Saving Skylab

The Saturn V performed its final mission in style, and 10 minutes after liftoff on 14 May 1973 Skylab was in its planned orbit, 436 km above the earth. During the next half hour a series of commands from the instrument unit would bring the laboratory to life. First a radiator cover was jettisoned so that the refrigerators could be switched on. Next the four sections of the payload shroud peeled away; Skylab officials, recalling the failure of a similar cover on Gemini 9, breathed a sigh of relief. With deployment of the telescope mount from the forward end of the stack to its flight position astride the docking adapter 16 minutes into the flight, Skylab passed a crucial hurdle. The move cleared the path for the Apollo spacecraft to reach its docking port. Within minutes the telescope mount’s four solar wings, resembling the sails of a Dutch windmill, opened. Meantime, the spaceship had assumed a solar inertial attitude, its long axis in the plane of the orbit and the telescope mount pointing toward the sun. Thus far there had been only one curious indication, a report from Houston that the meteoroid shield had deployed prematurely. When nothing more was heard, officials at the launch site dismissed the indication as a false telemetry signal. After the telescope mount had moved into its proper position, there was time to relax while awaiting deployment of the workshop’s solar arrays.¹

THE ACCIDENT

The relaxation was short lived. About half an hour after liftoff, Flight Director Donald Puddy in Houston reported erratic signals from both the meteoroid shield and the workshop solar arrays. The solar wings were scheduled to deploy 41 minutes after launch, when Skylab had passed beyond the range of the station at Madrid. Tension mounted as officials listened for news from the tracking station at Carnarvon, Australia. The information was confusing. One telemetry signal indicated that the array had released for deployment but was not fully extended, while temperature signals suggested that both wings were gone, a conclusion reinforced by the absence of voltage signals. The failure of backup commands from both Goldstone, California, and Madrid seemed to confirm the worst fears. The solar panels were the main topic of discussion at the postlaunch briefing at Kennedy.²

By late afternoon, it appeared that Skylab had at least two major problems. If the workshop solar panels were indeed gone, Skylab had lost nearly half of its electrical power. The workshop and ATM array each provided about 5 kw of usable power. Apollo fuel cells could produce an additional 1.2 kw for 20 days; after that the command-service module would draw electricity from Skylab. The system had been designed with power to spare; even without the workshop panels, officials believed an adequate mission was possible until Apollo’s fuel cells ran dry. Then the crew would be forced to curtail most experiments for the last week. The second and third crews would be hampered for much longer periods.³

The power shortage drew most attention at an evening press conference; little was said about an even more serious problem, the apparent loss of the micrometeoroid shield. No one was particularly worried about damage from a meteoroid strike, since the chances of a hit were slim.^* But the shield’s secondary function, thermal control, loomed large in the aftermath of the launch. The shield had been designed to keep the workshop on the cool side of the comfort zone, heating being easier than cooling. The outside of the shield was a black-and-white pattern designed to absorb the desired amount of heat. The inside of the shield and the outside of the workshop were covered with gold foil, which regulated the flow of heat between the two. It was an admirable system as long as the shield stayed in place. Without it, the gold coating on the workshop would rapidly absorb excessive heat, making the interior uninhabitable.⁴

The shield had failed to deploy at the scheduled time and subsequent ground commands had no effect. While officials were debating further action, Saturn engineers discovered flight data indicating an anomalous lateral acceleration about a minute after liftoff. The data, coming just before the space vehicle reached its maximum dynamic pressure, suggested some structural failure. A short time later, workshop temperatures began rising, strong evidence that the shield was gone. Within a few hours, readings on many of the outside sensors exceeded 82°C, the maximum scale reading. Internal temperatures moved above 38°C. Working from the thermal model, Huntsville engineers figured that workshop temperatures would go as high as 77°C internally and 165°C on the outside, endangering food, film, perhaps even the structure itself. Mission Control therefore began maneuvering the exposed area out of direct sunlight, and some cooling occurred.⁵

A bleak picture confronted the Skylab team the evening of launch. Besides the overheating and the lack of power, the attitude-control system had problems. Responses from rate gyroscopes were not averaging properly, and the initial maneuvers had expended excessive amounts of nitrogen gas. No doubt engineers wished they could bring Skylab back for repairs. This was out of the question, of course. The chances of repairing it in space looked unpromising, but the attempt had to be made.⁶

The first decision was to delay the launch of the crew by five days. Huntsville began a series of analytical studies to predict likely temperatures in the workshop and assess their impact. Both Huntsville and Houston started investigating ways of deploying a thermal shield. At the same time, contractors and other NASA centers were encouraged to pursue independent studies.⁷

At Marshall, Center Director Rocco Petrone moved with characteristic vigor, giving carte blanche to a special task force under the direction of the deputy directors of the Astronautics and Astrionics Laboratories: “Whatever you need at the center is yours.” The team operated from the Huntsville Operations Support Center, with personnel largely drawn from the mission support groups. Marshall’s laboratories and contractors’ plants provided additional help. Computer time was soon in short supply. Eventually much of the work was done on Martin computers in Denver, and sometimes procurement had to search elsewhere.⁸

The accident drastically altered activities within Huntsville’s operations center. A normal 40-hour week had been planned for operations personnel, with a skeleton crew on duty the remainder of the time. Facing an emergency of undetermined length, officials quickly established an around-the-clock schedule, reinforcing the operations team with Skylab design engineers. The support groups directly affected by the accident (electrical power systems, attitude control, and environmental control) doubled in size, while overall numbers at the operations center increased from 400 to 600.⁹

At first, Eugene Kranz, chief of JSC’s Flight Control Division, tried to operate with his four flight-control teams, having each team work specific problems when not manning the consoles, so that individuals who worked out plans could then implement them. By the 15th, however, the scheme had become unworkable. Too many things required investigation, and the major problems demanded continuous attention. Two teams were directed to man the consoles around the clock, while the other two supported contingency planning: altering the flight plan and activation checklist, supporting development of a sunshade, and reducing power requirements of the workshop.¹⁰

If Huntsville and Houston bore the heaviest responsibility, the entire Skylab team was involved. From Huntington Beach to Cape Canaveral, workdays of 16-18 hours became normal, and people lost track of time. Tempers remained remarkably calm despite the long hours. Relations between Marshall and JSC were excellent, a condition that both sides attributed to the close working ties that had grown up during Skylab’s design and development phases. There was healthy competition between groups developing sunshades, but in looking back on the time, participants most often recalled the teamwork and the tremendous amount of work accomplished in such a short time. Huntsville officials referred to the period as “the 11 years in May.”¹¹

MANEUVERING FOR MINIMUM HEAT, MAXIMUM POWER

The electrical power situation, while bothersome, was not an immediate threat. But the workshop’s temperature had to be lowered fast. Separately, neither problem seemed insurmountable; together the loss of the solar panels and meteoroid shield posed a dilemma, for anything that reduced the effect of one malfunction increased the effects of the other. To produce electricity, Skylab needed to remain in a solar inertial attitude, with the sun’s rays perpendicular to the solar panels, but this position exposed the full length of the workshop. For a time Mission Control pointed the forward end directly at the sun, which lowered temperatures somewhat but also reduced power generation. Experiments with various attitudes showed the best compromise to be pitched up about 45° toward the sun. During the daylight portion of each orbit enough sunshine struck the solar panels to charge the batteries for the next period of darkness, and internal temperatures stabilized near 42°C.¹²

The search for a compromise attitude was complicated by steering problems. Nine rate gyroscopes served as the basic sensors for attitude control, measuring the rate of rotation around three axes. Several gyroscopes overheated the first day, producing off-scale readings and causing the flight controllers to discontinue the practice of averaging the information from two gyroscopes. Fortunately, at least one gyroscope in each axis worked satisfactorily. The gyros accumulated excessive errors, and because the errors were erratic, ground controllers could not compensate for them. During the first few weeks, the attitude-control team waged a constant battle to predict the movement of the rate gyroscopes. The problem was compounded, however, when Skylab left the solar inertial plane. Random errors sent spurious signals to the control-moment gyroscopes, frequently causing them to reach saturation (p. 172). Desaturation required a daylight pass in the solar inertial attitude. To reduce the amount of maneuvering required, Mission Control worked out some rough-andready substitute procedures: measuring roll attitude by reading temperatures on opposite sides of the workshop, determining pitch angle by the electrical output of the solar wings, and calculating Skylab’s momentum to determine if it was in the correct orbital plane.¹³

All these unscheduled maneuvers used up large amounts of attitude control propellant, and while there were possible solutions to the other malfunctions, the gas could not be replaced. Due to favorable launch conditions, Skylab had lifted off with an excess supply, but in the first three days the compressed nitrogen that powered the attitude-control thrusters was expended at an alarming rate. By 17 May, 23% of it was gone, twice the amount expected. The situation improved as flight controllers became more adept at maneuvering the workshop. Though the expenditure of nitrogen remained too high, the rate could be tolerated until the first crew was launched. On the 17th, that launch was delayed another five days.¹⁴

Ironically, while much of the workshop suffered from overheating, the airlock was too cold, dropping below 4°C on the 18th. The suit umbilical system located in the airlock used water to transfer heat from the astronauts’ suits during extravehicular activity. Despite attempts to warm the airlock with heaters, its temperature continued to drop, approaching freezing on the 21st. If a line in the umbilical system froze, it might crack the heat exchanger at the junction with the airlock’s primary coolant loop. On the 20th, flight controllers had rolled the vehicle a few degrees to expose the airlock to more sunlight. When there was no significant change in temperature, Skylab’s pitch was decreased to 40°. On the following day, the workshop was rolled to place the water loops under direct sunlight for one pass. These maneuvers warmed the airlock and produced more electricity, but sent workshop temperatures up as well. By the end of the 21st readings approached 54°C. Flight controllers juggled Skylab for the rest of the second week, trying to keep temperatures and power within safe limits. The stable condition expected at the end of the first week eluded them, but at least they prevented serious damage to the vehicle.¹⁵

Even with the workshop’s solar array gone, there was enough power to meet Skylab’s needs until the crew arrived-if the ship remained perpendicular to the sun’s rays. When sunlight struck the solar panels at less than a 90° angle, however, production decreased sharply. The estimated power requirement for the unmanned Skylab was 4.5 kw, a few hundred watts below the ATM power system’s maximum output. When it became apparent that maneuvers were essential, engineers turned off heaters and transmitters, reducing requirements to 3 kw. This proved sufficient until the second week, when high-angle maneuvers dropped Skylab’s electrical output below that level. On the 24th, 8 of the ATM’s 18 batteries stopped working because of excessive electrical demands. Returning the workshop to the solar inertial revived only 7 batteries. The loss pointed up the danger of further high-angle maneuvers.¹⁶

ASSESSING THE HEAT EFFECT

The rapid buildup of heat raised doubts about Skylab’s provisions. The day after launch, controllers began plans to restock the larders, assuming that the high temperatures would probably ruin all nonfrozen foods. Over in the food laboratories, however, tests conducted before launch had indicated that the canned food could withstand 54°C temperatures for at least two weeks and the dehydrated items would last even longer. New tests were started to confirm the earlier findings, baking one lot of food at 54° and a second batch at the temperature of the workshop’s food locker. Periodic sampling indicated that the heat was not altering the food’s mineral content or taste. To be on the safe side, the crew was given a quick course in food inspection. On the 22nd Houston officials concluded that the food was all right, and plans to restock the workshop were dropped.¹⁷

The initial prognosis on Skylab’s medical supplies was also pessimistic; it was thought that half of the 62 medications aboard the workshop might be ruined. During the following week, Houston’s medical team pared down the resupply list, relying on heat tests and information from pharmaceutical companies. At the same time, Huntsville officials debated the condition of film aboard Skylab. While the film for the solar telescopes was out of harm’s way in the docking adapter, that for earth-resource cameras and other experiments was stored in workshop vaults. The problem was one of dryness as well as heat-emulsion on the film would dry out in the low humidity. Salt packs placed in the vaults to provide moisture were not expected to last more than 4 days. Kodak engineers believed the crew could restore the film by rehumidifying the vaults, but that might take up to 20 days. Accordingly, plans were made to carry additional film on the Apollo spacecraft.¹⁸

During the early rise in temperature-to perhaps 150°C at some points on the workshop’s exterior-Huntsville engineers feared for Skylab’s structural integrity, but the spacecraft was pressurized without incident.¹⁹ A related problem involved the possible release of toxic gases into the workshop. The aluminum wall of the S-IVB tank was insulated on the inside with polyurethane foam. Well suited for temperatures several hundred degrees below zero, the material at 150°C could give off carbon monoxide, hydrogen cyanide, and to luenediisocyanate. The last item was the most dangerous, lethal in small concentrations. Chemical experts from industry and the academic world considered the hazard a long shot and McDonnell Douglas tests indicated that the concentration of toxic gas in the workshop’s large volume would not be dangerous. Nevertheless, the workshop was vented and repressurized four times. The crew would wear gas masks and sample the air upon first entering.²⁰

DEVISING A SUNSHADE

The Skylab maneuvers were an attempt to buy time until some way was found to shade the workshop. Chances of finding a solution were reasonably good, certainly better than the odds given by many newsmen. For one thing, not all of the exposed surface required protection; covering part of the area facing the sun would bring temperatures within satisfactory limits. Second, a shade would not require rigid tie-downs or strong material since there is no wind in space. But a solution had to be found quickly, before the workshop deteriorated beyond recovery. In the week after the accident, Skylab officials examined scores of ideas, ranging from spray paint and wallpaper to balloons, window curtains, and extensible metal panels. Of the various proposals, 10 seemed promising enough to carry through design and at least partial development.²¹

Huntsville officials began considering a replacement for the meteoroid shield a few hours after launch. Some of the early ideas were rather farfetched, but no suggestion was ignored if its “package was light and the deployment relatively simple.” Several concepts were discarded after the first review. The astronauts ruled out use of the astronaut maneuvering equipment, experimental gear in which the crew had little confidence. The idea of deploying a weather balloon through the scientific airlock was opposed by thermal engineers, who feared it might reflect enough heat to melt solder joints on the ATM solar panels; they preferred a flat shade with some distance between it and the workshop wall. A similar winnowing of ideas occurred in Houston when Max Faget’s engineering directorate met on launch night to brainstorm the problem. After debating a number of suggestions, staff members were assigned specific concepts for further study. Next day paint and wallpaper were eliminated as possible solutions. While spray paint worked surprisingly well in a vacuum chamber test, it posed serious logistical problems and a threat of contamination. Wallpaper was ruled out because of uncertainty about the condition of the workshop’s exterior.²²

From the initial discussions, three promising solutions emerged: extending a shade from a long pole attached to the telescope mount, deploying a shade from the maneuvering Apollo spacecraft, or extending a device through the scientific airlock on the workshop’s solar side. The extravehicular activity required by the first option was a drawback since NASA liked to train extensively for such operations. In its favor, the crew had practiced extravehicular work on the telescope mount; and if they had a portable foot restraint, astronauts could face the exposed area without difficulty. A shade deployed from the spacecraft offered the earliest repair and the least complex design. These advantages were offset by the difficulty of flying around the workshop. The scientific airlock provided the easiest operation. Astronauts could extend the shade from inside the workshop using a procedure already prepared for an experiment. The problem was to design a device that would fit through an opening 20 centimeters square and then expand to cover an area 7 meters square.²³

Faget’s group at JSC concentrated on rigging a shade from the Apollo spacecraft, since this seemed to have the best chance of meeting a 20 May launch date. Standing in the Apollo hatch, an astronaut would attach the shade at the aft end of the workshop. The spacecraft would move laterally to another point on the aft end, where he would secure a second corner of the shade. The CSM pilot would then slowly maneuver the spacecraft toward Skylab’s forward end, allowing the shade to play out. At the telescope mount, the astronaut would make a third attachment. This shade was soon called the SEVA sail, for Standup Extravehicular Activity.²⁴

Responsibility for the SEVA sail fell to Caldwell Johnson, chief of the spacecraft design division. He organized a development team and worked in the centrifuge building; for 10 days the group felt like goldfish in a bowl, as public tours to the centrifuge observed their activity from a mezzanine. Seamstresses stitched the orange material, parachute packers folded the sail for proper deployment, and design engineers attended to the various fasteners. Probably the biggest obstacle was getting exact data on Skylab, since some drawings were not current. In one or two instances, the engineers relied on photographs provided by McDonnell Douglas. Johnson faced an additional problem-warding off suggestions from other NASA officials, whose good intentions might have improved the design at the expense of the deadline. In spite of minor delays, the SEVA sail made rapid progress. At the Management Council meeting on the 16th, it was tentatively chosen as the first shade for deployment.²⁵

Opinion at JSC inclined against sending astronauts outside Skylab; Gemini’s extravehicular troubles were well remembered. At Marshall, on the other hand, EVA from the telescope mount was preferred, largely because of fears that debris might block the scientific airlock. On the evening of the launch, Huntsville engineers began designing a sunshade that looked like a window blind. Working steadily through the night, the group completed the design on the 15th and immediately started fabrication. Testing started the following evening at the neutral-buoyancy simulator. Russell Schweickart, commander of the backup crew, and Joe Kerwin, scientist-pilot of the prime crew, had flown from Houston to test several devices and determine how much an astronaut could see from the telescope mount. They entered the tank amid a circus atmosphere, newsmen peering through floodlights to watch the underwater activity. Before the work ended, Huntsville engineers concluded that they needed another design.²⁶

Schweickart and Kerwin changed from their tank suits and joined 75 Marshall engineers for a debriefing. The astronauts were still in quarantine, and the blue masks worn by the other participants gave the appearance of a surgical ward. Schweickart sketched ideas on a blackboard as the discussion proceeded. Simplicity was essential; launch was less than four days away and crew training, transport, and stowage would eyelets at the far end. After the shade was attached to both ropes, it could be pulled out much as one hoists a flag. The height of the poles above the workshop could be varied if necessary to avoid debris.²⁷

While Huntsville proceeded with its twin-pole sail, a Houston team was developing the parasol that would be the first sunshade. Its designer, Jack Kinzler, had not been among the officials initially contacted for ideas. Although his Technical Services Division enjoyed a reputation for building flight items on short order, it was not a part of Houston’s R&D engineering force. Kinzler had a practical bent, as well as a personal interest in saving the mission for his close friend and neighbor, Pete Conrad. The morning after the launch he began designing possible solutions. Having stowed many items in the Apollo spacecraft, he was familiar with the weight and size constraints. He was predisposed to use the scientific airlock since it would simplify operations for the crew. He soon hit upon a happy combination of coiled springs and telescoping rods to provide the means of deploying a large cover through a small porthole.²⁸

By the 16th, Kinzler’s inspiration was taking shape. He attached a parachute canopy to some telescoping fishing rods that were fitted in hub-mounted springs. Springs, poles, and canopy were then stowed in a container roughly the size of the airlock canister. Kinzler deployed the parasol with strings tied to the telescoping rods. As the fishing poles extended and locked in a horizontal position, the attached parachute formed a smooth canopy. Demonstrations quickly convinced Houston management of the concept’s merit, and Kinzler was encouraged to continue.²⁹

Selection of the prime shade was a major topic of a telephone conference of Skylab officials on the 19th. The decision to delay the crew’s launch the second time had eliminated the SEVA sail’s principal advantage. Flight controllers had reservations about it anyway-its deployment would cap a rugged 22-hour launch day for the astronauts. Furthermore, the Apollo thrusters might contaminate the telescope mount and its solar panels. Medical representatives favored the parasol, not wanting to chance an EVA early in the mission before the crew was acclimated to space. Deke Slayton stressed that using the scientific airlock was “the most direct approach and the least difficult [operation]for the crew.” Schneider believed Huntsville’s twin-pole sail had the best chance of success, but Kraft wanted to eliminate it because it was 25 kg overweight. During a second status briefing that night, JSC’s director recommended further development of the SEVA sail in case Huntsville’s should fail neutral-buoyancy tests. The group approved Kinzler’s parasolConrad’s preference-placing it ahead of the twin-pole sail.³⁰

Confident that its twin-pole shade would work in space, the Huntsville group designed it for easy deployment in the neutral-buoyancy tank. As Schweickart recalled, “our real challenge . . . was convincing management that we could do it.” In several instances “we set about designing the equipment [to] look good.” In spite of the tight schedule, Marshall observed its traditional steps of design and development, including preliminary and critical design reviews, bench, checks, and static and dynamic structural tests. Huntsville aimed at completing its shade by the 22d, when NASA management would review the deployment in the neutral-buoyancy simulator. A tank test on the 18th confirmed the shade’s feasibility, but also indicated that the pole sections could separate under stress. After the locking nut was modified, the shade’s weight was reduced, and teflon inserts were placed in the eyelets to reduce friction, the dress rehearsal in the tank went off without a hitch.³¹

In Houston, Kinzler’s parasol was nearly made over. The fiberglass fishing poles were replaced by stronger aluminum rods, the coiled springs by a “rat-trap” spring. The canopy had to be enlarged when Huntsville’s thermal engineers calculated the exact requirements. Perhaps the biggest change involved the shape of the frame. The airlock was found to be considerably off-center of the area to be shaded. Since there were distinct advantages in packing and deploying a symmetrical frame, Kinzler designed all four arms to the same length, 6.5 meters, letting the rods on two sides extend beyond the off-center canopy.³²

After the 17th, Director Chris Kraft concentrated most of JSC’s resources on the parasol. Faget’s engineering division provided design support while Donald Arabian, program operations manager, directed configuration control and testing. Arabian quickly expanded the parasol team beyond Houston, farming out specific requirements to North American and Grumman. During the second week, he and Kinzler supervised development, exercising joint veto power over changes. Both men recall a lot of “engineering after the fact.” If something looked like it would work, they built it and designed the details later.³³

Certain basic criteria governed the selection of shade material, the foremost being its thermal performance. The material also had to be lightweight, compact, deployable, noncontaminating, and stable over a wide range of temperatures. Materials were unacceptable if they tended to retain their stowed configuration when deployed. “What appeared to be a relatively straightforward design problem to some of the enthusiastic shield designers turned out to be a nightmare of complexity when all the . . . design criteria were addressed.” A spacesuit material consisting of nylon, mylar, and aluminum was selected. Less than 0.1 mm thick, it met all the criteria but one-nylon had a marked tendency to deteriorate under ultraviolet rays. Deterioration could be reduced by applying thermal paint to the nylon. The paint added considerable thickness to the material-no problem for the SEVA and twin-pole sails, whose containers had room to spare, but the parasol fitted tightly into a small container. Houston canvassed the country for information, finding no precise data on nylon’s long-term exposure in a vacuum, partly because NASA had avoided using nylon in space. Before the end of the first week, Houston opted to go without the paint; the second crew could replace the parasol if it deteriorated.³⁴

Huntsville had less confidence in the unpainted nylon. Several days after the accident, Robert Schwinghamer’s office began testing JSC’s shades as part of a program that involved a dozen materials and 49 tests. After 100 hours of solar-vacuum testing, nylon lost half its pull strength. Houston officials were not greatly worried by these results or similar findings at their own center; they believed the inner surface, aluminum, would reflect most of the heat in any case. The Huntsville studies, however, showed a decline in shielding performance as well as strength.³⁵

At the design certification review in Huntsville on the 23d, every major aspect of Skylab’s problem was covered, with particular emphasis on sunshade candidates and materials testing. Houston’s spokesman summed up the case against nylon: although test results varied, all showed the material deteriorating in time under exposure to ultraviolet rays. In executive session, Skylab’s top officials agreed to retain the parasol as their first choice but with a protective covering for the nylon. Houston, anticipating such a decision, had selected kapton, an ultraviolet-resistant tape. The twin-pole and SEVA sails, made from the same nylon-reinforced material, would be covered with thermal paint. Langley Research Center was directed to continue work on an inflatable device in the event there should be an unexpected hitch with the parasol.³⁶

The decision in Huntsville left JSC less than a day to modify its two shades. Wednesday evening crews began applying kapton to the parasol and spray-painting a SEVA sail. Caldwell Johnson’s team quickly ran into problems on the latter; contaminants in the paint required a lengthy straining process, and the oven-drying took longer than expected. By Thursday morning it was uncertain whether the SEVA sail would dry in time for the launch. Parasol modifications proved even more troublesome as the additional bulk of the kapton made stowage difficult and release even harder, raising serious doubts that the shade would work in space. Morale at the Houston center, at a high point a day earlier, plummeted. At a final review at Kennedy, the parasol, with its nylon unprotected, was reconfirmed as the primary device. The educated guess of most materials experts was that the nylon would last at least 28 days. Marshall’s twinpole shade would be deployed later if the parasol showed signs of deterioration.³⁷

In Houston, packing the parasol proved difficult, even without the kapton. In its final configuration, the extension rod was recessed more than expected. Kraft noted that the astronauts would have a difficult time connecting the sections of rod. The parasol team agreed to add a 5-cm sleeve. Manufacturing began as the parasol was delivered to Ellington Air Force Base; the new piece followed on a separate flight to the Cape, arriving just before final closeout of the spacecraft.³⁸

PLANS TO INCREASE SKYLAB’S POWER

NASA’s immediate electrical problem was to reduce power requirements; but for the long run, more power had to be provided. The ATM and Apollo electrical systems, though adequate for most of the first mission, would fall far short on the 56-day flights. Schneider put Houston and Huntsville to work on promising concepts. JSC examined a solar winged module to dock at the side port of the docking adapter; Marshall investigated variations of a portable array. The necessary hardware modifications precluded the use of either by the first crew, but there was a third option. Telemetry suggested that remnants of the meteoroid shield still held one of the two workshop arrays in place. Its release would solve the problem quickly. The debris might be cleared the first day, during a standup EVA from the Apollo hatch. I t was just a hope; Schneider told a press conference that “we’re not too optimistic.” More likely, NASA would have to settle for photographs that would improve the chances of deployment later.³⁹

The chief of Marshall’s Auxiliary Equipment Section was given the responsibility of developing tools to cut away debris. He started with tree-trimming shears from a Huntsville hardware store and then called the A. B. Chance Company of Centralia, Missouri, maker of tools for power companies. Chance officials agreed to display their complete line of tools in Huntsville the following day. Two items were selected: a cable cutter and a universal tool with prongs for prying and pulling. Both were modified for mounting on a 3-m pole.⁴⁰

While the tools were under development, Huntsville’s Space Simulation Branch prepared a Skylab mockup in the neutral-buoyancy tank, complete with loose wires, twisted bolts, and fragments of a meteoroid shield. Close by, supports were installed for a model of the command module, flown in from Houston. NASA officials evaluated the tools on the 21st, and the following day astronaut Paul Weitz practiced freeing a solar array. The tools had already left for Kennedy when the certification review ruled that the pointed tips of the cutters were a hazard. New heads with blunt tips were quickly prepared and the change made at the launch site.⁴¹

LAUNCH AND DOCKING

Final launch activities were interrupted by a lightning strike on the service structure’s mast that knocked a spacecraft gyroscope off line. The guidance and navigation system was quickly retested and the count resumed. The schedule was altered when the parasol’s delivery became problematic; propellants were loaded three hours early and final stowage delayed until 3:00 a.m. At that hour, the crew was preparing to board.⁴²

Liftoff on the morning of 25 May 1973 was flawless. By midafternoon the crew had reached Skylab and found it very much as expected. “Solar wing two is gone completely off the bird,” Conrad reported. “Solar wing one is . . . partially deployed. . . . There’s a bulge of meteoroid shield underneath it in the middle, and it looks to be holding it down.” Sunlight had blackened the gold foil on the workshop’s exterior. More important, the scientific airlock was virtually free of debris. During the inspection, Weitz had trouble televising the damaged area from the spacecraft’s cramped quarters, but Houston acknowledged “some pretty clear views.” Conrad completed the fly around, optimistic that the crew could free the array in standup EVA.⁴³

The astronauts ate dinner before trying to extend the array. Weitz manipulated the tools while standing in the open hatch, as Kerwin held his legs and Conrad maneuvered the spacecraft. When Apollo passed over the California tracking station 40 minutes later, the crew was having obvious difficulties. Absorbed in their problem, the astronauts were venting their frustration with four-letter words, while Houston repeatedly tried to remind them that communication had resumed. Conrad’s report was gloomy; the metal strip wrapped across the array beam, though only a centimeter wide, was riveted in place by several bolts that had apparently fastened themselves to the array as the shield tore away. Weitz had pulled the panel with all his strength but to no avail. Conrad summed up the situation as the spacecraft headed into the earth’s shadow: “We ain’t going to do it with the tools we got.”⁴⁴

The crew then expected to end the work day by docking. When Conrad attempted it, however, the probe did not engage the drogue. He tried two backup procedures with no more success. Flight controllers proposed docking with the circuit breakers open, but this also failed. By 9:00 p.m., the crew was down to its last alternative, donning the pressurized suits to attempt another repair by EVA. While practicing that emergency procedure in Houston, Conrad had jokingly told Kerwin that if events ever reached that stage, they were coming home. Faced with a real problem, Conrad radioed Mission Control, “We might as well . . . try the EVA. Because if we ain’t docked after that, I think you guys have run out of ideas.”⁴⁵

The procedure involved depressurizing the spacecraft, opening the forward tunnel hatch, and removing the probe’s back plate to bypass some of the electrical connections. Then, centering the probe and drogue, the crew used the Apollo’s thrusters to close on the docking adapter. When the two docking surfaces met, all 12 latches properly engaged. While the program managers held a midnight press briefing, the crew straightened up the Apollo cabin to close out a 22-hour day.⁴⁶

ACCOMPLISHING THE REPAIR

Despite the first day’s troubles, NASA officials remained optimistic about deploying the parasol. The crew entered the workshop in midafternoon on the 26th, having first activated the docking adapter and airlock. Weitz reported a dry heat, “like the desert.” The crew proceeded deliberately, leaving the workshop on occasion for relief from the heat. The operation took about two hours. After connecting the parasol canister to the scientific airlock and opening the port, the astronauts threaded extension rods and gradually extended the parasol. When the folded arms finally swung outward, spreading the fabric, the crew was disappointed. Conrad reported that “it’s not laid out the way it’s supposed to be.” He estimated that the wrinkled canopy covered only about two-thirds of its intended area. At Mission Control, however, the news of a clean deployment was greeted with cheers. Houston officials believed the wrinkles had set in during the cold of the lengthy deployment (the shade had been extended but unopened in the dark portion of the orbit) and they expected the material to stretch in the sunlight.⁴⁷

The workshop cooled considerably in the next three days. The temperature on the external surface dropped 55°C overnight. Internal temperatures reacted more slowly, falling 11°C the first day. The outline of the parasol could be traced by running a hand along the workshop wall; the uneven coverage left hot spots, including an area near Joe Kerwin’s sleeping compartment. By the 29th, engineers had concluded that the workshop would stabilize near 26°C, about 5°C above the desired level but still tolerable. Full-scale operations began that day with medical tests, solar observations, and preparations for the initial earth-resources pass. Power consumption ran very close to Skylab’s output of 4.5 kw, particularly when the crew operated the telescope mount, which drew 750 watts. At the evening news briefing, Flight Director Neil Hutchinson acknowledged that the power limitation was one of several problems complicating the early flight planning.⁴⁸

The 30th brought yet another crisis. The earth-resource maneuver involved taking Skylab’s solar panels out of direct sunlight and relying on batteries for power. As the spacecraft passed through the earth’s shadow, four batteries dropped off line. Despite repeated attempts, flight controllers could restore only three of them when the workshop returned to its solar inertial attitude. The loss, Skylab’s second in a week, reduced power capacity by another 250 watts and raised serious doubts about the soundness of the electrical systems.^* On the 31st, the Management Council moved the launch date for the second crew ahead two weeks because of the worsening conditions. The group discussed possibilities of freeing the solar array and set 4 June as the date for a decision.49

A team led by Rusty Schweickart had been studying the solar-array problem since the day after launch. Talks with the crew helped fill in some of the blurred televised pictures so that by the 29th, Huntsville had a reasonable facsimile of the jammed array. During the next four days, the group developed a difficult but feasible procedure. Exiting from the airlock port, two crewmen moved through the airlock trusses to the long antenna boom at the forward edge of the workshop. After attaching an eight-meter cable cutter to the debris, one astronaut used the pole as a handrail to reach the solar array. There he connected a beam-erection tether-a nylon rope with hooks at each end-between the solar wing and the airlock shroud; the tether would be used to break a frozen hydraulic damper on the array once the debris had been removed. The most difficult aspect of the operation was the lack of footholds which would allow the astronauts to work with both hands. By 2 June, however, Schweickart and Ed Gibson had demonstrated the procedure successfully in the water tank. What could be done there could usually be repeated in space.⁵⁰

In Huntsville on 4 June, the Management Council received a bleak picture of Skylab’s condition. If no more batteries failed, the first crew could probably complete the scheduled experiments. Without some additional power, the next two crews could not. Schweickart reviewed the procedure to free the solar array and showed films of his practice session in the tank. Some members expressed reservations. Attaching the cutter to debris eight meters away seemed a tricky maneuver at best, and there was no alternate way of securing the pole. Nor was it clear that the strap running over the solar array was the only thing preventing its release. Nonetheless, the group approved the attempt. The extravehicular activity was no more hazardous than other EVAs, and success promised large gains in power. Even failure might provide valuable information for a later attempt.⁵¹

That evening Schweickart gave the crew a brief description of the operation. After the crew was asleep, a list of tools, assembly instructions, and detailed steps followed over the teleprinter. The astronauts reviewed the procedure in their spare time and resolved a few questions during an hour-long session with Houston the following evening. On the 6th, the crew rehearsed the operation inside the workshop, communicating with Mission Control by television as well as radio. Kerwin donned his pressure suit for a more realistic simulation, and Conrad made several small changes in the beam-erection tether. Neither was particularly optimistic about their chances.⁵²

The crew opened the airlock hatch just before Skylab began a dark pass on the morning of 7 June. Conrad assembled the tools under the lights of the airlock shroud, and the two men moved to the antenna boom. When it was light enough, Kerwin tried to fasten the cutters. His initial attempts failed. In the Huntsville tank, Schweickart had placed his feet at the base of the antenna; on the flight model, cable connectors were in the way. As Kerwin recalled, “one hand was essentially useless wrapped around the antenna-and with the other hand I couldn’t control the pole. . . . Every time you would move it, your body would react and move the other way.” On several occasions Kerwin got the jaws of the cutter close to the restraining strap, only to have the pole move as he brought his hand from the antenna to open the cutters. When Houston lost communications at 11:42 a.m., Kerwin had been hard at-work over half an hour, his pulse reaching 150. Then Kerwin hit upon an idea that saved the day. He shortened the tether that ran from his suit to the antenna by doubling the line, thereby establishing a firm position against the edge of the workshop. Ten minutes later the crew notified Houston that the pole was fastened securely to the array.⁵³

Although the worst was over, the crew had more problems. At the solar array, Conrad could attach only one of two hooks on the erection tether; the holes on the array were a bit smaller than those on the ground model. After struggling with the second hook for a time, he decided to make do with only one. Kerwin cut the restraining strap without much trouble, but releasing the frozen damper proved more difficult. The two men working together finally succeeded. Asked by Houston how the array had deployed, Conrad laughingly responded:

I’m sorry you asked that question. I was facing away from it, heaving with all my might and Joe was also heaving with all his might when it let go and both of us took off. ...By the time we got settled down and looked at it, those panels were out as far as they were going to go at the time.

By the next day they were fully extended and producing nearly 7 kw of power.⁵⁴

INVESTIGATION BOARD

Congressional critics were quick to fault NASA for the accident. Senate space committee chairman Frank Moss called for NASA to investigate, which agency policy required in any event. On 22 May, Bruce Lundin, director of Lewis Research Center, was asked to head an inquiry. His committee first examined the flight data to establish the events surrounding the accident. Then, having settled on failure of the meteoroid shield as the primary cause of the accident, the board reviewed its development in great detail, concentrating on the management aspects of design, fabrication, and testing. The inquiry included visits to the three manned space centers and McDonnell Douglas’s Huntington Beach plant before the report was completed in early July.⁵⁵

The board examined 10 ways that the shield might have failed, but considered only 2 as likely. The first involved the space between the edge of the shield and the workshop wall. Although NASA had stipulated that the shield fit tightly against the tank, at launch the shield had gaps that exceeded design specifications by half a centimeter. Wind-tunnel tests confirmed that a buildup of pressure in these spaces could have led to the accident. Flight data, however, pointed toward the shield’s auxiliary tunnel as the probable cause of the accident. The tunnel, used as a conduit for wires, was designed to vent pressure as the launch vehicle rose through the atmosphere. But the tunnel had not been constructed as designed, and pressure could build up.⁵⁶

Lundin’s committee theorized that the pressure may have become high enough to lift the shield into the airstream one minute after launch. As the shield ripped away, it wrapped around one solar array and broke the latches on the other. Forces of gravity and aerodynamic drag held the array close to the workshop for over eight minutes until the spent S-II stage separated from the workshop. When the stage’s retrorockets fired, the exhaust tore the solar array from its hinge.⁵⁷

Why had NASA and McDonnell Douglas failed to detect the deficiency in six years of development and testing? The board blamed the error in part on the presumption by Skylab engineers that the shield would fit tightly, as specified in design criteria. The actual shield proved to be a “large, flexible, limp system” that could not be rigged to design specifications. The committee criticized NASA’s failure to treat the shield as a separate system with a project engineer responsible for all its details. There was no evidence that development had been compromised by a lack of time, money, or expertise. Instead, the committee attributed “the design deficiencies . . . and the failure to communicate within the project . . . to an absence of sound engineering judgment and alert engineering leadership concerning this particular system over a considerable period of time.”⁵⁸

Among the board’s specific recommendations, three had broad significance for NASA management. One called for the appointment of a project engineer on complex items that involved more than one engineering discipline. The second warned against undue emphasis on documentation and formal details: “Positive steps must always be taken to assure that engineers become familiar with actual hardware, develop an intuitive understanding of computer-developed results, and make productive use of flight data in this learning process.” Finally, the board encouraged the assignment of an experienced chief engineer to major projects such as the workshop or airlock. Freed from administrative and managerial duties, he would “spend most of his time in the subtle integration of all elements of the system under his purview.”^59/a>