1963 - 1964

With the signing of the lunar module contract, the Manned Spacecraft Center and Grumman began the design and development of a vehicle that would land two men on the moon and, subsequently, take them off. When NASA selected Grumman in late 1962 to build this final piece in Apollo’s stack, the landing craft was still a long way from a “frozen” hardware design. While the command and service modules were evolving from Block I to a more advanced Block II version during 1963 and 1964, the lunar module was also changing, moving toward the huge, spidery-legged bug that later landed on the moon.

External Design

Houston and Grumman engineers had spent a month in negotiations and technical groundwork before signing the contract on 14 January 1963. Although ratification by NASA Headquarters was not forthcoming until March, Grumman forged ahead, devoting most of the first three months to establishing a practical external shape for the vehicle.¹

Cooperation between customer and contractor got off to a fast start. In late January, officials from the Houston Apollo office visited Grumman to review early progress, to schedule periodic review meetings, and to establish a resident office at Bethpage similar to the one already operating in Downey. Then, following a tradition that had proved effective in other programs, the Houston office set up spacecraft and subsystem panels to carry out technical coordination. Thomas J. Kelly had directed Grumman’s Apollo-related studies since 1960, earning for himself the title “father of the LEM,” but the vehicle that finally emerged was a “design by committee” that included significant suggestions from the Houston panels, notably Owen E. Maynard’s group.²

Using Grumman’s initial proposal for the lunar module as the departure point for continuing configuration studies and refining subsystem requirements, the team that had guided the company through its proposal spearheaded the design phase. When the contractor assigned 400 engineers to this task, an optimistic air about how long it would take pervaded both Bethpage and Houston. The job took longer than the six to nine months originally anticipated, however, because of special efforts to guard against meteoroids and radiation and to incorporate criteria imposed by the unique lunar environment.

Basic elements in Grumman’s proposal remained the same: the lunar module would be a two-stage vehicle with a variable-thrust descent engine and a fixed-thrust ascent engine; and the descent stage, with its landing gear, would still serve as a launch pad for the second, or ascent, stage.^* But almost everything else changed. As a first step in defining the configuration, Grumman formed two teams to study the ascent stage. One group examined a small cabin with all equipment mounted externally, and the other studied a larger cabin with most equipment internal. The findings of the two teams pointed to something in between. The spacecraft that ensued was ideally suited to its particular mission. Embodying no concessions to aesthetic appeal, the result was ungainly looking, if not downright ugly. Because the lunar module would fly only in space (earth orbit and lunar vicinity), the designers could ignore the aerodynamic streamlining demanded by earth’s atmosphere and build the first true manned spacecraft, designed solely for operating in the spatial vacuum.³

At a mid-April 1963 meeting in Houston, Grumman engineers presented drawings of competing configurations, showing structural shapes, tankage arrangements, and hatch locations. Grumman and Houston officials then worked out the size and shape of the cabin, the docking points, and the location of propellant tanks and equipment. The basic structure and tankage arrangement was cruciform, with four propellant tanks in the descent stage and a cylindrical cabin as the heart of the ascent stage, which also had four propellant tanks. Still to be resolved were questions of visibility, entrance and exit, design of the descent engine skirt (which must not impact the surface on landing), and docking and hatch structures.⁴

In late April and early May, Maynard (chief of spacecraft integration in MSC’s Spacecraft Technology Division) summarized for Director Robert Gilruth the areas still open for debate, especially the landing gear and the position of the landing craft inside the launch vehicle adapter. Another sticky question, he said, was the overall size of the vehicle, which dictated the amount of propellants needed to get down to the moon and back into orbit. The lunar module structure, especially the descent stage, would be wrapped around the tanks; as the tanks were enlarged, the vehicle design would have to grow to accommodate them. There was one ray of light, however; Marshall was talking about increasing the lifting capability of the Saturn V launch vehicle from 40,800 kilograms to 44,200. With that capability, the target weight for the lander could be pegged at between 12,700 and 13,600 kilograms, instead of the 9,000 kilograms listed in the proposal.⁵

One early concern, though not directly connected with external design, was the firing of the ascent engine while it was still attached to its launch pad, the descent stage. The exhaust blast in the confined space of the interstage structures - called FITH for fire-in-the-hole - could have untoward effects. Some observers feared that the shock of engine ignition might tip the vehicle over. And what would happen if the crew had to abort during descent, shed the descent stage, and return to lunar orbit? This would require extra fuel, posing yet another weight problem. Scale model tests in 1964 allayed these misgivings to some degree, but the real proof had to wait for a firing test in flight of a full-scale vehicle.⁶

Although the descent structure, with its four propellant tanks, appeared practical from the standpoint of weight and operational flexibility, the ascent stage was harder to pin down. Nearly two years passed before the cabin face, windows, cockpit layout, and crew station designs were settled. By late 1963 Grumman engineers had begun to worry about the weight and reliability of the four-tank arrangement, with its complicated propellant system. They recommended changing to a two-tank model, and Houston concurred. Redesign delayed the schedule ten weeks at an added cost of $2 million, but the system was much simpler, more reliable, and lighter by 45 kilograms. Yet the change brought its own problems. Because oxidizer was heavier than fuel, four tanks had allowed the engineers to put one tank of each on either side of the cabin for balance. With only two tanks, some juggling had to be done to maintain the proper center of gravity. The fuel tank was moved farther outboard than the oxidizer, giving a “puffy-cheeked” or “chipmunk” appearance to the front of the vehicle.⁷

Also shaping the face of the ascent stage were its windows. Windows were basic aids for observation and manual control of the spacecraft, and the pilots expected to use them in picking the landing site, judging when to abort a mission, and guiding the spacecraft during rendezvous and docking with the command module.

The importance of visibility was recognized early in Houston’s studies and stressed in Grumman’s original proposal. In both, large windows afforded an expansive view. Grumman had featured a spherical cabin like that of a helicopter, with four large windows so the crew could see forward and downward. This design was discarded because large windows would require extremely thick glass and a strengthening of the surrounding structure. The environmental control system would have trouble maintaining thermal balance. Two smaller windows could replace the four large ones, but the field of view would have to remain very much the same. To get the required visibility with smaller and fewer windows, Grumman had to abandon its spherical cabin design. The new cylindrical cabin had a basically flat forward bulkhead cut away at various planar angles; the large, convex windows gave way to small, flat, triangular panes (about one-tenth of the original window area) canted downward and inward to afford the crew the fullest possible view of the landing area.⁸

Grumman’s change to a cylindrical cabin posed another problem. A spherical shape is simple from a manufacturing standpoint, because of the relative ease in welding such a structure. The new window arrangement and front face angularity made an all-welded structure difficult. The Grumman design team wrestled with the new shape and in May 1964 adopted a hybrid approach. Areas of critical structural loads would be welded, but rivets would be used where welding was impractical. Grumman neglected to inform Houston of the switch in manufacturing processes, but a Houston engineer noticed the combination of welding and riveting while on a visit to Bethpage.

Toward the end of May, there was a series of reviews and inspections of Grumman’s manufacturing processes. NASA representatives looked at welding criteria, mechanical fastening techniques, and the behavior of sealant compounds under temperature extremes and a pure oxygen atmosphere. The contractor demonstrated that its part-riveted structure showed very low oxygen-leak rates in testing. Although Manned Spacecraft Center officials tentatively approved the change, they left an engineer from the MSC Structures and Mechanics Division in Bethpage to watch Grumman closely. Marshall experts visited Grumman from time to time to extol the virtues of an all-welded design and to warn of the problems of mechanical fabrication. But the peculiarities of the lunar module made a mix of the two techniques almost inevitable.⁹

Tailoring the Cockpit

The lunar module’s interior was as different from that of other manned spacecraft as its exterior. And it also took two years to design. A home on the moon required some very special features besides visibility: equipment and procedures for rendezvous and docking, environmental control for living, an easy means for leaving and reentering while on the moon, and the capability of operating in a low-gravity or no-gravity environment.¹⁰

With an internal volume of 60 cubic meters, the lunar module would be the largest American spacecraft yet developed. It would also be the most spacious, except for the command module when the pilot was there alone. To lessen already formidable crew training demands, Houston pressed Grumman to make the cabin instruments and displays as similar as possible to those of the command module. Complete duplication was impossible, however, because the two craft were so unlike. Ground rules were laid down governing the degree of redundancy required in controls and panels. Although these controls would be duplicated on each side of the cockpit, some of the instrument displays would have to be shared by the crewmen. Above all, Grumman was told, the spacecraft must be designed so that the hover and touchdown could be flown manually and so that no single failure of the controls or displays could cause a mission abort.¹¹

Because the lunar module was a means of transportation, as well as shelter and living quarters for the crew while on the moon, cockpit design presented interesting problems to human factor engineers. The man-machine interface embraced such items as stowage of space suits and personal equipment and room for the pilots to move about within the cabin. In a mockup in mid-1964, two crewmen demonstrated that they could put on and take off their portable life support systems with suits either pressurized or deflated, reach for and attach umbilical hoses, and recharge their backpacks. The MSC Crew Systems Division drew up a document governing spacecraft-spacesuit interface and change procedures. This was used by NASA to supplement spacecraft specifications and interface control documents. It was also an important managerial tool between Grumman and North American and their major associates, MIT and Hamilton Standard (developers of the guidance and navigation system and the life support system).¹²

The astronauts were an essential “subsystem” on the lunar module, and they were very much in evidence at Bethpage, as well as at Downey, where they helped in the design of the command module. Scott Carpenter, Charles Conrad, and Donn F. Eisele drew the lunar module as their special assignment, and William F. Rector, the lunar module project officer, frequently called upon them for help. He also urged other astronauts to take part in the periodic mockup reviews and significant design decisions: “They should be [part] of it,” Rector said. “They’re going to fly it.” This was not an unusual arrangement; astronauts, being both engineers and test pilots, have played an active role in the design and development of every manned American space vehicle.^*

Conrad probably worked more on the vehicle’s basic design than any other pilot, as the configuration evolved. Rector relied on him to sound out the crews on cockpit features - controls, switch locations, and visibility, among others. One innovation which Grumman favored, and which Conrad was instrumental in getting incorporated, was electroluminescent lighting. An inherent problem in both aircraft and spacecraft had been light intensity that varied from panel to panel. This uneven lighting made it difficult for a pilot to scan his instruments rapidly and to adjust quickly to low-level exterior light conditions. Electroluminescence, a wholly new concept that used phosphors instead of conventional filament bulbs, afforded an evenness in intensities hitherto unequaled in any flying craft. At the same time, it weighed less and used far less power than incandescent lighting. Conrad also got this new system into the Block II command module.¹³

The seating arrangement in the lunar module was perhaps the most radical departure from tradition in tailoring the cockpit. It soon became apparent that seats would be heavy, as well as restrictive for the bulky space suits. Bar stools and metal cagelike structures were also considered and discarded. Then an idea dawned. Why have seats in the lander at all? Its flight would be brief, and the g loads moderate (one g during powered flight and about five on landing). Since human legs were good shock absorbers, why not let the crew fly the lunar module standing up?

This concept was bandied about rather casually at first by two Houston engineers, George C. Franklin and Louie G. Richard. Franklin then went with Conrad to talk to Howard Sherman and John Rigsby at Bethpage. These Grumman employees, in turn, passed the idea along to Kelly and Robert Mullaney. At this point, the seat and window problems merged. Standing up, the crew would be close enough to the windows to get a larger field of view (one engineer estimated it at 20 times greater) than with any seating arrangement yet suggested. Moreover, since cockpit designers would not have to worry about knee room, the cabin could be shortened, saving 27 kilograms and improving the structure. Conrad called it a “trolley car configuration,” and said, “We get much closer to the instruments without our knees getting in the way, and our vision downward toward the moon’s surface is greatly improved.”

Grumman technicians later devised a restraint system to hold the pilots in place during weightless flight and prevent them from being jostled about the cabin during landing. Resembling the harness used by window washers and linked to a pulley and cable arrangement under constant tension, it was augmented by handholds and arm rests and by Velcro strips to keep the pilots’ feet on the floor.¹⁴

Hatches and Landing Gear

The lander originally had two docking hatches, one at the top center of the cabin and another in the forward position, or nose, of the vehicle, with a tunnel in each location to permit astronauts to crawl from one pressurized vehicle to the other. (Extravehicular transfer between craft remained an emergency backup method.) After injection into a translunar trajectory, a course toward the moon, the command module pilot would turn his ship around, fly up to and dock with the lander’s upper hatch, and then back the two vehicles away from the spent S-IVB third stage. This top-to-top docking arrangement aligned the thrust vector of the service module propulsion engine with the centers of gravity of the two spacecraft, thus avoiding adverse torques or tendencies to tumble during firings for midcourse corrections and injection into lunar orbit. The crew would enter the lunar module through this hatch. When the lander returned from the moon, however, the front hatch would be used for docking and crew transfer. With no windows in the top of the lander, the lunar pilots would be flying blind if they docked with the upper hatch. One of Grumman’s human factor experts later said, in an apt analogy, “It’s nice to see the garage . . . when you drive into it.”¹⁵

By spring 1964, NASA and Grumman engineers were thinking of deleting the front docking procedure and adding a small window above the lunar module commander’s head. This overhead window might add seven kilograms weight and some extra thermal burden, but cabin redesign would be minimal. The added weight would be offset by eliminating the front tunnel and the extra structural strength needed to withstand impact loads in two areas. Eliminating forward docking had another advantage. The hatches could now be designed for a single purpose - access to the command module through one hatch and to the lunar surface through the other - which certainly simplified the design of the forward hatch. NASA directed Grumman to remove the forward docking interface but to leave the hatch for the astronauts to use as a door while on the moon.¹⁶

Once the location of the hatches was settled, getting the astronauts out and onto the lunar surface had to be investigated. Using a cable contraption called a “Peter Pan rig” to simulate the moon’s gravity, Grumman technicians looked into ways for the crews to lower themselves to the lunar surface and to climb back into the spacecraft. When astronaut Edward White, among others, scrambled around a mockup of the lander, using a block and tackle arrangement and a simple knotted rope, he found that both were impractical. In mid-1964 a porch, or ledge, was installed outside the hatch and a ladder and handrail on the forward landing gear leg. When the astronauts discovered they had trouble squeezing through the round hatch in their pressurized suits and wearing the bulky backpads, the hatch was squared off to permit easier passage.¹⁷

All these design features, although unusual, appeared to be compatible with the lunar environment - at least the engineers did not entertain any special worries. But the landing gear was different. The design of the legs and foot pads depended on assumptions about the nature and characteristics of the lunar surface. In the absence of any firm knowledge and with scientific authorities differing radically in their theories, how should one design legs to support a craft landing on the moon?

Grumman had first considered five legs but, during 1963, decided on four. The change was dictated by the weight-versus-strength tradeoff that had produced the cruciform descent stage, with its four obvious attachment points. The revised gear pattern also greatly simplified the structural mounting of the vehicle within the adapter. Four legs set on the orthogonal axes of the lander (forward, aft, left, and right) mated ideally with the pattern of four reaction control “quads” (the basic four-engine package). The quads were rotated 45 degrees so the downward-thrusting attitude control engine fired between the two nearest gear legs, overcoming a severe thermal problem of the five-leg arrangement.¹⁸

While Bethpage was wrestling with the legs, Houston decided it had been too optimistic about the load-bearing strength of the lunar surface in the request for proposals. The resulting revision placed heavier demands on the landing gear, and Grumman had to enlarge the foot pads from 22 to 91 centimeters in diameter. The bigger feet made the gear too large to fit into the adapter. A retractable gear therefore replaced the simpler fixed-leg gear. Retractability also figured in the shift from five to four legs - the fewer to fold, the better.

Leg experts at Grumman had to change the geometry of the undercarriage, devise the best structure for impact absorption and stability upon landing, and choose the most suitable folding linkages. A broad program of computer-assisted analysis at Houston and Bethpage was used to determine the worst combinations of conditions at impact. The studies were reinforced by drop tests of lander models at Houston, Bethpage, and Langley. There were also plans to drop-test full-sized test articles to check out the new designs.¹⁹

During 1963 Grumman engineers continued to worry about the nature of the lunar surface and to carry on theoretical and simulation studies of lunar geology and soil mechanics, with the support of such consulting firms as the Stevens Institute of Technology in New York and the Arthur D. Little Company in Massachusetts. Much of this work covered the interaction between vehicle and surface at the moment of landing. What would happen to the landing gear at touchdown? Would the lunar dust that might be kicked up by the descent engine exhaust obscure the landing site? Would soil erosion affect the stability of the lander? Washington also assisted in this research. In mid-1963, Bellcomm surveyed all that was being done inside and outside NASA and suggested that a backup gear be developed, in case the surface should be more inhospitable than it appeared.²⁰

But Grumman could not wait on the outcome of these studies. At meetings in Houston in October and November, contractor engineers described gears that tucked sideways (lateral folding) for stowage in the adapter; a tripod arrangement (radial), with three struts meeting at the base just above the footpad, that tucked inward; and a cantilevered device, with secondary struts for extra strength that folded inward against the vehicle for stowage and braced the leg when deployed for landing. Houston and Bethpage selected the cantilevered version. Somewhat narrower than the radial one, it was, in many ways, more stable. It had other advantages: less weight, shorter length for easier stowage, and a simpler, and therefore more reliable, folding mechanism.

A landing gear for the lunar surface had to be designed for varying landing conditions, such as protuberances, depressions, small craters, slopes, and soil-bearing strength. To achieve the necessary stability, the landing gear had to be able to absorb a diversity of impact loads. Houston and Bethpage met this challenge by using crushable honeycomb material in the struts, so the gear would compress on impact. A principal advantage of honeycomb shock absorbers was their simplicity. Since they had to work only once, the more common hydraulic shock absorbers and their complexities could be avoided. Subsequently, crushable honeycomb was also applied to the large saucerlike foot pads to improve stability further for landing.²¹

Engines, Large and Small

When Grumman began designing the lunar module in January 1963, its major subcontractors began work on the vehicle’s integral subsystems: Bell Aerosystems, ascent engine; Rocketdyne Division of North American, descent engine; The Marquardt Corporation, reaction control system; and Hamilton Standard Division of United Aircraft Corporation, environmental control. Identifying rocket engines as the most critical subsystem, Grumman started their development first. The lander had 18 engines: 2 large rockets, one for descent to the moon and another for return to lunar orbit, and 16 small attitude control engines clustered in quads and pointing up, down, left, and right, around the ascent stage.²²

During the spring of 1963, Grumman hired Bell to develop the ascent engine, basing the selection on Bell’s experience in Air Force Agena development and hoping that the technology from that program might be applicable to the lunar module. Grumman placed heavy emphasis upon high reliability through simplicity of design, and, in fact, the ascent engine did emerge as the least complicated of the three main engines in the Apollo space vehicle (the descent and service module engines were the other two).^* Embodying a pressure-fed fuel system using hypergolic (self-igniting) propellants, the ascent engine was fixed-thrust and nongimbaled, capable of lifting the ascent stage off the moon or aborting a mission should a landing not be feasible.

There was one major concern about the ascent engine, and that was the usual worry about the ablation material burning off too fast and causing damage to the thrust chamber. Some ablative material eroded during firing tests at Bell’s plant near Niagara Falls and at the Arnold Engineering Development Center in Tennessee. But this erosion was not severe enough to warrant changes in the combustion chambers. In late 1964, Arnold was also the site of a fire-in-the-hole fifth static firing test on a full-scale vehicle to supplement Grumman’s previous scale-model test. The FITH flight test had to wait for later trials at White Sands.

Not everything went well with ascent engine development, however. About a year after the program began, the subsystem manager in Houston discovered that Grumman and Bell were using testing criteria left over from the Air Force Agena program. Since the Agena was unmanned, these were less stringent than NASA demanded for manned spacecraft. More rigorous standards were belatedly imposed by Houston, and a problem was revealed. In “bomb stability” tests, where the engine had to recover from combustion instability caused by an explosive charge within the combustion chamber, the ascent engine “went unstable” (failed to return to normal operation), and structural damage followed. This problem would have to be resolved before the engine could be trusted to bring a crew back from the lunar surface.²³

The lunar module descent engine probably was the biggest challenge and the most outstanding technical development of Apollo. A requirement for a throttleable engine was new to manned spacecraft. Very little advanced research had been done in variable-thrust rocket engines - NASA’s principal effort in this field, the hydrogen - fueled RL-10 used in the S-IV stage of the Saturn, antedating work on the lunar module engine by only a few months. Rocketdyne proposed a method known as helium injection, introducing inert gas into the flow of propellants to decrease thrust while maintaining the same flow rate. Although Bethpage and Houston agreed that this seemed a plausible approach to throttleability, it would be a major advance in the state of the art, and the MSC Apollo office directed Grumman to carry out a parallel development program and select the better design.

On 14 March 1963, Grumman held a bidders’ conference, attended by representatives from Aerojet-General, Reaction Motors Division of Thiokol, United Technology Center Division of United Aircraft, and Space Technology Laboratories, Inc. (STL). In May, STL (which had lost out in the original bidding for the engine) was selected to develop the competitive motor. STL proposed a pressure-fed hypergolic system that was gimbaled as well as throttleable. The engine’s mechanical throttling system used flow control valves and a variable-area injector, in much the same manner as does a shower head, to regulate pressure, rate of propellant flow, and the pattern of fuel mixture in the combustion chamber.

With two subsystem contractors working on such radically different throttling techniques, NASA planners, as Rector later said, “thought one or the other would stub his toe real quick . . ., that it would be obvious that we should go one [way] or the other - but it wasn’t happening. They were both . . . pretty good. . . .” STL and Rocketdyne continued this head-to-head competition for the final-and lucrative-engine development and qualification contract through the end of 1964.²⁴

In November 1964, Joseph Shea, Apollo spacecraft manager in Houston, told NASA Apollo Program Director Samuel Phillips in Washington that he had established a committee^** of propulsion experts from Grumman, the Marshall and Lewis centers, NASA Headquarters, and the Air Force to review the contractors’ efforts and recommend a choice. Selection of one firm over the other rested with Grumman and MSC, in the final analysis, and, Shea stated, “I do feel that we should have the intelligence at our disposal to appreciate all ramifications of [Grumman’s] final recommendation.”

Panel members visited both companies the week of 7 December 1964, but their findings were largely inconclusive. The progress of each firm was nearly identical. Both contractors, although experiencing minor troubles with injector designs, demonstrated satisfactory structural compatibility between injector and thrust chamber. After a year and a half, neither helium injection nor mechanical throttling had proved superior over the other. On 5 January 1965, Grumman decided to stick with Rocketdyne.²⁵

Manned Spacecraft Center Director Gilruth appointed a five-member board^*** to weigh Grumman’s recommendations, review the findings of the earlier committee, and study a technical comparison prepared by Houston’s Propulsion and Power Division. On 18 January this review board, in a surprising move, reversed Grumman’s action and named STL instead of Rocketdyne. The board said that the

recommendation of STL is based upon the assessment that STL is in a more favorable position [and] is capable of supplying more management and superior resources to this program without interference of other similar programs. . . . there are potential benefits to be gained for the Gemini and Apollo attitude engine programs at NAA by the cancellation of the [Rocketdyne] descent engine development.^****

This decision, unusual because Houston rarely vetoed a recommendation for a subcontractor made by a prime contractor, was sustained by Phillips at Headquarters. Shea and Contracting Officer James L. Neal then directed Grumman to proceed with STL.²⁶

Grumman chose Marquardt to build the lunar module’s third engine system, the small 100-pound-thrust attitude control thrusters. In 1960, Warren P. Boardman and Maurice Schenk of Marquardt had visited Robert Piland and Caldwell C. Johnson at Langley to discuss their firm’s propulsion work. Piland and Johnson were intrigued with the idea for a bipropellant thruster that promised to be far superior to the monopropellant engine then used in Mercury. Testing of Marquardt’s product - a dual-valve, pulse-modulated engine with a radiation-cooled combustion chamber - at the Lewis Research Center paved the way for its incorporation into Apollo. Marquardt at first supplied engines for both the command and service modules. In mid1962, NASA decided to use the Marquardt engine for the service module only, because the command module thrusters would be buried within the heatshield, making radiation cooling impossible. Rocketdyne would supply the command module thrusters, which were similar to those it was already developing for Gemini.

Marquardt would furnish attitude control engines and mounting structure and perform some tests of the propellant system. Grumman would provide tanks (purchased from Bell), propellant lines, and the pressurization system. Apollo officials had expected that the service module thrusters, with only slight modifications, could also be used in the lander, but common use proved difficult. The end results, though beneficial, fell far short of Houston’s anticipations. Differing functional requirements, as well as unique environmental and design constraints, precluded direct incorporation of the service module thruster. Houston, however, complained that Grumman failed to take advantage of all the common-use technology available and attributed delays in procurement of many thruster components to this failure.²⁷

After thruster tests at Bethpage and at Marquardt’s Magic Mountain Facility in California during the first half of 1964, a technical problem emerged: the engine spiked, or backfired, at ignition, and a rapid rise in temperature and pressure caused the engine to explode. The spiking appeared so significant that Grumman wanted to develop a backup engine through another source, but Houston refused permission. Marquardt eliminated spiking by installing a small, tubular “precombustion” chamber inside the engine.²⁸

Environment and Electricity

Grumman selected Hamilton Standard to supply the environmental control system for the lunar module. Like AiResearch’s unit in the command module, it was a “closed-loop” atmospheric circulation system, using supercritical oxygen and nonregenerative removal of carbon dioxide to provide a pure oxygen atmosphere. The system also had a liquid-circulating network and heat-absorbent panels to maintain a comfortable temperature inside the cabin. By mid-1964, Hamilton Standard had finished the design phase and begun fabrication and testing. Occasional problems arose during development, but none that threatened the manufacture of a successful subsystem.²⁹

United Aircraft Corporation’s Pratt & Whitney Aircraft Division, a legendary name in aircraft powerplants, was also a pioneer in research on fuel cells using hydrogen and oxygen as reactants to generate electricity. Grumman picked this firm in July 1963 to develop the power system for the lander. The fuel cell program was laden with technical and managerial problems. Many of the lander’s components operated with considerable independence, but the electrical power system had a complex interrelation with virtually every subsystem in the vehicle. The question of how many fuel cell stacks and how many tanks of reactant were needed to meet electrical requirements was, therefore, difficult to answer. In March 1964, Houston approved a three-cell, five-tank arrangement; by summer the fuel cell was in deep technical trouble. NASA and Grumman engineers concluded that it might take more than a year to get the cells working with the other systems properly. The lunar module, which had begun development a year late, did not have the time to spare.

Houston told Grumman in late 1964 to consider substituting batteries for fuel cells, and on 26 February 1965 Bethpage was ordered to make the change. Although the switch was not entirely welcome to the lunar module design team, it caused no appreciable delay. And to some it came as a distinct relief; the beauty of batteries lay in their simplicity, hence their reliability, in contrast to fuel cells. Some of the battery development cost would be offset by the cancellation of the Pratt & Whitney contract.³⁰

The “Sub-Prime” and the Radar Problem

Grumman contracted with Aerospace Communications and Controls Division of Radio Corporation of America (RCA) in Burlington, Massachusetts, for engineering support, radars, an inflight test system, and components of the stabilization and control system. RCA, the “sub-prime” contractor, was also to design and manufacture ground checkout equipment for these items. Although the two companies had worked together for years, the Grumman-RCA experience with the lunar module was fraught with difficulties. Electronics components became a pacing item in the development of the lander’s subsystems, causing unhappiness at NASA Headquarters and culminating in an investigation by the General Accounting Office.³¹

The extremely complex stabilization and control system was the source of much of the trouble. Design had to await definition of mission requirements and planning. To complicate matters further, Grumman did not buy the total system but merely procured parts, through RCA, from Minneapolis-Honeywell, which supplied similar items to North American for the command module. There was some commonality of parts, but the lander hardware had to be repackaged, often causing lengthy delays. Communications gear was purchased from Collins Radio and Motorola in the same manner. Tiring of this roundabout way of doing business, Houston finally decided to speed things up by supplying the television camera, originally intended for development by RCA, as government-furnished equipment. In mid-1964, the Westinghouse Electric Company was asked to submit a bid for the camera.³²

RCA’s role was further cut when inflight maintenance was canceled. At the outset of the program, the crews had been expected to perform basic repairs to electronics equipment in the lander, as well as in the command module, using spare parts stowed aboard the spacecraft. By mid-1963, Houston Flight Operations Director Christopher Kraft was arguing that the crewmen simply would not have time to repair faulty hardware during lunar module operations. Thomas Kelly was convinced that inflight maintenance would degrade reliability instead of improving it. This was probably true, since the electronic spares would be subjected to cabin humidity even when stowed. When George Mueller took over as manned space flight chief in Washington, he also had reservations about the plan. Inflight maintenance was deleted from the program and the crew was to rely on operational displays and the caution and warning system to detect malfunctions. Redundancy would be “wired in,” with duplicate or backup components the crew could switch to, and all electronics inside the cabin would be hermetically sealed to protect against moisture and contaminants.³³

Radar, tied into the guidance and navigation system, was one of the hardest pieces of the lunar module to qualify. Two sets would be used, one for landing, the other for rendezvous. Under its blanket subcontract for electronics, RCA was to design the system, manufacture the rendezvous radar, and buy the landing subsystem. After evaluating proposals from four bidders, RCA picked Ryan Aeronautical Company, developer of landing radar for Surveyor.³⁴

Development of the lunar module radar was not expected to be difficult, since no technological breakthrough was demanded for either system. Integrating these sets with the guidance and navigation system, however, was another matter. There were also problems in properly placing and insulating the antennas. Getting the precise ranging accuracy needed and overcoming the weight increases that resulted from meeting these requirements probably posed the biggest problem of all. A happy medium between optimum weight and desired reliability was elusive, and progress was practically nil.

During the final quarter of 1964, the chief of guidance and control in Houston warned Shea that the radar program was having trouble with weight, accuracy, reliability, thermal characteristics, and costs. Shea and William A. Lee, chief of MSC’s Apollo Operations Planning Division, began to think about omitting the rendezvous radar from both the command and lunar modules. Lee believed these units were doubly redundant, since rendezvous could be performed by the command module pilot with the aid of data relayed by the Manned Space Flight Network. Donald G. Wiseman, an instrumentation and electronics specialist in Houston, thought rendezvous could also be conducted by the lunar module crew, using ground, optical tracking, and S-band and VHF communications equipment ranging information in place of radar. Although not everyone agreed that the system should be eliminated, work was started on the development of an optical tracker.³⁵

Guidance and Navigation

Guidance and navigation was the most difficult of all the lander’s subsystems to develop, both technically and managerially. Development started off simply enough but turned into a complicated tangle. MIT and Houston officials wanted to use the basic command module arrangement in the lander to avoid developing an entirely new system. After Grumman was selected in November 1962, the contractor, the center, and MIT had tried to work out a configuration for the lander. In the middle of 1963, Houston asked Headquarters for permission to to procure lunar module guidance through existing agreements with MIT, AC Spark Plug, Kollsman, Raytheon, and Sperry. When Washington refused, time was lost in negotiating new contracts.³⁶

The biggest delay came from a dispute over whether to use the MIT unit in the lunar module. Grumman’s refusal to accept MIT’s word about the reliability of its system sparked the controversy. Lunar module manager James L. Decker in Houston shared this skepticism and asked Grumman to look into a more advanced system than the three-gimbal platform (pitch, yaw, and roll referencing system) MIT used. Meanwhile, David W. Gilbert, in charge of navigation and guidance in Shea’s office, insisted on getting the MIT unit into the lunar module. Grumman was caught between the two opposing factions. Neither of the Houston officials could get the other to change his mind - and the chasm deepened. Top management in Houston and in Washington then stepped in. Bellcomm would study the options, consult with all parties to the argument, and recommend a solution. In due time, NASA decided to stick with MIT and announced its decision, based on Bellcomm’s findings, on 18 October 1963.

But the announcement did not completely clear the air, and some rather strained feelings developed between Grumman and MIT. Early in 1964, however, the contractors recognized the necessity of working together on the areas where development progress affected both the lunar module and its guidance system. Set down in formal Interface Control Documents, agreements on these points would govern all future actions by both parties. At the end of February, Rector reported 29 meetings between the contractors (with 200 more to go, at this rate, he said) and 55 documents drafted, but almost no concessions by either party. In April, Manned Spacecraft Center managers realized that they would have to intervene to break up the logjam. At a two-day meeting in Bethpage on 25 and 26 June, Shea did just that. After scrutinizing the documents, he mediated the differences and forced the contractors to cooperate.³⁷

Mockup Reviews

At various stages of lunar module design, mockup reviews were conducted to demonstrate progress and ferret out weaknesses. These inspections were formal occasions, with a board composed of customer and contractor officials and presided over by a chairman from the Apollo office in Houston. Usually present were top management personnel from the NASA Office of Manned Space Flight in Washington and from the field centers, as well as a number of astronauts. The vehicle was thrown open for inspection, and the astronauts were expected to climb in, out, over, and around, to get a feel for the craft.

The first of these reviews, on “M-1” (a wooden mockup of the crew compartment), took place 16-18 September 1963. In general, the cockpit layout was acceptable, although the locations of some equipment and the arrangement of controls and instruments still had to be settled. The astronauts liked the visibility through the triangular, canted windows and the standup crew positions; but they wanted the instrument panel changed so both flight stations would have identical displays.³⁸

About six months later, 24-26 March 1964, Grumman showed its second model, “TM-1,” a wooden representation of a complete vehicle. Again attention centered on the cockpit arrangement: support and restraint systems, equipment layout, lighting provisions, location of displays and controls, and general mobility within the cabin and through the hatches. On this occasion, a number of changes were suggested. After evaluation and approval by the review board, these modifications were incorporated into the TM-1 to make up a “design freeze” for constructing an all-metal model, the final review mockup.

TM-1 was far more than just a means to get to the next, more advanced, mockup, however. For several months, Grumman designers used it to study astronaut mobility and spacecraft-spacesuit interfaces. Astronauts and company personnel got into and out of suits inside the cabin, practiced stowing and recharging backpacks, and checked out suit hose connections with the spacecraft’s environmental control system.³⁹

The most important mockup review, in October 1964, centered on “M-5” - a remarkably detailed model of a complete spacecraft, including some actual flight equipment inside the cockpit. Even before the inspection, its prospects for success were discussed in a senior staff meeting at Houston on 2 October. Comparing Grumman’s planned M-5 review with a review held a few days before on the Block II command module at North American, which one official considered “a good display for a salesman [but] a poor engineering tool,” Max Faget said that, in his opinion, North American representatives should go to Grumman to “see what a mockup should look like.” M-5 was the product of two years of configuration studies and the lessons of two previous inspections.

Formal review of M-5 led off with an examination on 5 and 6 October by the astronaut corps. On the following day, MSC Director Gilruth and virtually all the management, engineering, and Apollo leaders from Houston descended on Grumman to inspect the cabin, electrical wiring, plumbing, flight controls, displays, radars, propulsion systems (ascent, descent, and reaction control), environmental control system, communications system, structures and landing gear, and stowage for scientific equipment. No piece of the vehicle escaped the review party’s scrutiny and evaluation. The Mockup Review Board^* met on 8 October, examined the 148 proposed changes, and approved 120 of them. These were mostly minor, and none forced any major redesign. M-5 marked the culmination of the configuration definition.⁴⁰

The Lunar Module and the Apollo Program

Although configuration was not settled and major subsystems development was not begun until near the end of 1964, NASA had begun taking stock of where the lunar module stood in relation to other pieces of Apollo. Structural connections between the lunar module and other Apollo hardware were confined primarily to the command and service modules and the adapter. Unlike its scratchy relations with MIT, Grumman’s association with North American was smooth.^* Early meetings between the contractors were devoted to hardware designs and docking requirements. Initially, each manufacturer was to design and test all equipment mounted on his own vehicle, but in March 1963 North American assumed responsibility for the complete docking device as well as the adapter structure.

Late in 1963, design engineers from Downey recommended, and NASA approved, a center probe and drogue for docking. Stowage of the lander in the adapter was settled in October 1963, when the contractors and Houston agreed upon a truncated cone, 8.8 meters long, with the lunar module mounted against the interior wall by a landing-gear outrigger truss. Thereafter, detailed design focused on the dynamic loads expected during launch and on the deployment of the four panels for removal of the lander during flight. Grumman sent North American a mockup to use in confirming the structural mounting and panel opening characteristics.⁴¹

Lunar module ground testing to prove the practicality of the design and flight testing to verify the spaceworthiness of the flight vehicle also had to be worked into overall Apollo plans. Gilruth had stated that one fundamental requirement for mission success was employing “the kind of people who will not permit it to fail.” The basic reliability philosophy, he said, was “that every manned spacecraft that leaves the earth . . . shall represent the best that dedicated and inspired men can create. We cannot ask for more; we dare not settle for less.” As the lander grew larger and more complex, it became, in the eyes of some observers, the “most critical part of the [Apollo] vehicle.” The many things that could doom the crew made ground-testing all the more important. Reliability for the lander dictated either redundant systems or, where that was impractical because of weight and size, ample margins of safety.

Grumman’s basic plan for ground testing, set forth in May 1963, called for extensive use of test models and lunar test articles (called “TMs” and “LTAs” by the engineers), as well as for propulsion rigs to test propellant lines and for engine firing programs. Because the lander’s flight would be brief, Bethpage engineers adopted a practice of testing hardware until it failed, to provide an indication of strength and to gather information on failure points. Ground testing began with individual parts and subsystems and progressed upward, before the spacecraft was committed to flight.⁴²

Bethpage came up with a scheme for testing the lander in simulated flight by powering the vehicle with six jet engines, to overcome the pull of gravity, and using a modified descent engine to practice maneuvering the vehicle. Although the idea appeared workable, it would be both costly and complex. There were also suggestions for swinging the lander from a gantrylike frame at Langley or from a helicopter or a blimp at White Sands. After a second look, the last two were also scrapped. Grumman and Houston hoped that the lunar landing training vehicle being developed by Bell could test some of the flight components at least, but installing extra equipment might slow the development of the training vehicle. A few flight instruments and the hand controller might be incorporated at a later date into the training vehicle, which the astronauts would use to practice simulated lunar landings. Flight testing within the earth’s atmosphere was finally ruled out when Langley discovered in wind tunnel investigations that the Little Joe II-lander combination would be aerodynamically unstable.⁴³

Grumman had wanted some unmanned missions, using the Little Joe II and the Saturn IB launch vehicles, before men flew the lunar lander. Houston authorized the procurement of autopilots for unmanned spacecraft but did not actually schedule any such flights. After Mueller invoked the all-up concept, with each flight groomed as though it were the ultimate mission, Houston planners began to think about putting both the lander and the North American spacecraft aboard a single Saturn IB. One Houston engineer even went to Huntsville to ask von Braun about the possibility of increasing the launch vehicle’s payload capacity. And there was some discussion about strapping Minuteman missile solid-fueled rocket stages onto the launch vehicle to provide the extra boost needed!

In the meantime, ground testing would have to carry the burden of qualifying the lander until the Saturn was ready to fly the vehicle, which caused some realignment of the lunar module program. Eleven flight vehicles and two flight test articles were earmarked for Saturn development flights. NASA also decided that the first three flight vehicles must be able to fly either manned or unmanned.⁴⁴

In November 1964, Shea, Mueller, and Phillips decided on a tentative flight schedule. Saturn IB missions 201, 202, 204, and 205 would be Block I command module flights. There was no assignment for 203 at this time. Shea told the Houston senior staff that it looked as though an unmanned lander might be flown on 206. The first flight of a combined Block II command module and lunar module would be Mission 207 in July 1967. By that time, the Saturn V was expected to be ready to take over the job of flying the missions.⁴⁵

The lunar module had to be worked into Apollo facilities, as well as into flight schedules. Grumman had its own testing equipment in Bethpage and on the Peconic River, both on Long Island. But the lander’s propulsion systems would have to be tested at the Air Force’s Arnold Center and at White Sands. Fitting the lunar module into the launch complex at the Cape raised some interesting issues. One of the earliest was the rule that any vehicle flown from there must carry a destruct mechanism, in case a mission had to be aborted shortly after launch. The rule was based on a philosophy that it was better to explode propellants in the air than to have them burst into flame on the ground. Houston, however, refused to put a destruct button in the vehicle that was intended to land men on the moon, with the gruesome possibilities of a malfunction on the lunar surface that would either kill the astronauts outright or leave them stranded. Eventually, the Air Force Range Safety Officer agreed to drop this requirement for the lander.⁴⁶

A difficult task at all locations, Bethpage included, was getting ground support equipment (GSE) ready to check out the lunar module subsystems. Traditionally, GSE has been a problem, since it cannot be designed and built until the spacecraft design is fairly firm. Because the lander was the first of its kind and changed from day to day as the mission requirements changed, Grumman was even slower than other contractors in getting its checkout equipment on the line. Shea complained that “the entire GSE picture at Grumman looks quite gloomy.” He insisted that Grumman use some equipment that North American had developed for the command module. The situation had improved by the end of 1964, but much work was yet to be done over the next two years before the equipment could be considered satisfactory.⁴⁷

By mid-1964, both the lander and the command module were beginning to experience the weight growth that seems inevitable in spacecraft development programs. Von Braun promised Mueller in May that he would try to get an extra 2,000 kilograms of weight-lifting capability from the Saturn V, which eased some of the pressure on Gilruth’s team in Houston. Even so, the lander was getting dangerously fat, moving steadily toward its top limit of 13,300 kilograms. Most of the weight-reducing talent in Houston was busy with the command module, whose Block II configuration was not as well defined at the time as the lander’s. Several modifications in the landing vehicle were suggested, but any that limited either operational flexibility or reliability were resisted. Moreover, the lander was so unlike other spacecraft that projections were almost useless in estimating future weight increases. Containing this growth would be a major project during the coming year.⁴⁸

The years 1963 and 1964 had seen the lunar module move from the drawing boards to the manufacturing line. During 1965, hardware fabrication, assembly, and testing would begin. After that, it would take only a few steps to put the craft into space. These steps, though few after the spacecraft design had been “frozen,” would not be easy ones. There proved to be several more pitfalls to overcome. Some of these problems - difficulty with combustion in the ascent propulsion system, for example - were resolved only a short time before the mission that fulfilled Apollo’s goal of landing men on the moon.