Apollo: The Race to the Moon (22 page)

On November 7, 1962, Webb announced to the press that he was confirming the tentative decision of July, and that the Grumman Engineering Corporation had been chosen to build the lunar module. Eighteen months after the nation had decided to go to the moon, NASA had decided how.

Chapter 10. “It aged me, I’m sure”

On June 28, 1962, the same day Jim Webb learned that the centers had agreed on lunar-orbit rendezvous, NASA’s first F-1 engine destroyed itself on a test stand at Edwards Air Force Base in California.

The F-1 engine was the heart of the Saturn V and the technological achievement that more than any other made Kennedy’s lunar commitment possible. At the time Kennedy was considering his decision to go to the moon, Low’s task group was calculating that the spacecraft and rocket for a direct lunar ascent would weigh at least ten million pounds. While this didn’t necessarily mean that they had to build a rocket with more than ten million pounds of thrust (if they decided to go to the moon by earth-orbit rendezvous instead of direct ascent, they could split the load up), it did demand an engine that was a quantum leap ahead of the largest one in the American inventory, an uprated H-1 with 188,000 pounds of thrust.

Marshall had already let a contract with the Rocketdyne Corporation for such an engine, the F-1. Engineered to produce 1.5 million pounds of thrust, eight F-1’s in a single first stage (the contemplated Nova) could lift the required payload for lunar landing via direct ascent. Four F-1’s could achieve an earth-orbit rendezvous mission in two launches. Five F-l’s, producing 7.5 million pounds of thrust, could permit a lunar landing in a single launch by using lunar-orbit rendezvous.

The question in the spring of 1961 was whether the F-l was in fact feasible—not even the Russians with their large boosters had an engine half its size. On April 6, 1961, a week before Gagarin’s flight, a prototype thrust chamber for the F-l had achieved a peak thrust of 1,640,000 pounds. But this was a brief onetime test of a prototype, which was far from being a functioning engine. Kennedy’s May 25 decision to go to the moon had to be based on the prediction that eventually the F-l would work. If it didn’t, the Apollo Program would have to come to a halt until an alternative could be developed.

This prediction was by no means a sure thing. NASA’s rocket engineers had chosen 1.5 million pounds of thrust as the goal for the F-l not because they knew it was within their capabilities, but because that’s what they needed for a launch vehicle that could be used to build space stations or go to the moon, the missions that von Braun had in mind. “They were cavaliers and they were pioneers,” said an American engineer of the Germans. “They were willing to say, ‘Well, okay, it’s going to be one and a half million pounds.’ They didn’t go through mountains and mountains of computer programs to try to figure it out. A lot of personal judgment went into it.”

Despite the progress that had been made by the spring of 1961, the F-l was still a gamble. The early development work that had produced the brief 1.64 million pounds of thrust had also been troubled by persistent problems. A number of senior advisers, including at least one member of the President’s Science Advisory Committee, were arguing that the F-l was just too big to work.

Rocketdyne, the contractor for the F-l, was located at Canoga Park in the San Fernando Valley north of Los Angeles. The company had a long and successful history of collaboration with von Braun’s team, and during the remainder of 1961 and into the first half of 1962, it looked as if the F-l would be another thoroughgoing success. But the initial tests of the engine were done in small increments, with little high-frequency instrumentation, and throughout this optimistic period the engine was shut down at the first hint of problems.

The Rocketdyne engineers tested their engine at Edwards Air Force Base, 120 miles east of Canoga Park, where a gigantic test stand for the F-l had been anchored deep in the floor of the Mojave Desert. It was there, on June 28, 1962, that they let the F-l run for a few seconds longer than before and it destroyed itself. “Destroy” in this case meant melting through thick, high-strength steel plate in a quarter of a second. But that wasn’t the worst of it. The F-1 failed not because of a faulty weld or a flaw in an alloy. Those kinds of failures were comparatively easy to remedy. Instead, the F-l destroyed itself because of combustion instability.

In principle, liquid rocket engines are simple, far simpler than the internal combustion engine. Liquid fuel is pumped into a combustion chamber in the presence of liquid oxygen and a flame. It burns. That’s all there is to it. There are no crankshafts to turn, no pistons to drive. The burning fuel produces energy in the form of gases that exit through the rocket’s nozzle. The force the gases produce against the top of the engine is called thrust. The thrust is transmitted through the rocket’s structure and, if it is greater than the weight of the rocket, the rocket lifts off. Put in its most basic terms, for any rocket to work there are two things that must be done extremely well: The propellants must be brought together, and then they must burn smoothly.

In the F-l, just pumping the propellants to the combustion chamber raised unprecedented demands. The F-l used liquid oxygen (LOX) and R.P.-1, a form of kerosene. The pumps, one for the fuel and one for the LOX, had to deliver the kerosene from the tankage to the combustion chamber at the rate of 15,741 gallons per minute, and the LOX at the rate of 24,811 gallons per minute. Driven by a 55,000-horsepower turbine, the pumps had to operate at drastically different temperatures: 60 degrees Fahrenheit for the fuel, –300 degrees for the LOX, while the turbine itself ran at 1,200 degrees. To complicate matters, the whole assembly had to be light and compact enough to fit on board the rocket and nonetheless sturdy enough to resist the pressures, vibrations, and other stresses of launch and flight.

Developing the pumps was still not as hard as solving the second basic problem of rocket engines: making the propellants burn smoothly once they had reached the combustion chamber. The pumps brought the kerosene and the LOX to a circular metal slab three feet in diameter and about four inches thick, weighing 1,000 pounds, called the injector plate. The kerosene arrived by a circuitous route, first acting as a coolant for the engine shell by passing through a labyrinth of tubing on the walls of the combustion chamber and nozzle. It made for a complicated pumping system, but it saved weight—an engine made of metal strong enough to withstand the temperatures of the F-1 throughout launch would have been prohibitively heavy.

The injector plate was pocked with 6,300 holes less than a quarter of an inch in diameter through which the kerosene and LOX entered the combustion chamber. Most of the propellant streams were arranged in groups of five. Two of the five, both kerosene, impinged on each other at a carefully defined distance below the top of the plate, forming a fan-shaped spray. The other three in each five-hole group were of LOX. These also impinged on one another, forming another fan. The two fans intersected. There, given the presence of a flame, they would combust.

In the F-l, the combustion chamber was a barrel about thirty-six inches wide and thirty inches long, closed at one end by the injection plate and opening into a nozzle at the other end. A few seconds before ignition, four small pre-burners in the combustion chamber—pilot lights, in effect—were lit, providing a flame at the point of impingement. As the pumps screamed up to speed, valves snapped open and more than a ton of kerosene and two tons of liquid oxygen burst into the combustion chamber. Per second. The gases produced by their ignition roared out through the throat, the open bottom of the barrel, into the cone of the nozzle below. In the course of the few seconds from ignition to full power (mainstage), the interior of the combustion chamber went from ambient temperature to 5,000 degrees Fahrenheit. At the face of the injector plate, pressure went from zero to 1,150 pounds per square inch. Given that combination of propellants, pressures, and nozzle design, the force generated totaled 1.5 million pounds. In the first stage of a Saturn V, five F-l s were to ignite simultaneously and sustain mainstage combustion for 150 seconds.

By the early 1960s, creating an engine to withstand the temperatures and the pressures of the F-l was, thanks to new metallurgical and engineering techniques, not a formidable problem. The difficulty was to achieve what the engineers called a “smooth flame front,” in which the kerosene and oxygen combined and burned at a uniform temperature across the face of the injector plate.

Achieving this stable combustion with an injector plate three feet in diameter created unprecedented problems. If, for example, the holes in the plate were drilled so that one side of the flame front had a slightly higher oxygen content than the other side, the high-oxygen area would get hotter and produce higher pressures on that side. In a smaller combustion chamber, this imbalance might not create difficulties. But in the F-l, there was plenty of room for a racetrack effect to get started, in which a higher pressure on one side of the chamber would bounce, starting a wave front that would begin careening around the perimeter of the barrel. Within milliseconds, the heat fluxes inside the chamber would be bounding back and forth across the combustion chamber, reinforcing each other, going out of control, and destroying the engine.

“The slightest thing could trigger it,” said one of F-l’s engineers of combustion instability. This was a vexing situation, because the inside of an F-l combustion chamber during launch was prone to develop a variety of “slightest things.” If the pumps cavitated and failed to supply the propellants to the injector plate at an absolutely uniform rate, the streams of propellant and LOX impinged at the wrong points and could disrupt the burning process. Thermal shocks as the engine went from ambient temperature to 5,000 degrees could disrupt the burning process. Acoustical shocks that hit the chamber at the moment of ignition were the most troublesome of all. With the sole exception of a nuclear explosion, the noise of a Saturn launch was the loudest noise ever produced by man. The only sound in nature known to have exceeded the noise of a Saturn V was the fall of the Great Siberian Meteorite in 1883. Sound waves of such force tended to disrupt the burning process.

Combustion instability had been a problem in rocketry ever since Peenemünde days, and it had been solved for each generation of rockets through fixes that were more or less ad hoc. With the F-l, all the usual difficulties of fixing combustion instability were compounded by the enormous size of the engine. No one had ever tried to cure combustion instability under remotely comparable conditions.

When the F-1 destroyed itself out at Edwards, Jerry Thomson was working as chief of Marshall’s Liquid Fuel Engines Systems Branch. Thomson, a soft-spoken, small-town boy from Bessemer, Alabama, had come to Marshall by way of the Marines, where he had been a seventeen-year-old rifleman in World War II, and Auburn University, where he had gotten his degree in mechanical engineering. Early in the 1950s, Thomson had chanced into the rocket engineering business and discovered that he had an aptitude for designing combustion chambers. By 1962, his aptitude was such that “somehow,” as Thomson put it, he “got picked to resolve this combustion instability situation.” NASA headquarters and Marshall considered the problem so critical that Thomson was told to turn over the operation of his branch to his deputy and move out to Canoga Park.

“It aged me, I’m sure,” he later said with a sigh. While Thomson represented NASA, Paul Castenholz, a propulsion engineer, and Dan Klute, a Ph.D. in mechanical engineering research, headed up the contractor team for Rocketdyne. Like Thomson, Castenholz and Klute had a special talent for the half-science, half-art of combustion chamber design, and like Thomson they had been pulled off their management duties as senior engineers at Rocketdyne to work full time on the combustion instability problem. In all, about fifty engineers and technicians were assigned to what they called the Combustion Devices Team, augmented by technical support back at Marshall and consultants from universities, other NASA centers, and the Air Force. Within Rocketdyne, Castenholz remembered, they got whomever they wanted, to do whatever they asked, whenever they needed it—the Combustion Devices Team had the highest priority in the company.

As a process, the attack on the F-l’s combustion instability was a model of government-contractor collaboration. Neither side blamed the other for the failures; both sides worked together as if for a single employer. The process was fine. Its only defect was that for months thereafter it showed few signs of producing a solution.

At first, the team thought they could resolve the problem without having to redesign the whole system. They concentrated on the system’s hydraulics—the flow rates of the fuel and LOX and the patterns of the holes. When they had completed a modified version of the combustion chamber, they tried it out at the Edwards test stand. After a few tests it too went unstable and did not recover. They redesigned again, tested again, and again the engine failed.

There was no pattern to the instability, and it would occur “for reasons we never quite understood,” Thomson said. “It could initiate soon after you got the engine started, the engine would go unstable and destroy itself, or it might occur toward the end of the run. It just depended on the conditions that perpetuated the instability.”

By January 1963, their tests had completely destroyed two more engines, and Brainerd Holmes called the NASA-Rocketdyne team up to Washington. Holmes was prepared to ask Congress for funds to begin a parallel development effort on another system, he informed them. This was not the time for false pride. The program depended on making the right decision. He went around the table, asking each man what he thought. Each in turn told Holmes that a parallel effort wouldn’t be necessary. They would fix the F-1 themselves. And back they went to Canoga Park.

They set a new goal for themselves. It was useless, apparently, to try to design the F-l so that it never began to go unstable. The engine was too big and subject to too many disturbances. Now they would consider it tolerable for the engine to initiate instability—all it had to do was achieve dynamic stability, which meant that it would correct itself. After instability began, the engine must (according to their goal) damp it out within 400 milliseconds.