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

by Kelly, Thomas J.

  Once during a simulation of Apollo 9 this carefully contrived virtual reality was shattered abruptly. During a quiet period in the mission, the LM (mission call-name Spider) failed to reacquire communications with the tracking station at Woomera, Australia, after passing over the Indian Ocean and the Australian outback. The ground team began analyzing the LM’s communications systems and could see no problem. Since Spider was flying formation at a distance with CM Gumdrop, CapCom asked Dave Scott, the CM pilot, to hail Spider using Gumdrop’s radio. Scott hesitated and delayed for several minutes in following through on the request. Then Comdr. Jim McDivitt’s voice came sheepishly over the net: “I’m sorry fellows; you caught us. Rusty and I sneaked out for a few minutes to get a sandwich; we thought we’d make it back before acquiring the next ground station. Sorry about that!”

  The simulations allowed us to practice our assigned roles and sharpen our real-time problem-solving skills. As the senior Grumman representative, my job in SPAN was to evaluate each LM problem as it arose and to marshall as much expert help in the solution as time allowed. I collected their inputs, made my own evaluation, and presented this to the NASA senior person in SPAN (usually Owen Maynard or Scott Simpkinson) as Grumman’s official recommendations. Any deviation from normal performance was called an anomaly, and each anomaly was written up on a discrepancy report form as it occurred and dispositioned when the explanation of the cause was agreed upon. This might occur weeks later, after extensive computer analyses or laboratory tests had been performed.

  The last mission simulation I took part in was about two weeks before scheduled launch. By then we were part of a finely honed mission support team, ready to assist the NASA flight director and his flight controllers in dealing with the unexpected as it came up in flight. We knew the astronauts’ lives would be depending upon LM for the first time in space.

  Apollo 9 was also LM’s first flight together with the command and service modules. It was an ambitious ten-day mission with the goal of performing in Earth orbit the entire sequence of events required on a lunar mission, except for the actual landing. It also provided the first and only flight test of the spacesuit and backpack to be used in exploring the Moon, during extravehicular activity, or spacewalks, from both the LM and the CM.

  You Have Twenty Minutes …

  Before it was even launched Apollo 9 provided one of the most difficult tests of my ability to assess conflicting data and provide NASA with a sound recommendation. About four hours before scheduled liftoff at KSC, an anomaly was noticed during prelaunch filling of the LM’s descent helium tank. The helium was used to pressurize the descent propellant tanks, causing the fuel and oxidizer to flow into the rocket engine without pumps. To minimize tank weight, the helium was stored at high pressure and extremely low temperature (nominally 1,540 pounds per square inch at minus four hundred degrees Fahrenheit), which put it in what thermodynamicists call the supercritical state, where it is as dense as a liquid and yet completely fills its container like a gas. This required a tank of advanced design with a highly effective vacuum jacket and insulation, made for us by the Airesearch Division of Garrett Corporation. When filling the tank the servicing crew weighed the amount of helium delivered and verified the tank’s thermal performance by reading the predicted combination of temperature and pressure from a thermodynamic chart.

  When Spider’s helium tank was loaded, the resulting temperature and pressure fell above the predicted curve, suggesting that excess heat was leaking into the helium. Since this could be caused by an insulation defect in the helium servicing cart or its vacuum-jacketed hoses as well as the flight tank, our first recommendation was to empty the flight tank and refill it, using a different servicing cart and hose set. I was at my post in the SPAN Room at the Mission Control Center in Houston discussing the problem with the NASA propulsion people there, as well as on the phone with our Grumman people at KSC. It took more than an hour to detank and refill, and although the result was somewhat improved over the previous, it was not quite within limits. We decided to wait half an hour and see what happened, reasoning that if there were an insulation defect in the tank it should continue to pick up excess heat and move farther from the allowable curve. Instead of moving away from allowable, the pressure/temperature combination drifted toward it and forty-five minutes later was within limits.

  It was the worst kind of anomaly; a discrepancy that corrects itself! Is it the hidden clue to a real problem with the hardware, or just a minor and meaningless aberration in a system so complex that all of its variations may never be understood? If the tank really had a heat leak we should scrub the launch and change it, because otherwise we could not be sure of conducting all the descent engine firings required by the mission. It would delay the launch by about five days.

  George Low, NASA’s Apollo spacecraft director, came into the SPAN Room himself to confer directly with me and Owen Maynard, NASA’s LM Engineering manager. This showed the gravity of the situation, since I had never seen him in SPAN before. Low was a careful, thorough engineer and a decisive manager; he had led the Apollo program’s technical rebuilding after the dark days of the Apollo 1 fire. After we briefed him on the helium anomaly, he fixed his steely blue eyes on me and asked for Grumman’s recommendation: Did we proceed with the launch or scrub and change the tank? He said I had twenty minutes to let him know.

  I spent most of the twenty minutes on the phone with people who could contribute to the decision, including Grumman’s corporate chief engineer, Grant Hedrick, who had a sixth sense for pinpointing the cause of technical problems, and the experts at Airesearch, who had searched all prior test records on that tank and found no anomalies. Then I went up to the VIP viewing area behind Mission Control and huddled in a corner of the dimly lit room with Joe Gavin. Summarizing the data and the tradeoffs, I recommended to Gavin that we launch. After a thoughtful pause, he concurred. I hurried back to the SPAN Room and repeated this recommendation to Owen Maynard. When he agreed as well, I called George Low and told him that Grumman officially recommended that NASA proceed with the launch, as we believed the tank to be sound. He asked what Gavin and Hedrick thought. I said they agreed, and Low then said NASA would proceed with the launch. He thanked me for meeting my deadline.

  NASA did not mention the issue for the remainder of the mission, but I monitored the helium tank’s performance almost continuously whenever it was accessible on the screen.

  A Great Flying Machine

  Nothing that followed in the actual flight put so much pressure on me personally, and for the most part the flight went well. It was our Grumman support team’s first direct experience with astronauts on a real mission, and I found it exciting to have men whom I knew up in space flying our machine. The giant three-stage Saturn 5 booster lifted off on schedule and performed flawlessly except for some Pogo longitudinal vibrations in the S2 stage, placing the spacecraft into exactly the planned Earth orbital altitude. The critical maneuvers of command and service module separation from the spacecraft/LM adapter, and rotation and docking to the LM, went perfectly. Upon command the LM was separated and pulled away by the CSM, while the S4B stage was jettisoned into a lower orbit and burnup in the Earth’s atmosphere. After six hours of checking out the CSM and its systems, McDivitt fired the service propulsion system (SPS), and the powerful rocket engine boosted the heavily laden CSM/LM combination into a higher orbit. He sounded relieved that the dormant LM was still there after the force of the first burn. Three additional SPS firings were successfully accomplished, increasing the crew’s confidence in the capability of their spacecraft. Following these operations the crew settled down for a meal and sleep; the first Apollo mission on which the three astronauts were allowed to sleep simultaneously. I took advantage of the quiet time and shift change to hand the SPAN duty over to Howard Wright.

  I was back to the SPAN Room early the next morning, listening to the crew puffing as they donned their spacesuits to enter the LM. The crew channel went dead. We did not l
earn until the postflight briefings that Schweickart had suddenly vomited.6 After some delay he entered the LM and flipped dozens of switches to activate its systems. He commented that the LM was quite noisy, particularly its environmental control system. McDivitt joined him, and after they unpacked the television camera in the LM cabin we watched them on worldwide TV. Our friend McDivitt promptly embarrassed us by pointing out to the world a washer and other bits of manufacturing debris floating through the cabin under zero gravity. It was a chastisement we deserved, and it motivated us to still more stringent efforts to clean the cabin and all closed compartments of the LM during assembly and test.

  McDivitt and Schweickart extended the LM’s landing gear, which locked smartly into place upon command. They checked out the LM’s systems and fired the LM descent engine for more than six minutes at full thrust while in the docked condition, simulating much of the powered descent burn that would be required to bring LM down from lunar orbit for landing. The crew controlled the engine manually and demonstrated digital autopilot attitude control.7 All LM operations went perfectly, including the performance of the suspect descent helium pressurization system. When McDivitt and Schweickart rejoined Dave Scott in the command module, they felt that their LM would be up to the challenges ahead.

  The fourth day in orbit all three astronauts donned their space suits and opened the hatches of both spacecraft. Schweickart and Scott performed spacewalks from the LM and the CM respectively; the former using his backpack8 for space life support, while Scott and McDivitt’s spacesuits were connected to their spacecraft by flexible umbilical hoses. Unknown to us, Schweickart was instructed to take the spacewalk a step at a time and end the exercise immediately if he felt nauseous. Plans for him to float freely in space on his tether were dropped, but he did use the handholds to climb up the front face of the LM ascent stage near the docked connection between Spider and Gumdrop, where he could clearly see Dave Scott standing in Gumdrop’s open hatchway. He and Scott photographed each other in their celestial perches like ordinary tourists. During the spacewalk Schweickart went by the call name Red Rover since he was a third spacecraft himself, communicating through the backpack’s radio. Although its duration was halved to one hour, the spacewalk and backpack demonstration were completely successful.

  The fifth day in orbit was the crucial part of the mission for the LM—the demonstration of LM’s flight maneuverability, and its ability to rendezvous in orbit from a far distance. My colleagues and I scrutinized the instrumentation readouts on our consoles carefully as the crew reactivated Spider’s systems. Hundreds of pressure, temperature, voltage, current and other measurements located in all the systems were sampled several times a second, giving us detailed real-time information on the LM’s health and performance. Any measurement that strayed out of preset normal limits triggered caution and warning alarms in the cockpit and on the ground consoles. With all systems activated, Spider looked good to the crew, to the flight controllers, and to me. Over the net came Flight Director Gene Kranz’s crisp voice: “Apollo 9, you’re ‘go’ for LM sep” (lunar module separation).

  No longer joined at the head to Gumdrop, Spider cavorted briefly, testing her reaction control system, and then pirouetted slowly before Gumdrop’s windows, preening for Dave Scott’s inspection. He pronounced her beautiful. After forty-five minutes of maneuvering within 3 miles of Gumdrop, McDivitt fired the descent engine, putting more distance between the two spacecraft. He encountered “chugging” of the descent engine between 10 and 20 percent thrust, which smoothed out completely above 40 percent. Subsequent firings increased the separation distance to over 110 miles, where the pilots could no longer see each other’s spacecraft. Spider’s crew then separated the ascent from the descent stage while igniting the ascent engine in an orbital simulation of lunar liftoff, and successfully completed orbital rendezvous with Gumdrop, which could be seen from over 45 miles away. Everything worked perfectly except Spider’s tracking light, which failed at stage separation. McDivitt and Schweickart spent more than six hours in Spider apart from Gumdrop, exercising all the systems. They proved in flight that the LM could leave the CSM, find its way back to it, and dock safely.9

  Spider performed so consistently well that I never felt any apprehension as I watched each critical event of the mission click off like clockwork. I could hardly believe that this agile machine, dancing so gracefully through space, was the same crotchety beast with the broken wires and structural cracks that had given us fits for over two years of ground testing. Was our LM design and construction really good after all, or were we just lucky? I was not sure, but thought it was some of both. I was glad we had practiced all those mission simulations. Although the anomalies during the real mission were benign compared with the deviltry concocted by NASA’s simulation director, we handled them effectively, and we had an excellent follow-up system in place to assure that the entire list was closed out before the next mission, which would be even more demanding.

  After they returned to Gumdrop, the crew closed Spider’s docking hatch for the last time and set her free. Ground control radioed a firing signal to Spider’s ascent engine to park her in a highly elliptical orbit, and the crew watched her depart until no longer visible. As in the coming lunar missions, Grumman’s job was done, even though the mission was still in progress and the crew in space. I joined the rest of my colleagues at the airport and flew back to New York, where four days later I watched Gumdrop’s triumphant splashdown in the Atlantic near Puerto Rico. An hour later the crew was safely aboard the USS Guadalcanal, grinning broadly and obviously glad to be home safely. At a jubilant ceremony in Washington, Vice President Spiro Agnew awarded NASA Service Medals to Gen. Rip Bolender and Lew Evans to acknowledge the lunar module’s successful initiation as a manned space flying machine. The way was now clear for the final rehearsal, a mission that would take the complete Apollo spacecraft around the Moon.10

  The best accolades to my ears came at the crew’s postmission debriefing in Houston. McDivitt and Schweickart were enthusiastic about the LM’s performance: “That LM is a great flying machine. And when it’s just the ascent stage alone, it’s very quick. It snaps to the controls like a fighter plane, or a sports car. It was super to fly!”

  LM Brushes the Moon: Apollo 10

  “Son of a bitch!”

  Apollo 10 lunar module pilot Gene Cernan’s startled exclamation over the network snapped me out of the lethargy in which I had been dreamily scanning the LM’s readouts on the console monitor. Suddenly there were wild excursions in pitch and yaw, and much reaction control system thruster firing. Snoopy seemed to be throwing a fit, thrashing about in space. Howard Wright was already on the phone with Jack Russell, Grumman’s stabilization and control subsystem engineer in nearby Building 45.

  “Ask them to be sure the AGS is still in attitude hold,” Russell advised immediately.

  Wright and I passed this word on to Scott Simpkinson, NASA’s senior person in the SPAN Room. We could hear the crew’s heavy breathing over their open mike on Snoopy as Comdr. Tom Stafford struggled to regain control. Stafford jettisoned the LM descent stage, the next planned activity before igniting the ascent engine to leave low lunar orbit and rendezvous with John Young in Charlie Brown. Despite Snoopy’s being over thirty degrees off its proper flight attitude, this critical function was executed flawlessly—explosive bolts and nuts, guillotine cutter, deadface connector fired in unison, and the descent stage fell away stably as Stafford ignited his RCS thrusters. The released ascent stage continued to thrash spasmodically about all three axes. A warning light said they were approaching gimbal lock of the guidance platform. Stafford took over manually and worked the attitude control switches, and Snoopy calmed down.

  Playing back the last few minutes’ data, the flight controllers determined that the AGS had been mistakenly switched to automatic mode while the crew was correcting a minor rate gyro disturbance, and Snoopy’s tantrum was the result. Although AGS was normally a backup to the pri
mary navigation and guidance system, to be used only if PNGS failed, for Apollo 10 it was being used to control the ascent into rendezvous orbit to demonstrate its capability in flight. In the automatic mode it searched for and locked onto the command and service modules, producing the unwanted attitude gyrations. Stafford’s switching had already restored it to the correct “attitude hold” setting.

  Over the net I heard CapCom Charlie Duke tell the crew they had corrected an improper AGS switch setting that had caused the disturbance, and everything on Snoopy looked good. They were cleared to fire the ascent engine for rendezvous orbit insertion. The whole unsettling episode had taken about three minutes.

  The ascent engine ignited smoothly, and Snoopy ascended from skimming the forbidding mountains of the Moon at fifty thousand feet as its orbital velocity increased. Soon we would know if the orbital mechanics, rendezvous procedures and communications that had been demonstrated in countless simulations, and in Earth orbit on Apollo 9, would work as well in close proximity to that enigmatic gray eminence, the Moon. Its uneven mass concentrations (mascons) could perturb the analysts’ orbital calculations; its lack of shielding atmosphere left Apollo’s radios open to solar radiation interference. These concerns, together with the desire for close reconnaissance of the tentatively selected first lunar landing site, were NASA’s basis for performing the Apollo 10 low-altitude lunar-orbit mission, instead of going directly for a landing.

  Stafford and Cernan enjoyed their smooth upward ride, and commented that they could see the Moon’s features receding away from them. At the end of the ascent burn, they were in the correct attitude and flight path for rendezvous, 48 miles from Charlie Brown, on whom they had radar lock and visual sighting. Snoopy’s crew had first seen Charlie Brown from 100 miles away, and Young had seen them through his sextant at 155 miles. Steady communications with each other and Houston kept both crews aware of what was happening. Rendezvous closure and docking were routinely successful, as Stafford skillfully positioned the skittish lightened Snoopy close to Charlie Brown, and Young firmly thrust his probe into Snoopy’s drogue and was rewarded by the reassuring snap of twelve capture latches engaging.11 Lunar-orbit rendezvous, so long debated and studied with apprehension, proved to be a “piece of cake.”