Apollo: The Race to the Moon (61 page)

At this point, Liebergot was assuming that the reading for O2 Tank 2 pressure was another instrumentation problem caused by the same electrical difficulties that had produced the appearance of low fuel cell pressure. The tanks, metal spheres twenty-six inches in diameter, contained nothing but liquid oxygen and a few simple and reliable pieces of equipment—a heating element, two fans, and a few sensors. Nothing much could happen to them. And even if something had gone wrong with O2 Tank 2, that still left O2 Tank 1, which would automatically feed all three fuel cells and maintain full power even if Tank 2 quit working. If Liebergot wanted to treat the oxygen tank as the cause, he had to hypothesize a failure that (a) had disabled Tank 2, (b) left Tank 1 still functioning, but nonetheless (c) somehow disabled two fuel cells. That didn’t make sense. The spacecraft wasn’t built that way. It was doubtful that such a failure could even be simulated—what could a SimSup use as a cause? The alternative and much more plausible explanation was that Liebergot was seeing a problem in the electrical system. Perhaps it was a pure instrumentation phenomenon, and the readings were lying to him. Perhaps the problem was producing interference in the power flows, in which case they would find some way to bypass the errant circuitry.

Liebergot’s instrumentation theory was instantly reinforced. The Main Bus B undervolt warning light on his panel blinked off. Swigert reported the good news from Odyssey: The voltage was looking good. Still, he added, “We had a pretty large bang associated with the caution and warning there.” Shortly thereafter, Lovell reported that the jolt must have “rocked the sensor” for the O2 quantity in Tank 2. It had been down around 20 percent, but now was “full-scale high” again.

And so it went for the first few minutes—conflicting data, a stream of seemingly unrelated problems. Around the MOCR and in the back rooms, controllers tried to put the pieces together, and they just didn’t fit.

5 minutes. The difference between EECOM’s situation and everyone else’s was that the other guys’ problems were fairly minor. EECOM’s problems were either trivial (if it was bad instrumentation) or life threatening (if the electrical and life-support systems of the C.S.M. were going dead). Like Bales on Eleven and Aaron on Twelve, Liebergot was the man sitting out in the open, the man with the best opportunity to make the mistake that would, in another MOCR euphemism, “blow the whole mission.” The difference between Liebergot’s situation and Bales’s in Eleven or Aaron’s in Twelve was that Bales and Aaron had had the option of stopping everything immediately and coming home—Armstrong could have called off the landing, Conrad could have fired the escape tower. Liebergot was stuck. If it wasn’t an instrumentation problem, this crisis had no easy way out. There might be no way out at all.

Kranz inquired of Liebergot how things were going. Liebergot replied that the crew were “flipping their fuel cells around”—trying to reconfigure them. Kranz was impatient: “Well, let’s get some recommendation here, Sy, if you got any better ideas.”

Just then Swigert reported the inexplicable and disconcerting news that they were getting an undervolt on Main Bus A, the one that until now had still been working. “Sy,” Kranz said, the exasperation in his voice now unmistakable, “what do you want to do? Hold your own and—” Kranz broke off to listen to the spacecraft, which was adding that Main B was “reading zip,” distributing no power at all. Still nothing from Liebergot. “Sy,” Kranz said, “have you got a sick-sensor type problem there or what?”

Liebergot had been busy talking to his back room. He now came onto Flight’s loop and recommended that the crew try to reconnect the fuel cells that, Liebergot hoped, had been thrown off line by the mysterious jolt. CapCom passed the instruction up to the crew while Liebergot returned to the EECOM loop that connected him to his back room. Remembering that O2 pressure reading, Liebergot called to Sheaks. “Larry, you don’t believe that O2 Tank 1 [sic] pressure, do you?” Sheaks came back confidently. “No, no. Surge tank’s good. Manifold’s good, E.C.S. is good.” Reassured, Liebergot went back to the screens and to a discussion of the electrical problem with Dick Brown.

Odyssey came back on the line. “Okay, Houston, I tried to reset and Fuel Cell 1 and 3 are both showing gray flags. But they are both showing zip on the flows.”

“I copy, Flight,” Liebergot said unhappily, indicating that he had heard the crew’s transmission. This had to be an instrumentation problem. They couldn’t possibly be having a simultaneous failure in two independent fuel cells—especially since, he had just been reassured, the oxygen was doing fine. And yet every time he got a new reading on the fuel cells, their condition seemed to be deteriorating.

“Okay, what do you want to do?” Kranz asked.

Liebergot suppressed an impulse to reply, “I want to go home.” He passed along a new configuration of linkups between fuel cells and buses that would reveal this problem for what he hoped it was: some blown circuitry that they could work around.

John Aaron was at home, shaving off the day’s stubble after spending a long shift in the Control Center. Arnie Aldrich, chief of the C.S.M. Systems Branch, called on the phone from SPAN. “John, there’s something funny going on here,” he said to Aaron. “A lot of the guys think it’s instrumentation problems and flaky readouts.” Aldrich had a long cord on his phone, so Aaron asked him to walk along the rows of consoles and read him some data. Aaron would later realize that being away from the MOCR, not glued to his own console, made it easier for him to see. In fact, standing in the quiet of his bedroom with shaving cream on his face, listening to the numbers that Aldrich reported, it was all quite clear. “Arnie, that’s not an instrumentation problem,” he said to Aldrich. “It’s a real problem.”

Over in Building 45, Sid Jones, one of the MER’s shift leaders, had been presiding over a half-empty MER when Swigert reported that Apollo 13 had a problem. It took him only a few minutes to decide that they were going to need all the help they could get. He called Don Arabian, who had left just an hour earlier, and some of the key systems people. Owen Morris, ASPO’s chief of engineering for the lunar module, got back quickly. It was easy to see that people were upset by what was happening, but things hadn’t settled down to the point that much was getting done. Morris called the LEM people up to his desk beside the dais and started handing out assignments. He was also on the phone with Bethpage.

10 minutes. Sitting in the back room, pulling up screens, Dick Brown was beginning to wonder whether it was instrumentation after all. He called to the front room.

“Let’s throw a battery on Bus B and Bus A until we psyche it out,” he said to Liebergot. “We’re getting undervolts.” Brown’s recommendation meant that the command module would be running off battery power. Liebergot suggested the less radical change of limiting the battery to Bus B and leaving the fuel cells on Bus A. “I want to psyche out what those fuel cells are doing here,” Brown repeated. He confessed his growing fear: “We might have a pressure problem in the fuel cells—it looks like two fuel cells simultaneously.”

“That can’t be!” Liebergot protested.

“I can’t believe that right off the bat,” Brown agreed. “But they’re not feeding currents.”

John Aaron’s performance on Twelve crossed Liebergot’s mind. Aaron had seen that unbelievable mess of parameters on his screen and immediately he had come back to Flight—Do this! Do that!—and everything had been straightened out in minutes. Seconds, even. Liebergot felt as if it were taking him forever to figure this thing out. He was disappointed in himself, and at the same time had a sense of being at the edge of an abyss.

Liebergot pushed these thoughts to the back of his mind and, one hand clutching one of the metal handles to the electronics drawer on his console, began the first of a series of attempts over the next hour to resuscitate the electrical power. Occasionally he had to remind himself to swallow.

12 minutes. Guido Will Fenner reported to Kranz that the spacecraft’s attitude was still changing, when it shouldn’t be. “He ought to stop it,” Fenner told Kranz, “he” meaning the crew. Fenner was thinking about one of the perpetual concerns during the Apollo flights, gimbal lock.

All of the spacecraft’s course corrections, orbital insertions, and entry maneuvers depended on knowing not only precisely where the spacecraft was in space, but also the precise attitude of the spacecraft. Whenever the crew performed a maneuver, they had to know, within fractions of degrees, in which direction the nozzle of the engine was pointing. The guidance platform, encased within its three gimbals, provided that information. Unfortunately, there were certain circumstances under which the three gimbals could not perform their function of keeping the guidance platform steady. In particular, if the spacecraft were to get into a position where two of the three gimbals in the guidance system were lined up in the same plane, the gimbals would be unable to let the platform swing free. This was “gimbal lock,” and would result in the platform losing its alignment—in effect, becoming lost.

A fourth gimbal would have taken care of the problem—no matter what crazy position the spacecraft got itself into, the platform would have remained steady as long as the guidance system remained powered up. Gemini had used a four-gimbal system. But a fourth gimbal would have been heavy. It would have added substantially to the bulk of the system, because it would have had to fit around the other three. And the Instrumentation Lab at M.I.T., which designed the system, was comfortable with three gimbals—that’s what the Lab had used for Polaris. Three gimbals were enough for any maneuver the spacecraft might be required to make, as long as everyone paid attention to what was happening. After a long battle with the astronauts (Jim McDivitt threatened not to fly in a spacecraft without a fourth gimbal), a three-gimbal system was used.

In the event that the spacecraft did accidentally go into gimbal lock, it would be possible to realign a platform from scratch, but, as on Apollo 12, it was a complicated, tedious job. And most certainly, Fenner wanted to avoid gimbal lock with a spacecraft that already seemed to have more problems than it could handle. Now, he was warning Kranz that the spacecraft was moving unpredictably. Pushed by an unidentified force, Odyssey was vulnerable to gimbal lock.

14 minutes. Lovell reported seeing something venting from the service module—a gas of some sort. He didn’t say so, but in his own mind he was pretty sure it was oxygen—the O2 Tank 2 pressure was reading zero and there was a big sheet of what looked like white smoke out his window. It added up. He was also pretty sure that it was only a matter of time until the C.S.M. went dead. For Lovell, the moment when he looked out the window and saw the venting was the moment when he stopped being disappointed at losing the lunar landing and started wondering how they were going to get home. For CapCom Jack Lousma, a fellow astronaut, Lovell’s report of venting was the most chilling moment of the flight. The problem could not be just instrumentation or an electrical screwup. Something violent and destructive had happened to the service module more than 200,000 miles away from home.

Kranz understood the implications too, and knew that everyone else in the room had heard Lovell’s calm report. He decided it was time to give a little speech to the controllers. “Okay, now, let’s everybody keep cool,” he began. “We got the LEM still attached, the LEM spacecraft’s good, so if we need to get back home we got a LEM to do a good portion of it with.” He went through the priorities: Don’t blow the command module’s internal batteries, don’t do anything to blow the remaining Main Bus. “Let’s solve the problem,” he concluded, “but let’s not make it any worse by guessing.” Kranz spat out the word as if it were an epithet.

A few seconds after Kranz had finished, Brown called Liebergot on the EECOM loop. He sounded curiously formal, but that was because he had something momentous to say: “EECOM, this is E.P.S. I think we ought to start powering down.” Powering down meant that they would start taking some equipment off line, making it temporarily inaccessible to the crew. It was not something one did to a spacecraft in flight unless there was a compelling reason, but Liebergot immediately agreed—he had been coming to the same conclusion independently. “Let’s get the power-down list, here,” Brown said. For the first time that night, Brown’s voice shook. During the next minute, the only sound on the EECOM loop was the rustle of pages as each of them looked through his copy of the emergency power-down checklist.

It was then that Liebergot looked up at his screen, and suffered yet another shock: The pressure on his other O2 tank was falling.

In preparing for disasters during flight, the world of flight control began with certain verities. One was that structural materials like lines and bulkheads had a reliability not just of .9999 or even .999999, but 1. Most of them had margins that protected them well beyond the worst-case design requirements. A second verity was that while two of the mechanisms of the spacecraft might plausibly fail simultaneously, it was not plausible that two redundant elements within the same system would fail simultaneously. This confidence was based on simple mathematics. With thousands of parts in the spacecraft, even very small probabilities of failure in the individual components added up to a reasonably large probability that some two of those thousands of mechanisms would fail at the same time. But if there are two completely independent oxygen tanks and each of them has a reliability of .9999 (for example), then the chance that both of the oxygen tanks will fail is .0001-squared, or one in a hundred million. The likelihood of a dual failure within any of the handful of major systems—electrical, oxygen, water, propulsion, and the like—was therefore exceedingly small.

Until the moment that he saw the pressure on O2 Tank 1 falling, Liebergot (if he believed his unbelievable screens) had been looking at a double failure in the fuel cells plus an oxygen tank failure. Now, if he accepted the newest information before him, he was confronted with a quadruple failure, a pair of double failures within systems. As Swigert would write later, “If somebody had thrown that at us in the simulator, we’d have said, ‘Come on, you’re not being realistic.’”