Part II : 1950 -1957
Thrust Chamber Design and Fabrication
 The initial failure of the Lewis experiments with liquid hydrogen was primarily one of thrust chamber design. The key to a successful design lies in the injector which can mean high or low performance, durable operation or quick burnout. The function of the injector is to mix the fuel and oxidizer thoroughly and uniformly for complete combustion, while the propellants also cool the injector face. With a good injector, combustion chamber design becomes a matter of providing sufficient volume for the reaction to go to completion, sufficient wall strength to contain the pressure, and sufficient cooling to keep wall temperature within its working limits. The design of the nozzle involves a compromise between providing the optimum contour for complete gas expansion and size, the latter limited by vehicle design.
Rocket experimenters exploring the performance of various propellant combinations usually used either a water-cooled thrust chamber and nozzle or uncooled types that could withstand high temperatures for a few seconds. Most effort was concentrated on obtaining an injector yielding high performance. Following this, the next step was to cool the chamber and nozzle with the fuel. We have already seen that both the Jet Propulsion Laboratory and Ohio State University succeeded in operating regeneratively-cooled hydrogen-oxygen thrust chambers during the 1940s. The Lewis experimenters were trying the same but with larger engines (22 and 89 kN), combining regenerative cooling with thrust chambers of light weight to approach a practical flight design. These objectives had been spelled out in 1952 and reaffirmed each year.
 The first injector used for liquid hydrogen-liquid oxygen at Lewis in 1954 was a like-on-like impingement where jets of the same fluid impinge, breaking the streams into droplets. Mixing is obtained by locating the impinging streams of fuel and oxidizer near each other so that the resulting droplets mix well. Ohio State University used this type and it was popular among rocket experimenters. JPL used an injector with impinging hydrogen jets and an oxygen spray. Aerojet's best injector was a multiple-tube, concentric type where each jet of hydrogen was surrounded by a sheath of oxygen (fig.10). The three successful runs by the Lewis group in February 1956 used a "tube bundle" injector where a large number of small tubes carried the hydrogen and oxygen into the chamber with a fine degree of mixing.
By all experience and design principles, the hydrogen-fluorine injector used in the first Lewis laboratory run in March 1956 should have worked well. It consisted of four rings of hydrogen holes producing streams parallel to the combustion chamber axis, alternating with four rings of similar holes for fluorine. The holes were small and mixing was good; but when tried, the operator summed the results in four cryptic sentences in his log:
Parallel to these experiments, more detail studies were under way at the Lewis laboratory on fundamentals of injector design. Such work had been in progress since the early 1950s, but it was not until 1956 that experiments in this basic work focused on hydrogen. Carmen M. Auble studied six types of injection methods for hydrogen-oxygen in a small (900 N) thrust chamber.35 Gaseous hydrogen, chilled to the temperature of liquid nitrogen (77 K), simulated the physical characteristics of hydrogen after it served as a coolant prior to injection. Not surprisingly, Auble found correlation between the mixing and spreading of the propellants and performance over a range of propellant mixture ratios, all his designs doing well at fuel-rich ratios. Increasing the temperature of the hydrogen to room temperature was beneficial. Compared with hydrocarbon fuels, hydrogen needed a fifth as much volume for comparable combustion efficiency. Separate and parallel jets, as used in the hydrogenfluorine run, did as well as injectors that promoted mixing. Auble found, however, that combustion efficiency was controlled more by the degree of oxygen vaporization than by hydrogen dispersion and mixing.
Late in 1957, Marcus F. Heidmann and Louis Baker, Jr., extended Auble's investigation, combining it with earlier analyses of propellant vaporization as a ratecontrolling step in combustion. They investigated fourteen injectors for hydrogenoxygen in an engine of the size that Auble had used. Their investigation confirmed that the degree of oxygen atornization was the primary factor affecting combustion efficiency.36
Concurrent with injector and performance studies were several investigations of fabrication techniques for lightweight and cooled combustion chambers and nozzles. In 1953, John E. Dalgleish, a fabrication expert, and A. O. Tischler, a rocket researcher, worked together on lightweight thrust chambers using an electroforming technique.37 In 1954, Tischler placed orders in the shop for two other types. One used tubes formed according to the contour of the combustion chamber and nozzle and  brazed together-a method used by several rocket manufacturers starting with Reaction Motors. The other type was similar except that, instead of tubes, channels were formed and then brazed or welded together with a closure over the channels to complete the coolant passage and strengthen the whole assembly. Both of these experimental types were still in the shop two years later as they had been given a low priority.
Until 1956, the primary responsibility for designing thrust chambers rested with an engineering service group headed by William A. Anderson. He developed a fabrication technique consisting of an inner shell of spun metal, wire spacers to form spiral coolant passages on the outside of the shell, and a welded "clam-shell" outer wall to enclose the coolant passages. A variation of this method was to form the outer shell of square wire brazed together. The Anderson design was successfully used on engines of 4.5 kilonewtons and was the prime design for larger engines until 1956-1957.
Obtaining experimental engines was hampered by increasing congestion in the fabrication shop. The NACA shops were unexcelled in advanced fabrication techniques and willingly accepted all challenging work, but delivery was sometimes delayed by an avalanche of orders or work of higher priority. In 1956, the shops had orders for over a dozen thrust chambers of various designs and delivery was delaying experimentation. Steps were taken to reduce the number of designs, and Silverstein assigned Edward Baehr, a gifted design and fabrication engineer, to assist the rocket group. Baehr made a major contribution to the rocket effort by choosing a design something like Tischler's channel-wire wound type and successfully fabricating it. It consisted of a number of longitudinal channels of varying depth according to the coolant velocity required. These were bonded together to make up the chamber and bound by stainless steel wire wrapping which was brazed to make a fluid-tight and strong outer skin (fig.17). This design was used in 1957 and subsequently.38
Since the early 1950s, Lewis associate director Abe Silverstein had been interested in liquid hydrogen as a fuel for both jet engines and rockets. In the spring of 1957, he decided that it was time to hold a research conference on results of the laboratory's investigations. That conference, plus additional emphasis on rocket research at the laboratory, meant unprecedented support for the rocket group, and they made the most of it. Silverstein became more involved in rocket problems and on 9 August 1957 held one of his famous after-hours staff conferences on the subject of injector design. The informal session was held in the control room of the new facility-complete with beer and pretzels, compliments of Silverstein-and ran past midnight. Everyone contributed his views.* The author remembers stressing the concepts of mixing on a very fine scale coupled with uniform mixing except for a fuel-rich cooling region at the chamber walls. These concepts, not new, were adopted along with other design ideas such as selecting angles of jet impingement well away from the injector face, avoidance of recirculation of reactants across the injector face, and fuel-cooling of the injector face. In September, Silverstein held another meeting on injectors as well as other rocket design problems for experiments intended to be reported at the coming conference.39
* Attendees were: Silverstein, W.T. Olson, Edward Bachr, Vearl Huff, M.F. Heidmann, A.O. Tischler, Howard Douglass, George Kinney, William Anderson, and the author.