CONTINUING CHALLENGES

Unfortunately, the reliability demonstrated during the PFRT program did not continue at Edwards. Early in the flight program, vibrations, premature chamber failures, pump seal leaks, and corrosion problems plagued operations. Potentially the most serious problem was a 1,600-cycle vibration. Fortunately, the natural frequencies of the engines dampened the vibration below 100 g. However, between 100 and 200 g, the vibration could be dampened or could become divergent, depending on a complex set of circumstances that could not be predicted in advance, and the vibration always diverged above 200 g.[80]

The vibrations caused a great deal of concern at Edwards. On 12 May 1960, as the program was trying to get ready for the first XLR99 flight, the Air Force called a meeting to discuss the problem. Although Reaction Motors had experienced only one vibration shutdown every 50 engine starts at Lake Denmark, personnel at Edwards reported that there had been eight malfunction shutdowns out of 17 attempted starts. The vibration began when the main-propellant valves opened for final chamber start, although the engines had not experienced vibrations during the igniter phase. Since the demonstrated rate of occurrence had jumped from 2% at Lake Denmark to 47% at Edwards, nobody could ignore the problem. Engineers discovered that the 1,600-cycle vibration corresponded to the engine-engine mount resonant frequency, and that Reaction Motors had not seen the vibration using the earlier non-flight-rated engine mounts at Lake Denmark. As a temporary expedient, Reaction Motors installed an accelerometer that shut the engine down when the vibration amplitude reached 120 g, a move the company believed would permit flight­testing to begin.-181

The engine (serial number 105) used at Edwards differed only slightly in configuration from those used at Lake Denmark; for example, it used an oxidizer-to-fuel ratio of 1.15:1 instead of 1.25:1. The desired operating ratio at altitude was 1.25:1, and this is what Reaction Motors had used during their tests. However, to simulate the 1.25:1 ratio on the ground, the engine had to run at 1.15:1 to compensate for atmospheric and propellant density differences at the lower altitude. Reaction Motors had tested this reduced oxidizer-to-fuel ratio only twice at Lake Denmark, and had not encountered vibrations either time. The company recommended a series of actions, including checking for purge gas leaks at the PSTS, changing the propellant ratio back to 1.25:1, and performing more engine test firings.-82

By the beginning of June 1960, the problem did not seem to be getting any better. The Air Force conducted two tests with 17 starts on engine 105 at Edwards, with two vibration shutdowns using the ground orifice (1.15:1 ratio). When engineers reinstalled the flight orifice (1.25:1 ratio), three of five starts resulted in vibration shutdowns. Reaction Motors conducted 18 starts on engine 104, and three of the four initial starts resulted in vibration shutdowns, but all restarts were successful.82

A series of minor changes made to engine 104 by Reaction Motors seemed to ease the problem, and between the middle of July and the middle of August 1960, the engine accumulated 25 starts at Edwards without any vibration-induced shutdowns. In fact, only a single malfunction shutdown of any type was experienced, which was attributed to a severe "throttle chop" that the turbopump governor could not keep up with. Other XLR99s had experienced similar problems, and Reaction Motors warned the pilots to move the throttle slowly to avoid the situation.-1841

The Propulsion System Test Stand was the unlikely name for a non-flight X-15 fuselage that was used to test rocket engines. At least two of the fuselages were manufactured, one for Reaction Motors and one for Edwards AFB. Here technicians install an XLR99 in the PSTS in preparation for a test. (NASA)

Still, as late as the meeting of the Technical Advisory Group on 9-10 November 1960, the vibration problem persisted and the Air Force launched an effort to solve the problem. This program used two engines (006 and 012) at Lake Denmark and completed a series of baseline tests by the end of November that showed a 30% incidence rate of vibration shutdowns with the flight orifices installed. Reaction Motors found that modifying the liquid-oxygen inlet substantially lowered the incident rate of vibration shutdowns. Since this modification did not seem to have any other noticeable effect on the engine, the Air Force adopted it as a temporary fix.[85]

Separately, Reaction Motors determined that o-ring deterioration at the casing joint caused fuel pump seal leaks. Replacing the o-ring was difficult because it took technicians two or three shifts to remove the turbine exhaust duct, stator blades, rotor, and inlet housing; just to remove the exhaust duct necessitated the removal and re-safety-wiring of 60 bolts. Thus, although the o – ring failure itself was not serious, since it simply resulted in a steam leak, the repair required removing the engine from the aircraft, performing a time-consuming engine disassembly, and revalidating the engine installation. This process directly contributed to early flight delays using the XLR99.[86]

Ironically, the corrosion problem appeared to be the result of the unusually long engine life. With a few exceptions, the materials used by Reaction Motors for the turbopump were compatible with the various propellants, but those in contact with the hydrogen peroxide were experiencing more corrosion than desired. There were also some instances of galvanic action between the magnesium pump case and steel parts with decomposed peroxide as an electrolyte. As one

researcher noted, "the only thing really compatible with peroxide is more peroxide." There were no obvious fixes, so the program lived with the problem.[87]

The premature failure of the thrust chambers was of more concern. To insulate the stainless-steel cooling tubes from the 5,000°F flame, Reaction Motors used a 0.005-inch-thick, flame-sprayed Nichrome®-881 undercoat with 0.010 inch of oxygen-acetylene flame-sprayed Rokide Z zirconia as an insulating, erosion-resistant top coating. In service, the Rokide coating began to spall or flake due to thermal cycling from the large number of engine starts, and from vibration effects from an unstable flame. For instance, by January 1961 about 50 square inches of Rokide coating had peeled off engine 108 at Edwards, including 14 inches during a single vibration shutdown. The loss of the coating exposed the cooling tubes to the heat and erosive effects of the flame, overheating the ammonia coolant within the tubes and reducing the amount of cooling available. The superheated ammonia vapors also attacked the stainless steel and formed a very brittle nitrided layer. At the same time, the combustion gases began to melt and erode the tube surface. As this condition continued, the effective thickness of the tube wall gradually decreased until it burst. Raw ammonia then leaked into the chamber, causing more hot spots and eventually the complete failure of the chamber.-1891

In January 1961 the X-15 Project Office and the Materials Central Division of the Aeronautical Systems Division at Wright Field initiated a study of methods to improve the chamber life of the XLR99. Two possible approaches were to attempt to improve the Rokide coating system, or to develop an improved coating. The Air Force contract with Reaction Motors already included an effort to improve the Rokide coating, but researchers expressed little faith that this would achieve any measurable results. This resulted in the Air Force initiating a program to develop an alternate coating. In the meantime, engineers at the NASA Flight Research Center (FRC) surveyed other rocket engine manufacturers to find out whether they had developed workable processes. Both Rocketdyne and Aerojet were doing extensive laboratory testing of ceramics applied with plasma – arc devices, but neither had put the process into production. Both companies indicated that their experience with flame-sprayed alumina and zirconia had been unsatisfactory. Instead, Rocketdyne was working on metal-ceramic graduated coatings, and Aerojet was investigating the use of refractory metal (molybdenum and tungsten) overcoats on top of ceramics.-1901

At the time, the Air Force already had a contract with the Plasmakote Corporation to study graduated coatings in general, and this contract was reoriented to solving the XLR99 problem specifically. A second contract, this one with the University of Dayton, was reoriented to provide realistic techniques for laboratory evaluations of the coatings.-1911

A graduated coating consisted of sprayed layers of metal and ceramic; the composition changed from 100% metal at the substrate to 100% ceramic at the top surface. This removed the traditionally weak, sensitive interface between the metal and ceramic layers. Researchers produced the coatings by spraying mixed powders with an arc-plasma jet and gradually changing the ratio of metal and ceramic powders, with most of the coatings using combinations of zirconia with Nichrome, molybdenum, or tungsten. The FRC recommended adopting the new technique immediately as a way to repair damaged chambers at Edwards. They noted that engine 101 had been patched using Rokide coating, but the engine would soon need to be repaired again since the coating was not lasting. The Air Force and NASA decided that the next patch on engine 101 would use the new process, and NASA built a special fixture at the FRC to allow the chamber of a fully assembled engine to be coated.-921

Before the new coating was applied, NASA tested an existing Rokide chamber for 5.5 minutes, and 25 square inches of Rokide coating was lost during the test. Engineers then applied a

graduated coating segmented into areas using several different top coats, including tantalum carbide, titanium carbide, titanium nitride, zirconia with 10% molybdenum, and zirconia with 1% nickel. This chamber ran for 5.75 minutes, and only 3 square inches of the new coatings were lost. However encouraging, the tests were of relatively short duration and researchers did not consider them conclusive. One thing that became apparent during the tests was that it would be extremely difficult or impossible to reclaim failed chambers if the coating wore thin or was lost, since the internal damage to the tube might be sufficient to cause it to fail with no visible damage.[93]

One of the most significant issues experienced by the XLR99 during the flight program was the premature failure of the thrust chambers. Researchers eventually traced this to the spalling or flaking of the Rokide Z zirconia coating that had been applied to the inside of the chamber as an insulator. Although improved coatings were eventually developed, the Flight Research Center also developed an in-house capability to recoat the chambers when necessary, resulting in a significant cost savings compared to sending the chambers back to Reaction Motors or procuring new chambers. (NASA)

The Technical Advisory Group met on 11-12 January 1961 at the Reaction Motors facility at Lake Denmark. All in attendance agreed that chamber durability needed to be increased, and supported the development of a quick-change orifice to simplify ground runs. The group also recommended that the X-15 Project Office initiate the procurement of six spare chambers and sufficient long – lead material to construct six more. It could not be determined whether these chambers were actually procured.-194

Some documentation indicates that the XLR99 was redesignated YLR99 on 29 December 1961, although nothing appears to have changed on the engines themselves. The original source documentation from the period is inconsistent in its use of XLR99 or YLR99; this history will use XLR99 throughout simply to avoid confusion.1951

By March 1962, technicians at the FRC had the necessary equipment and training to recoat the chambers as needed. The cost of the tooling had come to almost $10,000, but the cost to recoat a chamber was only about $2,000-much less than the cost of procuring a new chamber from Reaction Motors. The coating finally approved for use consisted of 30 mils of molybdenum primer in the throat and 10 mils elsewhere, followed by 6 mils of a graduated Nichrome-zirconia coating and then 6 mils of a zirconia topcoat. NASA used this coating process for the duration of the flight program with generally satisfactory results.1961

As is the case with almost any new technology, some things can never be fully understood. One of the harder things to grasp when dealing with complex mechanical devices is component matching (or mismatching), i. e., why some items will work in a particular assembly and other seemingly identical items will not. For example, during the initial checkout of engines 108 and 111 at Edwards, both engines exhibited excessive vibrations. NASA replaced the igniter in engine 108 with a spare that reduced the vibration to acceptable levels. The igniter that had been removed from 108 was then installed in 111 and its vibration was reduced to acceptable levels. Compatibility was not a particular problem, but scenarios such as this did point out some puzzling inconsistencies.-1971