA and contractor officials decided not to replace the tanks.
The spacecraft was launched on April 11, 1970, and the mission was quite routine for the first two days. At 30 hours and 40 minutes after launch (30:40 ground elapsed time, or g.e.t.), the crew ignited their main engine to put the spacecraft on a hybrid trajectory, a flight path that saved fuel in reaching the desired lunar landing point.* At 46:40** the crew routinely switched on the fans in the oxygen tanks briefly. A few seconds later the quantity indicator for tank number two went off the high end of the scale, where it stayed. The tanks were stirred twice more during the next few hours; and at 55:53, after a master alarm had indicated low pressure in a hydrogen tank, the Mission Control Center (MCC) directed the crew to switch on all tank stirrers and heaters. Shortly thereafter the crew heard a loud "bang" and felt unusual vibrations in the spacecraft. Mission controllers noticed that all telemetry readings from the spacecraft dropped out for 1.8 seconds. In the CM, the caution and warning system alerted the crew to low voltage on d.c. main bus B, one of two power distribution systems in the spacecraft. At this point command module pilot Jack Swigert told Houston, "Hey, we've had a problem here."
Because of the interruption of telemetry that had just occurred, flight controllers in the MCC had difficulty for the next few minutes determining whether they were getting true readings from the spacecraft sensors or whether the sensors had somehow lost power. Before long, however, both MCC and the crew realized that oxygen tank number two had lost all of its contents, oxygen tank number one was slowly losing its contents, and the CM would soon be out of oxygen and without electrical power. Among the first actions taken were shutting down one fuel cell and switching off nonessential systems in the CM to minimize power consumption; shortly after, the second fuel cell was shut down as well. When the remaining oxygen ran out, the CM would be dead; its only other power source was three reentry batteries providing 120 ampere-hours, and these had to be reserved for the critical reentry period.
An hour and a half after the "bang," MCC notified the crew that "we're starting to think about the lifeboat" - using the lunar module (LM) and its limited supplies to sustain the crew for the rest of the mission. Plans for such a contingency had been studied for several years, although none had anticipated a situation as grave as that of Apollo 13. Many of these studies were retrieved and their results were adapted to the situation as it developed.
Shortly after the accident, mission commander James Lovell reported seeing a swarm of particles surrounding the spacecraft, which meant trouble. Particles could easily be confused with stars, and the sole means of determining the spacecraft's attitude was by locating certain key stars in the onboard sextant. Navigational sightings from the LM were difficult in any case as long as it was attached to the command module, and this would only complicate matters. Flight controllers decided to align the lunar module's guidance system with that in the command module while the CM still had power. That done, the last fuel cell and all systems in the command module were shut down, and the crew moved into the lunar module. Their survival depended on this craft's oxygen and water supplies, guidance system, and descent propulsion engine (DPS) . Normally all course corrections were made using the service propulsion system (SPS) on the service module, but flight controllers ruled out using it, partly because it required more electrical power than was available and partly because no one knew whether the service module had been structurally weakened by the explosion. If it had, an SPS burn might be dangerous. The DPS would have to serve in its place.
When word got out that Apollo 13 was in trouble, off-duty flight controllers and spacecraft systems experts began to gather at MSC, to be available if needed. Others stood by at NASA centers and contractor plants around the country, in touch with Houston by telephone. Flight directors Eugene Kranz, Glynn Lunney, and Gerald Griffin soon had a large pool of talent to help them solve problems as they arose, provide information that might not be at their fingertips, and work on solutions to problems they could anticipate farther along in the mission. Astronauts manned the CM and LM training simulators at Houston and at Kennedy Space Center, testing new procedures as they were devised and modifying them as necessary. MSC director Robert R. Gilruth, Dale D. Myers, director of manned space flight, and NASA administrator Thomas O. Paine were all on hand at Mission Control to provide high-level authority for changes.
Soon after the explosion, the assessment of life-support systems determined that although oxygen supplies were adequate, the system for removing carbon dioxide (CO2) in the lunar module was not. The system used canisters filled with lithium hydroxide to absorb CO2 as did the system in the command module. Unfortunately the canisters were not interchangeable between the two systems, so the astronauts were faced with plenty of capacity for removing CO2 but no way of using it. A team in Houston immediately set about improvising a way to use the CM canisters, using materials available in the spacecraft. Flight controllers, meanwhile, were addressing operational problems. Their first critical decision was to put the crippled spacecraft back on a free-return trajectory, which was accomplished by firing the LM descent engine at 61:30. Mission Control then had some 18 hours to consider the remaining problems; the next was a possible course adjustment to change the spacecraft's landing point on earth. If this was to be done, it was scheduled for "PC + 2"-two hours after pericynthion (closest approach to the moon), after the spacecraft emerged from behind the moon. In the interval, Houston worked out a new flight plan that would minimize the consumption of oxygen, water, and electricity while keeping vital systems operating.
The alternatives for the PC + 2 maneuver were worked out by about 64 hours g.e.t. A major consideration was the total time to splashdown. Left on its free-return course the command module would return at about 155 hours g.e.t. to a landing in the Indian Ocean. Three options would bring it back in the mid-pacific and could reduce the total mission time to as little as 118 hours. The fifth possibility returned the spacecraft in 133 hours, but to the South Atlantic. For one reason or another, all but one of these choices were discarded. The free-return (no course correction) choice was abandoned, since there was no known reason not to use the LM descent propulsion system. Recovery in either the Atlantic or the Indian Ocean was far from ideal; the main recovery force was deployed in the mid-Pacific and there was not enough time to move it or to make adequate arrangements elsewhere. Two options giving the shortest return time (118 hours) had other drawbacks. Both would require using virtually all of the available propellant, and it was not prudent to assume that no additional course corrections would be required. One of them involved jettisoning the service module, which would expose the CM heat shield to the cold of space for 40 hours and raise questions about its integrity on reentry. After five and a half hours of weighing the choices and their consequences, flight directors met with NASA and contractor officials and presented their findings and recommendations. The decision, made some ten hours before the scheduled engine burn, was to go for mid-Pacific recovery at 143 hours.
During all of these deliberations the atmosphere in the lunar module was gradually accumulating carbon dioxide as the absorbers in the environmental control system became saturated. Members of MSC's Crew Systems Division devised a makeshift air purifier by taping a plastic bag around one end of a CM lithium hydroxide cartridge and attaching a hose from the portable life-support system, allowing air from the cabin to be circulated through it. After verifying that this jury rig would function, they prepared detailed instructions for building it from materials available in the spacecraft and read them up to the crew. For the rest of the mission the improvised system kept the CO2 content of the atmosphere well below hazardous levels.
The decision to recover in the Pacific fixed the time line for the remainder of the mission and imposed some rigid constraints on preparations for reentry. The final course correction had to be made with the LM engine; command module systems had to be' turned on and the guidance system aligned; the service module had to be' discarded; and when all preparations had been made, the lunar module would be cut loose. In all these preparations the power available from the CM's reentry batteries was a limiting factor. From the PC + 2 burn until about 35 hours before reentry the sequence of activation of CM systems was worked out, checked in the simulators, and modified. Fifteen hours before beginning reentry the revised sequence of activities was read to the crew, to give them time to review and practice it.
The husbanding of expendable resources, particularly electrical power, paid off on the morning of landing, when it was discovered that power reserves in the LM were adequate to allow use of it in the CM. Some of the early CM activities could then be done at a less hurried pace. The Apollo 13 command module splashed down within a mile of the recovery carrier with about 20 percent of its battery power remaining. Three weary, chilled astronauts came aboard the U.S.S. Iwo Jima on April 17 and were flown to Hawaii for an emotional reunion with their families.
Mission Control teams and their hundreds of helpers were no less drained. The usual cigars were lighted up after recovery, but the splashdown parties that evening were subdued: most of those who went quit early and went home to bed. Their efforts were recognized the next day when President Richard M. Nixon, on his way to Hawaii, stopped in Houston to present the Presidential Medal of Freedom, the nation's highest civilian award, to the entire team.
NASA immediately convened an investigation board*** to determine the cause of the accident and postponed Apollo 14 until its results were in. Lacking the spacecraft itself - the service module had been jettisoned before reentry, and the crew had been able to take only a few rather poor photographs of it - the board initially had only the data from inflight telemetry to work with. When it became clear that the fault lay in oxygen tank number two, the board carefully reviewed its entire history, from fabrication to launch, as recorded in the detailed documentation that followed every piece of equipment from plant to launch pad. Under the board's direction, MSC and other NASA centers conducted tests under simulated mission conditions to verify its findings. The investigation, which concluded in a few weeks, turned up a highly improbable sequence of human error and oversight that led inexorably to the failure in flight.
Board Chairman Edgar M. Cortright, director of Langley Research Center, explained the board's findings to congressional committees in June. The accident, he reported, was not a random malfunction but resulted from an unusual combination of mistakes as well as "a somewhat deficient and unforgiving design." As the board's report reconstructed the events leading up to the accident, the tank left Beech Aircraft's plant on May 3, 1967, after passing all acceptance tests. It was installed as part of a shelf assembly in service module no. 106 on June 4, 1963, having passed all tests conducted at North American Rockwell during assembly. Design changes in the service module, however, necessitated removing the entire shelf from SM 106 for modification. During removal, which was accomplished by use of a special fixture that fit under the shelf to lift it upward, workmen overlooked one bolt that held down the back of the shelf, with the result that the removal fixture broke, dropping the shelf two inches. The board concluded that this incident might have jarred loose a poorly fitting fill tube. Subsequent tests did not detect any flaws, and after modification the shelf was shipped to Kennedy Space Center for installation in SM 109, the Apollo 13 spacecraft.
Arrangement of fuel cells and cryogenic systems in bay 4.
What was not known was that this oxygen tank was fitted with obsolete thermostatic switches protecting its heating elements. Original specifications for the switches called for operation on 28 volts d.c.; in 1965 this was changed to 65 volts d.c. to match the test and checkout equipment at the Cape. Later tanks conformed to the new specifications, but this one, which should have been modified, was not, and the discrepancy was overlooked at all stages thereafter.
Oxygen tank no.2 internal components.
In a normal checkout of a normal tank, this would not have mattered, because the switches would not have opened during normal operation. But the improvised procedure used when this tank failed to empty (the result of a loose fitting, as noted above) raised the temperature in the tank above 80 deg.F (27 deg.C), at which point the switches opened. Tests conducted during the investigation showed that the higher current produced by the 65-volt power source caused an arc between the contact points as they separated, welding them together and preventing their opening when the temperature dropped. This went undetected during the detanking procedure at the Cape; it could have been noticed if anyone had monitored the heater current, which would have shown that the heaters were operating when they should not have been. But all attention was on the specific malfunction, and no one was aware that the heaters were on continuously for eight hours on two separate occasions. The result, as tests showed, was that the heater tube reached 1,000 deg,F (538 deg.C) in spots, damaging the Teflon insulation on the adjacent fan-motor wiring and exposing bare wire. From that point on, the board concluded, the tank was hazardous when filled with oxygen and electrically powered. Teflon can be ignited at a high enough temperature in the presence of pure oxygen, and the tank contained small amounts of other combustibles as well.
Unfortunately for Apollo 13, the tank functioned normally for the first 56 hours of the mission, when the heaters and the fans were energized during routine operations. At that point an arc from a short circuit probably ignited the Teflon, and the rapid pressure rise that followed either ruptured the tank or damaged the conduit carrying wiring into the tank, expelling high-pressure oxygen. The board could not determine exactly how the tank failed or whether additional combustion occurred outside the tank, but the pressure increase blew off the panel covering that sector of the service module and damaged the directional antenna, causing the interruption of telemetry observed in Houston. It also evidently damaged the oxygen distribution system, or the other oxygen tank, as well, leading to the loss of all oxygen supplies and aborting the mission.
The board pointed out that although the circumstances of the tank failure were highly unusual and that the system had worked flawlessly on six successful missions, Apollo 13 was a failure whose causes had to be eliminated as completely as possible. It recommended that the oxygen tanks be modified to remove all combustible material from contact with oxygen and that all test procedures be thoroughly reviewed for adequacy.
Compared to the AS-204 fire in 1967, Apollo 13 was only a frightening near-miss, and because its cause was localized and comparatively easy to discover, it had fewer adverse effects on the program. Only the skill and dedication of hundreds of members of the often-celebrated "manned space flight team" saved it, however, and the accident served to remind NASA and the public that manned flight in space, no matter how commonplace it seemed to the casual observer, was not yet a routine operation. The same lesson had to be learned once more sixteen years later, when on January 28, 1986, the space shuttle Challenger and all seven of its crew were lost a minute after launch. An unforgiving design and the failure of human judgment under pressure combined again to bring a program to a halt while corrective measures were taken.
* The "hybrid" trajectory was designed so that if the main propulsion engine failed, the attitude control rockets on the spacecraft could change the flight path enough to bring it back to a safe reentry after it rounded the moon.
** This and all subsequent times are in hours and minutes g.e.t. Launch time was 12: 13 p.m. Eastern Standard Time on April 11.
*** Board members were: Edgar M. Cortright, director, Langley Research Center, chairman; Robert F. Allnut, assistant to the administrator, NASA Hqs.; Neil Armstrong, MSC; John F. Clark, director, Goddard Space Flight Center; Brig. Gen. Walter R. Hedrick, Jr., Hqs. USAF; Vincent L. Johnson, deputy associate administrator for engineering, Office of Space Science and Applications; Milton Klein, manager, AEC- NASA space nuclear propulsion office; and Hans M. Mark, director, Ames Research Center.