Moon Lander: How We Developed the Apollo Lunar Module (Smithsonian History of Aviation and Spaceflight)

by Kelly, Thomas J.

  Carbee and I led some memorable design meetings in structures and propulsion to review and revise the configuration. The Grumman LM design as proposed just barely fit the constraints imposed by NASA at its estimated weight of twenty-two thousand pounds. When this increased to twenty-five thousand pounds during negotiations, the carefully constructed proposal design house of cards collapsed and had to be reinvented.

  Arnold Whitaker was a brilliant systems engineer with an instinctive feel for how to solve problems through rigorous analyses that yielded pragmatic results. He was short, slim, and bald with flashing blue eyes and a sharply pointed nose, and his probing intellect was evident in his persuasive, logical approach to problem solving. Arnold had a disconcerting habit of thinking carefully and deliberately before he answered a question or presented an opinion.

  Whitaker and Carbee were the same age, four years older than I. Whitaker served in the navy during World War II but never saw combat. An aeronautical engineer from MIT, he obtained a master’s in applied mathematics from Adelphi University. After a brief stint at Chance Vought Aircraft, Whitaker came to Grumman when Vought moved from Connecticut to Texas. He became group leader of Research Control Analysis, developing techniques for analyses of aircraft control systems. Swept-wing supersonic aircraft with hydraulically powered ailerons, elevators, flaps, and rudders demanded increased sophistication in mathematical analysis and computer simulation, and Whitaker was at the forefront of Grumman’s efforts in this area. He worked on problems that commanded attention throughout the industry, such as control surface flutter, control reversal at transonic speeds, and powered control system instabilities.

  Whitaker was a forceful leader as well as a gifted analyst. He drove studies and analyses hard until they produced answers designers could use in practical systems. He became project engineer for the Eagle missile, an advanced two-stage solid-rocket air-to-air weapon won by Grumman in a navy competition. I remember him analyzing a missile separation dynamics problem at the blackboard with his project group leaders, filling the board with sketches and equations, striding back and forth, head down, thinking, stabbing the air with his cigarette when offering a new approach: an impressive performance that left each participant with an assignment and due date for his part of the puzzle.

  The navy canceled the Eagle project just before it was to enter flight testing due to a change in strategic planning direction. It decided to have a compatible series of airborne missiles to perform various missions, launched from a subsonic missile-carrying airplane called the Missileer. Whitaker went from leading a real project to heading Grumman’s proposal team for its successor, the TFX(N) missile.

  Whitaker played a major role in all aspects of LM Engineering because the analytical sections, including Systems Analysis and Integration, Structural Analysis, Thermal, Weights, Dynamic Analysis, Electronics, and Systems Test, were under his leadership. He and Bob Carbee contributed to many of the most creative sessions during which the LM design took shape and matured.

  The LM proposal configuration had cylindrical structure in the ascent and descent stages and a five-legged fixed landing gear. The descent-stage structure housed the descent propulsion system (DPS), the main elements of which were the descent rocket engine, six spherical propellant tanks (three fuel and three oxidizer), three spherical helium pressurant tanks, and associated plumbing. The descent stage also supported the landing gear, provided compartments to carry scientific equipment to be deployed on the Moon, and carried the landing radar. The ascent-stage structure contained the ascent propulsion system, consisting of ascent rocket engine, four spherical propellant tanks (two each fuel and oxidizer), a spherical helium pressurant tank and associated plumbing, and the reaction control system, which had sixteen small rocket thrusters mounted in clusters of four, plus spherical fuel, oxidizer, and pressurant tanks. The ascent stage also housed the crew compartment with its seats, instrument panels, controls, and helicopter bubble windows; the electrical power system (EPS), with its fuel cells, hydrogen and oxygen tanks, busses, and switchgear; the environmental control system, with its water and oxygen tanks, fans, and plumbing; and all the electronic systems, including GNC, communications (with steer able and omnidirectional antennas), instrumentation, and rendezvous radar. The ascent stage had two docking hatches into the crew compartment, one upper (or overhead) and one forward. All this was squeezed into a package so tight that the least change of ground rules or assumptions would cause it to burst.

  At twenty-five thousand pounds the increased diameter of the descent propellant tanks meant that a five-legged fixed landing gear that would provide LM with adequate tipover stability would no longer fit within the spacecraft/LM adapter. We had to accept the added complication of a folding, extendable landing gear, but we gained an increase in the allowable diameter of the LM descent-stage structure.

  The proposed descent-stage structural arrangement was inefficient for carrying the loads between the LM and the SLA during boost. We needed deep structural beams spanning the full distance across the SLA. This suggested that the basic descent-stage structural arrangement should be a cruciform rather than a cylinder. A cruciform structure in turn demanded four propellant tanks rather than six, and to contain sufficient propellant they must be cylindrical rather than spherical. The cruciform shape also provided a natural place on each end of the cross to mount a rugged tubular framework with which to attach the LM to the SLA and to mount the folding landing gear. The empty areas between the cross were covered diagonally with light structure to serve as storage bays for equipment that could be left on the Moon, or which would be used by the astronauts during lunar exploration. In a relatively short time the redesign of the LM descent stage was basically completed.

  The redesign of the ascent stage took much longer because it involved a greater number of complex interrelated factors, including the ascent propulsion system (APS), the crew compartment, and the electronics packaging approach. The optimum solution to these variables was established by performance, reliability, and weight considerations rather than geometry, which had dominated the descent-stage redesign.

  One of the most basic variables in the ascent-stage configuration, the number of propellant tanks in the APS, was not seriously challenged until mid-1963 because I and my associates were so ingrained with the principle of symmetry in design. The proposed four-tank APS was conventionally symmetrical. As we saw more clearly how critical the APS was to survival of the crew we reexamined the design to maximize its reliability. We decided to emphasize design simplicity: because the APS itself and most of its major components could not be made redundant due to weight restrictions, APS reliability would be assured by making it as simple as possible. Simplicity, coupled with ample safety factors and extensive ground testing, would be the key.

  Our proposed design already had gone a long way toward simplicity. It was helium pressure-fed (no pumps); hypergolic (no igniters); ablatively cooled (no intricate liquid-cooling passages), and operated at a single fixed level of thrust (simple controls). Minimizing the amount of plumbing, components, and joints was also a virtue in APS design because it reduced the chance of leakage, which, due to the APS’s small margins on the amount of propellant required to achieve lunar orbit from the Moon’s surface, could be disastrous.

  Manning Dandridge, LM Propulsion section head, urged me to consider other tank options, including a two-tank version. Dandridge was a tall, ruddy-faced man who possessed an engaging, confident manner and a breezy, informal style of presenting arguments that only hinted at the careful analysis and judgment behind his approach to engineering. A mechanical engineering graduate of Stevens Institute, he designed and tested fuel systems controls and precision components at Aerotec Corporation and a ramjet engine control system for the Navajo missile at Wright Aeronautical Division. Since joining Grumman he had led the propulsion design work on our missile and space proposals, including the successful bids for OAO and the Eagle missile. He spent a year at Aerojet General as Grumma
n’s resident for Eagle rocket-engine development. Outside of work he was a Life Master bridge player, also skillful at cribbage, and he turned his expertise at mathematically based games to good use in solving engineering problems.

  Carbee and I held several meetings with our propulsion and structures engineers in which we considered both the four-tank design, either spherical or cylindrical, and the two-tank design, all tanks to be constructed of aluminum or titanium. The two-tank design gave the LM ascent stage an unusual asymmetric appearance because, to maintain the center of gravity on the center-line of the rocket engine, the distance between the fuel and oxidizer tanks had to be in inverse ratio to their weights when fully loaded. (This ratio is the product of the oxidizer/fuel [O/F] density ratio times the O/F ratio required for proper combustion.) The result, Dandridge admitted, looked like LM had the mumps—on one side only.

  Appearance notwithstanding, the two-tank design had simplicity on its side and offered a savings in weight, even greater in titanium. I decided to adopt the titanium two-tank configuration, and we made the case for it to NASA.

  NASA was cautious about making such a highly visible change and insisted that we do further analyses of off-design conditions. A symmetrical four-tank design remained properly balanced even if the system operated somewhat off the nominal O/F ratio due to differences in tank pressures, line pressure drops, or other factors, but the two-tank design did not. Operation off the nominal O/F caused an unbalanced moment that had to be offset by the RCS thrusters to maintain control of the LM’s attitude and trajectory. We analyzed many off-nominal cases to assure ourselves that control could be maintained at all times and that the extra RCS propellant expended would not negate the weight reduction of the two-tank design. Betting that the outcome of these analyses would be positive, I ordered the two-tank design into our M-1 mockup for the review with NASA in September 1963 but did not obtain NASA’s approval until December.

  Innovative Electronics Packaging

  Another exciting area for engineering innovation was electronics packaging. Led by Bob Carbee, instrumentation subsystem engineer Ben Gaylo, GNC subsystem engineer Jack Russell, and RCA Engineering manager Frank Gardiner, we made early decisions that shaped LM’s electronics. LM was being designed at a time of pivotal changes in the electronics industry. Circuits constructed of individual transistors were just starting to be replaced by solid-state integrated circuits (ICs) etched on silicon, the beginnings of “circuits on a chip” technology. The new ICs were being developed rapidly for commercial and military use, but they had not accumulated enough operational time to have a solid statistical reliability record. All indications were that the reliability of ICs would be superior to transistor circuits, in addition to saving weight, volume, electrical power, and cooling. I felt we should “bet on the come” and use ICs extensively. After consulting with George Weisinger, our reliability systems engineer, our major subcontractor RCA, and NASA, we established a parts policy that allowed us to use ICs that were just entering the stringent mil-spec parts-qualification test program, as well as those already qualified. If a part subsequently failed the test program, we would have to find a substitute. In practice this rarely happened, and our progressive policy helped keep LM’s electronics abreast of the rapidly advancing state-of-the-art until the detailed designs were frozen in 1965.

  I asked Ben Gaylo and Jack Russell to develop an Electronics Packaging Specification for LM. This “spec” would establish the shape, geometry, cooling method, connectors, parts mounting, and other standards for the “black boxes” into which most LM electronics would be packaged. With helpful advice from RCA, they developed an advanced packaging design that met my requirements for simplicity, ruggedness, and maintainability. There were two main features of this design. First, wire-wrap mounting of circuits and components onto a mother board. This design, pioneered by the Sippican Company, used a special wire-wrapping tool to wrap together hardened pins from the circuits that protruded beneath the motherboard. The connections remained secure in high-vibration environments, as shown by many convincing tests, yet they could be unwrapped when required for maintenance replacement of the circuits.

  The second main feature was circuit cooling by thermal conduction, using copper strips etched into the circuits and motherboards to transfer heat from the heat-producing ICs or transistors to aluminum rails at the midpoint of the sides of the box, upon which the motherboards were mounted. The rails on the boxes in turn were bolted to extruded aluminum mounting rails on the LM’s supporting structure through which an ethylene glycol-water coolant mixture was pumped by the environmental control system. The space program pioneered the conduction cooling of electronics. For ground-based and aircraft applications the standard technique was air cooling, using fans or blowers, but in the vacuum of space, air cooling was impractical, and it was undesirable for the low pressure of the LM cabin.

  We pushed hard to prepare and issue the electronics packaging specification, which was a constraint to finalizing all our electronics procurements. When NASA approval was received, we included the spec in our procurement packages and used it in redesigning our electronics installations. In the ascent stage, we designed an electronics rack mounted outside of and behind the pressurized crew compartment. The rack contained parallel rows of cold rails to which the electronics boxes were bolted in vertical columns. Plumbing connections from the ECS provided water-glycol coolant to the cold rails. The area between the equipment rack assembly and the rear wall of the crew compartment was used for mounting water and oxygen tanks and other equipment; the entire area was known as the aft equipment bay.

  I enjoyed developing the LM electronic packaging design as we did it, and later I appreciated it even more. The LM electronics that was packaged according to our spec was among the most reliable, trouble-free elements of the program. Other electronics items that were not packaged in this manner, such as the development flight instrumentation (used only on the first three flight LMs) and the cockpit displays inside the cabin were harder to maintain and had higher failure rates.

  Crew Systems and the LM Cockpit

  The proposal LM had a small, pressurized interior resembling a helicopter’s crew station with aircraft seats and instrument panels, but with hand controllers replacing the conventional aircraft stick and rudder pedals. We also had forward and upper hatches, each with docking capability for redundancy. During the LM contract negotiations NASA provided us with preliminary information on the spacesuit and backpack designs and of their estimated requirements for crew equipment stowage and lunar-sample return provisions in the cabin. With these additional requirements the cabin volume we had provided was insufficient. A complete redesign of the LM crew compartment was required.

  John Rigsby and Gene Harms, the section head and deputy of Crew Provisions, and Howard Sherman, Human Factors section head, realized that Grumman had much to learn from NASA, McDonnell Douglas, and NAA about crew equipment, human factors, and medical considerations for manned spaceflight. They spent most of the first three months after LM go-ahead with these organizations, learning the accumulated American experience with humans in space. They tried on spacesuits and learned firsthand how difficult seemingly simple tasks became in a stiff, pressurized suit. They discussed the effects of zero gravity and the problems of extravehicular activities (EVA) with astronauts and crew systems engineers from Projects Mercury and Gemini. They saw the Gemini backpacks and working models of the prototype for Apollo. They discussed problems of lighting and equipment stowage, as well as eating, personal hygiene, rest, sleep, and other practical human considerations as affected by the space environment. They learned about different mockup construction techniques and schemes for simulating zero gravity in water tanks or with suspension devices. NASA scientists gave them their latest opinions about the nature of the lunar surface and the problems men would encounter exploring it. With this knowledge and their prior experience with crew systems for airplanes, they returned to Bethpage to redesign the L
M.

  Bill Rathke, Bob Carbee, and I had several creative sessions with them discussing the conflicting requirements in the design of the crew compartment. I was concerned about the large amount of window-glass area in our proposal design. Glass was very heavy, it was dubious as a structural material, and the large windows allowed too much heat transfer into and out of the cabin. Weight would be a problem because the revised cabin volume was more than twice that provided in the proposal. We agreed that the crew’s vision angles to the landing site that we had provided in the proposal must be maintained or increased.

  The solution came in a flash of insight as Rigsby, Harms, and Sherman were rehashing the cabin design around a drawing board that showed a three-view (drawing of an area viewed from three directions) of the interior: “What if we get rid of the seats?”

  It was a brilliant, paradigm-shattering question that led to a frenzied afternoon of sketching, reasoning, and debating among the three. The next morning they were in my office, clutching several rolled-up vellum drawings and excitedly announcing that they had a new crew compartment design to show. Rathke and Carbee joined me in the conference room as they proudly unrolled the drawings showing two astronauts inside a redesigned LM crew compartment without seats, in the positions they would take to perform every major activity required inside the LM: flying the spacecraft, docking the upper hatch to the CM, doffing and donning spacesuits and backpacks, resting while on the Moon and entering and exiting the LM via the upper and forward hatches. After briefly perusing the sketches we agreed it appeared to be a superior approach.