Thrust Chamber Experiments, 1947-1949
 From their previous work, Young and Gordon were confident that the flared tube configuration, with its very small combustion chamber, was the best design for the thrust chamber of 13.3 kilonewtons (3000 lb thrust). They intended to use a porous inner wall but were still undecided about the coolant. They decided to determine the relative merits of both water and liquid hydrogen as transpiration coolants. They also planned to study injection methods for liquid hydrogen. Stary was studying the same things at Ohio State University and had just made his first run using liquid hydrogen (p.20).
From mid-1947 to mid-1948, the Aerojet men made few thrust chamber tests. None was made with liquid hydrogen, for the liquefier was not yet in operation. The major experimental work was an investigation of the performance loss at sea level in operating a nozzle designed for maximum performance at altitude.
 The force produced by a nozzle from expanding exhaust gases is the result of a momentum force and two pressure forces. One of the pressure forces aids the momentum force and the other opposes it. An ideal nozzle is one that expands the exhaust gases from the pressure in the combustion chamber to the outside ambient pressure. The nozzle thereby maximizes the momentum force and the two pressure forces cancel each other. Since a rocket nozzle is a fixed design, the designer must choose a single ambient pressure for his design. If he chooses sea-level pressure, he gets less than optimum performance at altitude; if he chooses a lower pressure corresponding to some altitude, he theoretically loses performance at sea level. Since much of the operation occurs at reduced ambient pressure, the designer usually wishes to make the nozzle as large as mass and size restrictions permit. The question at Aerojet was: What penalty would result from sea-level operation of a nozzle designed for best operation at altitude? In experiments with a small rocket chamber they found, to their great joy, that the actual performance loss was much less than theoretically predicted their nozzle designed for altitude had only a 10 percent loss at sea level.*
Aerojet was still committed to transpiration cooling but had encountered a series of new and worrisome material problems. It was difficult to obtain porous materials of uniform permeability-but worse yet, the porous structure became clogged in unpredictable and nonuniform ways. These problems began to raise doubts about using the flared tube configuration as well as transpiration cooling. When the project received new funding and directions in mid-1948, Aerojet planned to use a group of thrust chambers of various sizes and shapes, as well as a variety of injection methods. The engineers believed regenerative cooling would be possible with either oxygen or hydrogen, or both. Preparations were made to study the heat transfer properties of oxygen and hydrogen by means of an electrically heated tube. All of these activities signaled a major change in direction by Aerojet, from emphasis on their flared tube design using transpiration cooling to a conventional plenum thrust chamber with regenerative cooling. It was about this time, mid-1948, that George H. Osborn became the chief test engineer.
The first Aerojet test with liquid hydrogen and oxygen was made on 20 January 1949 with a 1780-newton (400 lb thrust) chamber. By the end of March, 10 runs had been made with disappointingly low exhaust velocities-about 2920 meters per second or 82 percent of theoretical. Of equal concern was the unsteady operation, or "chugging," which indicated unstable combustion. The injector, designed by Osborn, used a diverging cone of liquid oxygen intersecting a converging sheet of liquid hydrogen. The only good news was a low heat transfer rate, which was attributed to incomplete combustion.
In the midst of all the bad experimental results came the worst news of all. On 2 March 1949, as previously mentioned, the Bureau of Aeronautics directed Aerojet to change the fuel from liquid hydrogen to anhydrous hydrazine, but allowed the experiments with liquid hydrogen to continue for the three months remaining in the contract. No evidence has been found that Aerojet protested this change- perhaps it  was welcomed after the first series of experiments with liquid hydrogen. However, the Aerojet designers were determined to do a creditable job with liquid hydrogen in the time remaining and the record shows that they did. The key was injector design.
Osborn was designing new injectors even before all the dismal results with the spray type were in. The second design was a "showerhead" type with 115 fuel and oxidizer holes across the face and 30 fuel holes around the circumference for film cooling. The film, or layer, of fuel-rich gas next to the chamber and nozzle walls kept them cool. The design gave low performance and failed structurally on 4 April, three months before the end of the contract.
The pressure on the team to succeed must have been great. Fortunately, Osborn had designed a third injector, called a multitube concentric orifice, in March and it proved to be highly successful. Liquid hydrogen was injected through a number of thin-walled tubes surrounded by an annular flow of liquid oxygen, as illustrated by figure 10. For....
 ....the 1780-newton (400-lb-thrust) chamber, 61 of these "tubes within tubes" provided a very fine degree of mixing. As in the previous design, axial orifices were spaced around the circumference for hydrogen film cooling. Two runs with this injector gave an exhaust velocity of 3590 meters per second, or virtually 100 percent of theoretical. The propellants mixed so well that combustion occurred very close to the injector face and burned it. Osborn sought to correct this with design changes, but the fix did not work as well as the original design. However, he knew how he wanted to design the 13-kilonewton (3000-lb-thrust) injector. When he signed the drawing for it on 5 May, there were less than two months left to complete the work. The injector had 489 sets of circular oxygen orifices surrounding hydrogen tubes, plus 60 hydrogen orifices for a fuel-rich layer at the walls. The thrust chamber, which had been designed and fabricated earlier, was a conventional plenum chamber, water cooled, with an inner liner of copper. The copper was machined from a solid billet and its size limited the nozzle design so that it was not ideal.** Starting on 27 May three successful runs were made with this engine at pressures from 24 to 31 atmospheres. Exhaust velocities of 3380 to 3520 meters per second were obtained, approaching 95 percent of theoretical performance. On 16 June, with two weeks to go before the contract expired, they attempted to make a fourth run, but an explosion occurred in the liquid hydrogen propellant system-the second in that system. Aerojet attributed the cause to contamination of the liquid hydrogen with solid oxygen. That ended Aerojet's rocket experiments with liquid hydrogen.
In reporting the results, Osborn and Wayne D. Stinnett included experiments by Gordon on heat transfer and injectors using a smaller, water-cooled engine where the multitube, concentric injector had initially proved successful. Heat transfer rates were reported as excessive for both engines, leading the authors to conclude that additional film cooling over that used in the larger engine would be necessary. Although they had not fulfilled the objective of a self-cooled, lightweight rocket engine using liquid hydrogen-oxygen, the investigators believed that their results were highly encouraging, and no fundamental difficulties were encountered. From their rapid progress during the last four months of the contract, there is little doubt that Aerojet was on the right track in thrust chamber design and with additional work would have been able to perfect self-cooling. Concurrent with their work, Dwight I. Baker at nearby Jet Propulsion Laboratory was doing just that.