8
Trimming Pounds and Ounces
Even before Grumman was placed under contract the weight of the LM had grown, and as we began the actual design of the craft, its weight continued to increase at an alarming rate. Our LM proposal design was estimated at 22,000 pounds, but in discussions with NASA at the contract negotiations it was agreed to increase LM’s weight to 25,000 pounds. The target weight, which was used for propellant-tank sizing, was increased to 29,500 pounds in February 1964. This was the basis for the tank sizes in the TM-1 and M-5 mockups. The estimated weight of the design soon exceeded the target, and we again faced the need to increase the target weight and the size of LM’s propellant tanks.
Because of the flight mechanics and rocket propulsion constants of the lunar mission, LM was extremely sensitive to weight growth. For every pound in the ascent stage, three pounds of propellant were required to land it on the Moon’s surface from lunar orbit, lift it back into lunar orbit, and rendezvous with the CSM. The weight growth factor was, therefore, 4 to 1: every pound of added ascent-stage inert weight increased the Earth launch weight of the LM by four pounds. Weight added to the descent stage but left on the Moon had a growth factor of 2.25 to 1. This was far greater than for aircraft, even for military missions with fighters and attack bombers, where growth factors seldom exceeded 15 percent, or 1.15 to 1.
The major factors that drove LM weight up during 1963 and 1964 were reliability requirements, mission operational requirements, and configuration definition. The reliability approach that we had adopted relied upon functional or component redundancy and ample safety factors where redundancy was not possible. This was a sound design philosophy, but it added weight compared to our proposal design, which had little redundancy.
Operational requirements became visible to us through the work of the Apollo Mission Planning Task Force. Their design reference mission provided the basis for estimating many operational requirements that affected LM weight: the duty cycle (on/off times) of LM equipment requiring electrical power, the number of cabin pressurizations with oxygen, the spacecraft attitude relative to the Sun and hence heat loads and cooling water capacity, the amount of consumables (water, oxygen, power, propellants) required on LM, and so on. There was a wealth of data in the DRM that had an effect on weight. As we dug into this, our understanding of mission operations requirements improved and the weight grew.
Configuration definition changed swiftly in 1963, when the ascent stage’s geometry was completely revised in the M-1 mockup. A year later the M-5 mockup essentially completed the definition of LM’s basic configuration. The weight estimates were revised based on this geometry and equipment arrangement, and invariably they were greater than before.
An empirical explanation for the large early weight growth on LM lies in the historical data accumulated on many aircraft and spacecraft programs, which shows that estimated weight, from design sketches and system descriptions, is typically 20 to 25 percent lower than weight in the final product, mainly due to the omission of many components and design details that are unknown to the estimators. Calculated weight, from engineering drawings, is usually 5 to 10 percent low because many small details, such as the exact number of fasteners, cut-outs, and installed components, are only approximated in the calculations. Not until actual weight is obtained by putting the various parts and assemblies on a scale is the bias toward low estimates corrected. As the LM design moved from sketches to drawings to fabricated parts, the weight increased with the increasing percentage of calculated and actual weights versus estimated.
No matter how we rationalized it, the inexorable growth in LM weight was threatening the whole Apollo mission. Caldwell Johnson was concerned that LM might become too heavy to do its job. Johnson’s opinion was not to be taken lightly. He had designed the Space Task Group’s own versions of the LM and constantly reviewed and critiqued Grumman’s emerging design, looking for improvements, simplifications, and weight savings. I obtained sound guidance from him, Maynard, and Faget in evaluating LM design choices.
In mid-October 1964 Carbee, Whitaker, and I had a lengthy meeting with Owen Maynard, Caldwell Johnson, and William A. Lee of the Apollo Operations Planning Division in NASA-Houston to review the LM’s weight status and tank resizing options. The NASA people had done their own analyses of LM’s weight growth history and status and found the outlook bleak. They projected 12 to 15 percent growth in LM’s weight from its then-estimated 30,200 pounds, which would exceed the maximum allocation for LM based upon a Saturn 5 boost capability of 32,000 pounds. They suggested possible LM design changes to reduce weight and modifications to the mission rules and trajectories that might allow for an increase in the fraction of the Saturn 5 payload weight allocated to LM.1 We had a month to study these suggestions and add to them, and to recommend a target weight for LM resizing and an improved weight-control approach.
We responded that LM should be resized to a target weight of 32,000 pounds at Earth launch, of which 10,800 pounds would be in the ascent stage. This fully utilized the allocated Saturn 5 payload for LM. We also recommended a number of weight-reduction items, including the use of supercritical helium instead of gaseous helium to pressurize the descent propellant tanks, reduction of LM hover time at landing from two minutes to one, and possible use of batteries instead of fuel cells for electrical power. Batteries might not save weight, but they promised great simplification of the electrical power system and increased reliability. We urged further study of this option. Also recommended for study was replacement of the LM rendezvous radar with star trackers or lasers. Mission changes that we suggested NASA consider included use of a non-free return trajectory to the Moon and reduction of the CSM’s midcourse correction propellant allowance, both of which would allow more of the Saturn 5 payload to go to LM. After extensive discussions, these recommendations were approved by Maynard, Lee, and Johnson, then, the next day, by Joe Shea. Shea made clear to me that this was all the weight allowance they had to give LM, and that I had better get control of the weight growth or the whole program would be in deep trouble.2
Rathke and I preached reduction and control at all our Engineering meetings and set up a system to identify and evaluate potential weight-reduction items. We devoted more time at the daily technical staff meetings to reviewing and deciding whether to incorporate these changes into the design, subject to approval by NASA and the Change Control Board. Still, the monthly LM weight-status report showed increases, and Grumman and NASA managements grew restive. Some of the largest increases were in equipment supplied by our subcontractors and vendors. In March 1965 Bill Rector met with me and strongly urged Grumman to set up an intensive weight-reduction effort, with emphasis on subcontractors and with continuous involvement with and direction by Grumman’s LM program office.3
Outside events overcame my single-minded dedication to LM: on 27 March 1965 Joan gave birth to a beautiful eight-pound baby boy. He was our fifth son and sixth child, and we named him Peter. I took two days off to mind our other children and take Joan and Peter home from the hospital. Not without pangs of conscience, because I was in the middle of finalizing the design of the newly adopted battery EPS and also the accelerating weight-reduction and drawing-schedule efforts. I soon disappeared into the vortex of work but I took my turn with middle-of-the-night baby feedings.
Shortly after Rector’s request for more vigorous and effective weight reduction by Grumman, Arnold Whitaker presented a plan for tightening control on subcontractor and vendor equipment weights. He recommended that we include the allowable weight-change limit as part of every change request that Grumman issued to its suppliers and require the subcontractor to estimate the weight impact as part of his change proposal. This would be followed up by specific accounting of the actual weight impact of the change. The subcontractor’s weight-control performance would be a factor considered in determining his incentive fee. Gavin and Mullaney promptly implemented Whitaker’s suggestion.4 However, as the reports came in from
this tightened control over the subcontractors, it was clear that LM again was increasing in weight. By July 1965 we were reporting LM as exceeding the control weight of thirty-two thousand pounds.
Rathke and I decided that an all-out “crash” effort was needed to reduce LM weight and to control further growth, despite the disruption this might cause to our equally critical drive to meet drawing-release schedules. We initiated a “scrape” program, led by Sal Salina, head of the Weight Control Section, Len Paulsrud, head of Vehicle Design, and Dick Hilderman, head of Structural Analysis, to whittle weight from the LM structure. By making more extensive use of chemical milling, precision machining, and material substitutions, we could obtain significant weight reductions without major changes in the structural arrangement of LM. Unfortunately, this was not enough for the twenty-five hundred to three thousand pounds of reduction we needed as a margin against further growth.
From Grant Hedrick we learned of a successful approach Grumman recently used on the Fin aircraft program that we could apply to LM. Grumman was a major subcontractor to General Dynamics, supplying the aft fuselage and tail assemblies for all Fins and the design and final assembly of the navy version of this joint air force-navy airplane. General Dynamics-Grumman had won the Fin contract within weeks after the LM award. Despite a very extensive pre-award preliminary design and mockup, as the detailed design proceeded, both the air force and navy versions outgrew their weight limits and engine capabilities. To save the Fin program, which was threatened with cancellation if its weight and performance goals were not met, General Dynamics-Grumman mounted a drastic response, called the Super Weight Improvement Program, or SWIP. Even though most of the airplane and system designs had been completed and released to manufacturing, an independent team of SWIP design experts was turned loose on the program to second guess every aspect of the design for weight savings. The SWIP team worked cooperatively with project Engineering, but they had a direct reporting channel to the Fin program manager to assure that none of their recommendations would be stifled. A dollars-per-pound criterion was established to set a threshold for accepting weight reductions; for the Fin it was five hundred dollars per pound. SWIP was successful on the Fin: weight growth stopped and there was some net reduction in weight. The cost of the changes was high but affordable, and the resulting impacts on the schedule could be endured. SWIP saved the Fin program.
Early in 1966 Roy Grumman retired as chairman of the board and was replaced by Clint Towl. Llewllyn J. “Lew” Evans was then appointed president of Grumman Aerospace Corporation. Lew was a lawyer who came to Grumman as chief counsel after working in the Legal Department. He had a flair for marketing and deal making and became vice president of Business Development, strengthening and expanding Grumman’s close working relations with the navy. A charismatic and inspiring leader, Lew lost no time in getting acquainted with Grumman’s newest major customer, NASA. He held internal reviews on the status of the LM program and was alarmed by the array of problems he found involving our technical, schedule, and cost performance. To provide direct corporate-level oversight of the program, he established the Executive and Technical Review Board (ETRB), chaired by Senior Vice President George Titterton. At its first meeting the ETRB urged Gavin to conduct a SWIP on LM and offered to make available the SWIP review team, which was then finishing its work on the F111.
In July 1965 I was put in charge of the LM SWIP activity. The Grumman SWIP review team, twelve engineers headed by Ed Tobin and his deputy, Paul Wiedenhaefer, would report to me for the duration of the exercise. They had unlimited access throughout the program and a direct reporting channel to Grant Hedrick. NASA management liked the plan; they were pleased that Grumman was taking forceful action to control weight. Bill Lee was assigned by Joe Shea to be my counterpart as co-chair of the SWIP team.
Anyone could make weight-reduction suggestions to the SWIP team; opinions were actively sought through the employee suggestion program. Grumman was responsible for evaluating SWIP items and making recommendations to NASA for approval. We held intensive weekly reviews with NASA, usually at Bethpage, which Bill Lee always attended, often accompanied by Maynard, Johnson, and other engineers. Internally the SWIP team met several times a week with LM Engineering and Manufacturing management, reviewing hundreds of SWIP weight-reduction items and suggestions and thousands of Grumman and subcontractor drawings. Each item was evaluated and dispositioned against the criterion, which after considerable study and discussion had been set at ten thousand dollars per pound for “round-trip” items. Not every SWIP item could make it in time for LM-1; some of the more difficult items were phased in at LM-3, LM-4, or even LM-5. Even so the disruption to the schedule was severe and required constant replanning and revision of the PERT networks and schedules as we fought to maintain forward motion on the program while accommodating some major design changes to save weight.
Tobin and Wiedenhaefer were thorough, persistent, and innovative in finding areas to reduce weight. They were curious about how every system worked on LM and how the critical loads, safety factors, and materials choices were arrived at, so they led us through a complete review and justification of the LM design criteria and choices. Some LM engineers bristled at being second-guessed on their designs, but Ed and Paul were so logical in their questions and approach that no one could take offense. At their suggestion, we began holding SWIP meetings at our major subcontractors’ facilities, too, as more than half our SWIP items were in subcontractors’ equipment.
Although the SWIP team worked full time implementing the weight-reduction effort, I supplied the technical leadership and made the decisions for Grumman. Joe Shea, Joe Gavin, and I kicked off the SWIP effort at a motivational meeting with a large audience in the Plant 25 main conference room, making clear that the future of the Apollo program could very well depend upon the success of our efforts. We introduced Tobin and Wiedenhaefer to the LM people, and they told briefly what they had done on the Fin program and how successful it had been. They were confident of being able to do it again for LM. I made clear that all LM Engineering managers would be heavily involved and would be accountable for delivering their portion of the required weight savings. In the active question and answer session that followed, our engineers clearly showed their desire to support this program but expressed reasonable concerns and asked for direction in how to balance conflicting priorities in their workloads. We promised to expand the SWIP guidelines beyond the dollars-per-pound criterion to include schedule and reliability criteria also.
My daily technical staff meetings and weekly SWIP meeting were the management focus of the effort. The SWIP meeting usually took three or four hours. Tobin and Wiedenhaefer went through the list of potential SWIP items, noting new additions and reviewing the status of each. Whenever an item was ready for decision, whether to implement, modify, or delete, the SWIP team would be joined by the “cognizant” LM subsystem engineer (called the “cog engineer”) or section head.5 Bob Carbee attended all SWIP meetings, contributing to the decisions and following up in implementing them in his subsystem design groups. At the technical staff meetings the SWIP team usually was not present, but we examined in detail the design issues involved in one or two SWIP items in the subsystem under discussion.
We started the SWIP effort by establishing target weights for each subsystem, broken down into all its parts and components. These were arrived at in reviews with the cognizant subsystem engineer, his Engineering section head, and representatives from his principal subcontractor and suppliers. Drawings and specifications for the subsystem and its major components were reexamined and avenues of possible weight reduction identified. Because most of the designs had been released to manufacturing and many of the system components had been through some amount of development testing, we also discussed the potential impact on schedule and cost of these changes. The outcome of these meetings was that the cog engineer, his section head, and his subcontractor accepted a weight-reduction “target to beat” for SWIP.
>
I concentrated on the “big-ticket” items that had major potential for weight reduction, although no item was too small to escape the fine net trolled by the SWIP team (items down to .1 pound were considered). Structure headed the big-ticket items because there was so much of it. Redesign, scrape, and materials substitutions were possible in virtually all parts of the LM structure, and thanks to the inspired and diligent efforts of the Vehicle Design Section we saved every possible ounce. The Materials Section played a key role in identifying substitute lightweight materials and in perfecting the chem-milling process in Manufacturing, which was a major technique for reducing the weight of structural parts.
Implementing supercritical helium pressurization of the descent propellant tanks was another big weight saver. Feasibility of this item depended upon the design ingenuity of the LM Propulsion Section and their cryogenic tank subcontractor, Airesearch Division of Garrett Corporation, and by the Fluids GSE Section and their cryogenic tank and component supplier Beech Aircraft. Neither I nor NASA were willing to approve the supercritical helium change until we saw a design for the GSE that convinced us that it would be practical to load and unload this hard-to-handle material on the Apollo launch pad at Kennedy Space Center.
I also gave major attention to the substitution of batteries for fuel cells in the electrical power system. This change was made to improve reliability because the use of batteries eliminated a very complex system of hydrogen and oxygen tanks, plumbing, and components—as well as the fuel cells themselves. From a SWIP standpoint, I had to assure that minimum weight increase resulted. We had to be very sure of what we were doing, both in verifying the battery suppliers claims and accurately assessing the weight of the fuel cell system. Our fuel cell subcontractor, Pratt and Whitney, was well along in building and testing development units both for LM and the CSM, which used a similar system. Grumman held a competition between battery suppliers Eagle Picher and Yardney, and obtained electrical test results and actual weights from both of them. On 26 February 1965 Shea approved the change to Eagle Picher batteries for LM.6
Moon Lander: How We Developed the Apollo Lunar Module (Smithsonian History of Aviation and Spaceflight) Page 17