Category Hypersonics Before. the Shuttle

X-15 Design and Development

Harrison A. “Stormy” Storms, Jr. led the North American X-15 design team, along with project engineer Charles H. Feltz. These two had a difficult job ahead of them, for although giving the appearance of having a rather sim­ple configuration, the X-15 was perhaps the most technologically complex single-seat air­craft of its day. Directly assisting Storms and Feltz was test pilot A. Scott Crossfield, who had worked for the NACA prior to joining North American with the intended purpose of working on the X-15 program. Crossfield describes Storms as a man of wonderful imagination, technical depth, and courage… with a love affair with the X-15. He was a tremendous ally and kept the objectivity of the program intact….” According to Crossfield, Feltz was a remarkable ‘can do and did’ engineer who was very much a source of the X-15 success story.”1

Storms himself remembers his first verbal instructions from Hartley Sould: “You have a little airplane and a big engine with a large thrust margin. We want to go to 250,000 feet altitude and Mach 6. We want to study aero­dynamic heating. We do not want to worry about aerodynamic stability and control, or the airplane breaking up. So if you make any errors, make them on the strong side. You should have enough thrust to do the job.” Adds Storms, “and so we did.”2

Crossfield’s X-15 input proved particularly noteworthy during the early days of the development program as his experience per­mitted the generation of logical arguments that led to major improvements to the X-15. He played a key role, for instance, in con­vincing the Air Force that an encapsulated ejection system was both impractical and

X-15 Design and DevelopmentBy the time of the first industry conference in 1956, this was the design baseline for the North American X-15. Note the tall ver­tical stabilizer, and the fact that it does not have the distinctive wedge shape of the final unit. Also notice how far forward the fuselage tunnels extend—well past the canopy. (NASA)

unnecessary. His arguments in favor of an ejection seat, capable of permitting safe emergency egress at speeds between 80 mph and Mach 4 and altitudes from sea level to 120,000 feet, saved significant money, weight, and development time.

There has been considerable interest over whether Crossfield made the right decision in leaving the NACA since it effectively locked him out of the high-speed, high-alti­tude portion of the X-15 flight program. Crossfield has no regrets: “… I made the right decision to go to North American. I am an engineer, aerodynamicist, and designer by training… While I would very much have liked to participate in the flight research pro­gram, I am pretty well convinced that f was needed to supply a lot of the impetus that allowed the program to succeed in timeli­ness, in resources, and in technical return…. I was on the program for nine years from conception to closing the circle in flight test. Every step: concept, criteria, requirements, performance specifications, detailed specifi­cations, manufacturing, quality control, and flight operations had all become an [obses­sion] to fight for, protect, and share—almost with a passion.”3

Although the first, and perhaps the most influential pilot to contribute to the X-15 program, Crossfield was not the only one to do so. In fact, all of the initially assigned X-15 pilots participated in the development phases, being called on to evaluate various operational systems and approaches, as well as such factors as cockpit layout, control sys­tems, and guidance schemes. They worked jointly with engineers in conducting the sim­ulator programs designed to study the aspects of planned flight missions believed to present potential difficulties. A fixed – base simulator was developed at North American’s Los Angeles facility, containing a working X-15 cockpit and control system that included actual hydraulic and control – system hardware. Following use at North American, it was subsequently relocated to the Flight Research Center4 (FRC) at Edwards AFB. Once flight research began, the simulator was constantly refined with the results of the flight test program, and late in its life the original analog computers were replaced by much faster digital units. For the life of the program, every X-15 flight was preceded by 10-20 hours in the simulator.

A ground simulation of the dynamic envi-



X-15 Design and DevelopmentПодпись:Подпись: VELOCITY 6600 FT PER SEC DESIGN ALTITUDE 260,000 FT LANDING SPEED 164 KN POWER PLANT*RMl MAX THRUST (40.000 FT) 57,000 LB MIN THRUST (40,000 FT) 17,000 LB WING AREA 200 SQ FT SWEEP c/4 25 0E6BEES THICKNESS 5 PERCENT ASPECT RATIO 2.5 WEIGHT LAUNCHING 31.275 LB BURN-OUT 12.971 LB PROPELLANT 13.304 LB X-15 Design and DevelopmentOne of the more con­troversial features of the North American design was the fuse­lage tunnels that car­ried the propellant lines and engine con­trols around the full monocoque propellant tanks, shown in this 1956 sketch. Originally these tunnels extend­ed forward ahead of the cockpit, and the NACA worried they would create unac­ceptable vortices. (NASA)

ronment was provided by use of the Navy centrifuge at the Naval Air Development Center (NADC) Johnsville, Pennsylvania. Over 400 simulated reentries5 were flown during an initial round of tests completed on 12 July 1958; Iven Kincheloe, Joe Walker, Scott Crossfield, A1 White, Robert White, Neil Armstrong, and Jack McKay participat­ed. The primary objective of the program was to assess the pilot’s ability to make emergency reentries under high dynamic conditions following a failure of the stability augmentation system. The results were gen­erally encouraging.6

When the contracts with North American had been signed, the X-15 was some three years away from actual flight test. Although most of the basic research into materials and structural science had been completed, a great deal of work remained to be accom­plished. This included the development of fabrication and assembly techniques for Inconel X and the new hot-structure design. North American met the challenge of each problem with a practical solution, and even­tually some 2,000,000 engineering man­hours and 4,000 wind tunnel hours in 13 dif­ferent wind tunnels were logged.

The original North American proposal gave rise to several questions which prompted a meeting at Wright-Patterson AFB on 24-25 October 1955. Subsequent meetings were held at the North American Inglewood plant on 28-29 October and 14-15 November. Major discussion items included North American’s use of fuselage tunnels and all­moving horizontal stabilizers (the “rolling – tail”). The rolling-tail operated differentially to provide roll control, and symmetrically to provide pitch control; this allowed the elimi­nation of conventional ailerons. North American had gained considerable experi­ence with all-moving control surfaces on the YF-107A fighter. In this instance the use of differentially operated surfaces simplified the construction of the wing, and allowed elimination of protuberances that would have been necessary if aileron actuators had been incorporated in the thin wing. Such pro­tuberances would have disturbed the airflow and created another heating problem.

One other significant difference between the configuration of the NACA design and that of the actual X-15 stemmed from North American’s use of full-monocoque propel­lant tanks in the center fuselage and the use

The interior layout of the fuselage did not change much after the 1956 conference. Note the helium tank locat­ed in the middle of the LOX tank. The hydro­gen-peroxide (H202) was used to power the turbopump on the XLR99 rocket engine.

X-15 Design and Development(NASA)

of tunnels on both sides of the fuselage to accommodate the propellant lines and engine controls that ordinarily would have been contained within the fuselage. The NACA expressed concern that the tunnels might cre­ate undesirable vortices that would interfere with the vertical stabilizer, and suggested that the tunnels be kept as short as possible in the area ahead of the wing. North American agreed to make the investigation of the tunnels’ effects a subject of an early wind tunnel-model testing program.7

During the spring and summer of 1956, sev­eral scale models were exposed to rather intensive wind tunnel tests. A 1/50-scale – model was tested in the 11 – inch hypersonic and 9-inch blowdown tunnels at Langley, and another in a North American wind tun­nel. A 1/15-scale model was also tested at Langley and a rotary-derivative model was tested at Ames. The various wind tunnel pro­grams included investigations of the speed
brakes, horizontal stabilizers without dihe­dral, several possible locations for the hori­zontal stabilizer, modifications of the vertical stabilizer, the fuselage tunnels, and control effectiveness, particularly of the rolling-tail. Another subject in which there was consid­erable interest was determining the cross­section radii for the leading edges of the var­ious surfaces.

On 11 June 1956, North American received a production go-ahead for the three X-15 air­frames (although the first metal was not cut for the first aircraft until September). Four days later, on 15 June 1956, the Air Force assigned three serial numbers (56-6670 through 56-6672) to the X-15 program.8

By July, the NACA felt that sufficient progress had been made on the X-15 devel­opment to make an industry conference on the project worthwhile.8 The first Conference on the Progress of the X-15 Project was held



X-15 Design and DevelopmentSeven different wind tunnels are represent­ed in this chart show­ing how the extreme front of the fuselage tunnels began to be modified. Note the large speed brakes on the vertical stabilizer.

“LAC on the chart is the Langley Aeronautical Laboratory, while “AAL” is the Ames Aeronautical Laboratory.


at Langley on 25-26 October 1956. There were 313 attendees representing the Air Force, the NACA, Navy, various universities and colleges, and most of the major aero­space contractors. It was evident from the papers that a considerable amount of progress had already been made, but that a few significant problems still lay ahead.10

A comparison of the suggested configuration contained in the original NACA proposal and the North American configuration pre­sented to the industry conference revealed that the span of the X-15 had been reduced from 27.4 feet to only 22 feet and that the North American fuselage had grown from the suggested 47.5-foot overall-length to 49 feet. North American followed the NACA suggestion by selecting Inconel X as the major structural material and in the design of a multispar wing with extensive use of cor­rugated webs.11

One of the papers summarized the aerody­namic characteristics that had been obtained by tests in eight different wind tunnels.12 These tests had been made at Mach numbers ranging from less than 0.1 to about 6.9, and investigated such problems as the effects of speed brake deflection on drag, the lift-drag relationship of the entire aircraft, of individ­ual components such as the wings and fuse­lage tunnels, and of combinations of individ­ual components. One of the interesting prod­ucts was a finding that almost half of the total lift at high Mach numbers would be derived from the fuselage tunnels. Another
result was the confirmation of the NACA’s prediction that the original fuselage tunnels would cause longitudinal instability; for sub­sequent testing the tunnels had been short­ened in the area ahead of the wing, greatly reducing the instability. Still other wind tun­nel tests had been conducted in an effort to establish the effect of the vertical and hori­zontal tail surfaces on longitudinal, direc­tional, and lateral stability.

It should be noted that wind tunnel testing in the late 1950s was, and still is, an inexact sci­ence. For example, small (3- to 4-inch) mod­els of the X-15 were “flown” in the hyper­velocity free-flight facility at Ames. The models were made out of cast aluminum, cast bronze, or various plastics, and were actually fairly fragile. Despite this, the goal was to shoot the model out of a gun at tremendous speeds in order to observe shock wave patterns across the shape. As often as not, what researchers saw were pieces of X-15 models flying down the range side­ways, Fortunately, enough of the models remained intact to acquire meaningful data.13

Other papers presented at the industry con­ference dealt with research into the effect of the aircraft’s aerodynamic characteristics on the pilot’s control. Pilot-controlled simula­tion flights for the exit and reentry phases had been conducted; researchers reported that the pilots had found the early configura­tions nearly uncontrollable without damping, and that even with dampers the airplane pos­sessed only minimum stability during parts

These charts show the expected tempera­tures and skin thick­ness for various parts of the Х-15’s fuselage.

X-15 Design and DevelopmentNote the large differ­ence between top-side temperatures and those on the bottom of the fuselage. (NASA)



of the programmed flight plan. A program utilizing a free-flying model had proved low – speed stability and control to be adequate. Since some aerodynamicists had questioned North American’s use of the rolling-tail instead of ailerons, the free-flying model had also been used to investigate that feature. The results indicated that the rolling-tail would provide the necessary lateral control.

Several papers presented at the conference dealt with aerodynamic heating. One of these was a summary of the experience gained with the Bell X-1B and X-2. The information was incomplete and not fully applicable to the X-15, but it did provide a basis for compari­son with the results of the wind tunnel and analytical studies. Another paper dealt with the results of the structural temperature esti­mates that had been arrived at analytically. It was apparent from the contents of the papers that the engineers compiling them were con­fronted by a paradox—in order to attain an adequate and reasonably safe research vehi­cle, they had to foresee and compensate for the very aerodynamic heating problems that were to be explored by the completed aircraft.

In addition to the papers on the theoretical aspects of aerodynamic heating, a report was made on the structural design that had been accomplished at the time of the conference. Critical loads would be encountered during the accelerations at launch weight and during reentry into the atmosphere, but since maxi­mum temperatures would be encountered only during the latter, the paper was largely
confined to the results of the investigations of the load-temperature relationships that were anticipated for the reentry phase. The selection of Inconel X skin for the multispar box-beam wing was justified on the basis of the strength and favorable creep characteris­tics of that material at 1,200 degrees Fahrenheit. A milled bar of Inconel X was to be used for the leading edge since that por­tion of the wing acted as a heat sink. The internal structure of the wing was to be of titanium-alloy sheet and extrusion construc­tion. The front and rear spars were to be flat web-channel sections with the intermediate spars and ribs of corrugated titanium webs.

For purposes of the tests the maximum tem­perature differences between the upper and lower wing surfaces had been estimated to be 400 degrees Fahrenheit and that between the skin and the center of the spar as 960 degrees Fahrenheit. Laboratory tests indicated that such differences could be tolerated without any adverse effects on the structure. Other tests had proven that thermal stresses for the Inconel-titanium structure were less than those encountered in similar structures con­structed entirely on Inconel X, Full-scale tests had been made to determine the effects of temperature on the buckling and ultimate strength of a box beam. Simply heating the test structure produced no surface buckles. Compression buckles had appeared when ultimate loads were applied at normal tem­peratures but the buckles disappeared with the removal of the load. Tests at higher tem­peratures and involving large temperature





The wing of the X-15 was constructed from Inconel X skins over a titanium struc­ture. Unlike many air­craft, there was not a continuous spar across both wings. Instead, each wing was bolted to the fuselage. (NASA)




X-15 Design and Development

differences had finally led to the failure of the test box, but it seemed safe to conclude that . thermal stresses had very little effect on the ultimate strength of the box.”

Tests similar to those conducted on the wing structure had also been performed on the hor­izontal stabilizer. The planned stabilizer struc­ture differed from the wing in that it incorpo­rated a stainless steel spar about halfway between the leading and trailing edges, and an Inconel X spar three and one-half inches from the leading edge. The remainder of the inter­nal structure consisted of titanium compo­nents and the skin was Inconel X sheet. Tests of the stabilizer had indicated that a design which would prevent all skin buckling would be inordinately heavy, so engineers decided to tolerate temporary buckles. The proposed sta­bilizer had flutter characteristics that were within acceptable limits.

The front and rear fuselage were semimono – coque structures of titanium ribs, Inconel X outer skin, and an inner aluminum skin insu­lated with spun glass. The integral propellant tanks in the center fuselage were of full monocoque construction. The full mono – coque design used only slightly thicker skins than the semimonocoque design, possessed adequate heat sink properties, and reduced stresses caused by temperature differences by placing all of the material at the surface. It seemed, therefore, that the resulting structure was ideal for use as a pressure tank. The thickness of the monocoque walls would also make sealing easier and leaks less likely.

The fuselage side tunnels presented yet another problem. As the tunnels would pro­tect the side portions of the propellant tanks from aerodynamic heating, the sides would not expand as rapidly as the areas exposed to the air, and another undesirable compressive stress had to be anticipated. It was thought that beading the skin of the areas protected by the tunnels would provide a satisfactory solution, but beading introduced further complications by reducing the structure’s ability to carry pressure loads. Ultimately, however, the techniques proved successful.

Like most rocket engines of the period, the XLR99 would use liquid oxygen as an oxi­dizer, and a non-cryogenic fuel, in this case anhydrous ammonia.14 Each of the two main propellant tanks was to be divided into three compartments by curved bulkheads; the two compartments furthest from the aircraft cen­ter of gravity were equipped with slosh baf­fles. Plumbing was to be installed in a single compartment, the compartment sealed by a bulkhead, and the process repeated until all the compartments were completed. The tank ends were to be semicurved in shape to keep them as flat as possible, to reduce weight, and to permit thermal expansion of the tank shell. This entire structure was to be of weld­ed Inconel X.

The expected acceleration of the X-15 pre­sented several unique human factors concerns early in the program. It was expected that the pilot would be subjected to an acceleration of up to 5g. It seemed advisable to develop a

X-15 Design and Development

One of the innovations proposed by North American was the use of monocoque propel­lant tanks, leading to the use of the contro­versial fuselage tun­nels. The forward-most part of the LOX tank was equipped with slosh baffles. (NASA)

side-stick controller that would allow the pilot’s arm to be supported by an armrest while still allowing him of full control over the aircraft.15 Coupled with the fact that there were two separate attitude-control systems on the X-15, this resulted in a unique control stick arrangement. A conventional center stick, similar to that installed in most fighter – type aircraft of the era, was connected to the aerodynamic control surfaces through a sta­bility-augmentation (damper) system. A side – stick controller on the right console was con­nected to the same aerodynamic control sur­faces and augmentation system. Either stick could be used interchangeably, although the flight manual16 describes using the center stick “during normal periods of longitudinal and vertical acceleration.” The center stick was occasionally omitted from flights later in the flight research program based on pilot prefer­ences. Another side-stick controller on the left console operated the so-called “ballistic con­trol” system17 (thrusters) that provided attitude control at high altitudes. The flight manual warns that “velocity tends to sustain itself after the stick is returned to the neutral posi­tion. A subsequent stick movement opposite to the initial one is required to cancel the orig­inal attitude change.”

At the time of the industry conference in 1956, the design for the X-15 side controller had not been definitely established but a summary of the previous experience with such controllers was available. Experimental controllers had been installed on a Grumman F9F-2, Lockheed TV-2, Convair F-102, and on a simulator. The pilots who had tried side controllers had reported no difficulty in maneuvering, but they generally felt that greater efforts would have to be made to eliminate backlash and to control friction forces; they had also urged that efforts be made to give the side controllers a more “natural” feel.

Another problem which had not been thor­oughly explored at the time of the 1956 con ference concerned the proposed reaction con­trols that would be necessary for the X-15 as dynamic pressures decreased to the point where the aerodynamic controls would no longer be effective. Analog computer and ground simulator studies were then under way in an effort to determine the best relationship between the control thrust and the pilot’s movement of the control stick. Attempts were also being made to determine the amount of propellant that would be required for the reac-





The X-15 contained two side-stick con­trollers; one for the aerodynamic controls (shown), and one on the other console for the reaction controls. Although the side-stick proved very success­ful on the X-15, it would be another 20 years before one was installed on an opera­tional aircraft (the General Dynamics F-16). (NASA)















X-15 Design and Development

tion controls. No significant problems were uncovered during these early investigations, but it was clear that the pilot would have to give almost constant attention to such a con­trol system and that pilots should be given extensive practice on simulators before being allowed to attempt actual flight.

Some of the anticipated difficulties in the field of instrumentation arose because available strain gauges were not considered satisfactory at the expected high temperatures and because of difficulties in recording the output of ther­mocouples. Large structural deformations of wings and empennage were to be recorded by cameras in special camera compartments. Another instrumentation problem arose because the sensing of static pressure, ordi­narily difficult at high Mach numbers, was compounded in the case of the X-15 by heat­ing that would be too great for any conven­tional probe and by the low pressure at the high altitudes to be explored. The answer was to develop a stable-platform-integrating- accelerometer system to provide velocity, alti­tude, pitch, yaw, and roll angle information.

Still another instrumentation difficulty was created by the desirability of presenting the
pilot with angle-of-attack and side slip infor­mation, especially for the critical exit and reentry periods. Any device to furnish this information would have to be located ahead of the aircraft’s own flow disturbances, be structurally sound at elevated temperatures, accurate at low pressures, and cause a mini­ma) flow disturbance so as not to interfere with the heat transfer studies that were to be conducted in the forward area of the fuse­lage. These requirements had resulted in the design of a ball-nose16 capable of withstand­ing 1,200 degrees Fahrenheit. A six-inch diameter Inconel X sphere located in the extreme nose of the X-15 was gimbaled19 and servo-driven in two planes. It had five open­ings: a total-head port opening directly for­ward and two pairs of angle-sensing ports in the pitch and yaw planes, located at an angle of 30 to 40 degrees from the central port. Pitch and yaw could be sensed as pressure differences and these differences were con­verted into signals that would cause the ser­vos to realign the sphere in the relative wind.

Based largely on urgings from Scott Crossfield, the Air Force agreed to allow North American to design an ejection seat and to make a study justifying the selection

X-15 Design and DevelopmentAlthough the ejection seat showed at the 1956 industry confer­ence did not resemble the final unit used in the X-15s, the basic concepts remained the same. Restraining the pilot’s head, arms, and legs during ejec­tion at high dynamic pressures presented one of the major chal-. lenges to seat devel­opment. (NASA)

of a seat in preference to a capsule system.20 Two main criteria had governed the selec­tion of an escape system for the X-15, and these criteria were not necessarily comple­mentary. The first requirement was that the system be the most suitable that could be designed while remaining compatible with the airplane. The second was that no system would be selected that would delay the development of the X-15 or leave the pilot without any method of escape when the time arrived for flight research. The four possible escape systems that were consid­ered included cockpit capsules, nose cap­sules, a canopy shielded seat, and a stable – seat with a pressure-suit. An analysis of the expected flight hazards had indicated that because of the fuel exhaustion and low aerodynamic loads, the accident potential at peak speeds and altitudes was only about two percent of the total.

Capsule-like systems had been tried before, most notably in the X-2 where the entire for­ward fuselage could be detached from the rest of the aircraft. Model tests showed these to be very unstable and prone to tumble at a high rate of rotation. They also added a great deal of weight and complexity to the aircraft,21

The final decision for a stable-seat with a pressure-suit was made because most of the potential accidents could be expected to occur at speeds of Mach 4 or less, because system reliability always decreased with sys­tem complexity, and finally, because it was the system that imposed the smallest weight and size penalties upon the aircraft. The selected system would not function success­fully at altitudes above 120,000 feet or speeds in excess of Mach 4, but designers, particu­larly Scott Crossfield, held that the aircraft itself would offer the best protection in the areas of the performance envelope where the seat-suit combination was inadequate.

Cockpit and instrument cooling, pressuriza­tion, suit ventilation, windshield defogging, and fire protection were all to be provided from a liquid nitrogen supply. Vaporization of the liquid nitrogen would keep the pilot’s environment within comfortable limits at all times. An interesting aspect of the cooling problem was an estimate that only 1.5 per­cent of the system’s capacity would be applied to the pilot; the remaining 98.5 per­cent was required for equipment. Cockpit temperatures were to be limited to no more than 150 degrees Fahrenheit, the maximum

X-15 Design and DevelopmentThis chart shows that 92 percent of the expected X-15 acci­dents would happen below Mach 2 and

90,0 feet. This esti­mate supported Scott Crossfield’s request to use an ejection seat and pressure suit instead of a more complex escape cap­sule. (NASA)

limit for some of the equipment. The pilot would not be subjected to that temperature, however, as the pressure suit ventilation would enable him to select a comfortable temperature level. Cockpit pressure was to be maintained at the 35,000 foot level.

The effects of flight accelerations upon the pilot’s physiological condition and upon his ability to avoid inadvertent control move­ments had not been completely explored, but it was recognized that high accelera­tions could pose medical and restraint diffi­culties. In addition to the accelerations that would be encountered during the exit and reentry phases of the X-15’s flights, a very high acceleration of short duration would be produced during the landings. This was a result of the location of the main skids at the rear of the aircraft. Once the skids touched down, the entire aircraft would act as if it were hinged at the skid attachment points and the nose section would slam downward. Reproduction of this landing acceleration on simulators showed that because of the short duration, no real prob­lem existed. There were, however, numer­ous complaints about the severity of the jolts both in the simulator and once actual landings began.

The final paper presented to the 1956 indus­try conference was an excellent summary of the development effort and a review of the major problems that were known at that time. The author, Lawrence R Greene from North American, considered flutter to be an unsolved problem, primarily because of a lack of basic data on aero-thermal-elastic relationships and because little experimental data was available on flutter at hypersonic Mach numbers. He pointed out that available data on high-speed flutter had been derived from experiments conducted at Mach 3 or less, and that not all of the data obtained at those speeds were applicable to the problems faced by the designers of the X-15. As it turned out, panel flutter was encountered early in the flight test program, leading to a change in the design criteria for high-speed aircraft. Another difficulty was the newness of Inconel X as a structural material and the necessity of experimenting with fabrication techniques that would permit its use as the primary structural material for the X-15. Problems were also expected to arise in con­nection with sealing materials, most of which were known to react unfavorably when subjected to high temperature condi­tions.22 Although North American did encounter initial problems in using Inconel

X-15 Design and DevelopmentDespite its perform­ance potential, the basic cockpit design of the X-15 was quite conventional, with the exception of the side – stick controllers. The engine instrumenta­tion on the lower left of the instrument panel would be differ­ent for the XLR11 flights. The addition of the MH-96 in the X-15-3 would necessi­tate some changes in the instrumentation. See page 63 for a photo. (NASA)

X and titanium during the construction of the X-l5, it was able to work through the diffi­culties with no major delays.

A development engineering inspection was held at the North American Inglewood plant on 12-13 December 1956. This inspection of a full-scale tnockup was intended to reveal unsatisfactory design features before fabrication of the aircraft got under way. Thirty-four of the forty-nine individuals who participated in the inspection were rep­resentatives of the Air Force; twenty-two of them from WADC. The important role of the Air Force was also evident from the composition of the committee that would review the requests for alteration.2’ Major E. C. Freeman, of ARDC, served as committee chairman, Mr. F. Orazio of WADC and Lieutenant Colonel Keith G. Lindell of Air Force Headquarters were committee mem­bers, and Captain Chester E. McCollough, Jr. of the ARDC and Captain Iven C. Kincheloe, Jr. of the Air Force Flight Test Center (AFFTC) served as advisors. The Navy and the NACA each provided a single committee member; three additional advi­sors were drawn from the NACA.

The inspection committee considered 84 requests for alterations, rejected 12, and placed 22 in a category for further study. The majority of the 50 changes that were accept­ed were minor, such as the addition of longi­tudinal trim indications from the stick posi­tion and trim switches, relocation of the bat­tery switch, removal of landing gear warning lights, rearrangement and redesign of warn­ing lights, and improved markings for sever­al instruments and controls.

Some of the most interesting comments were rejected by the committee. For instance, the suggestions that the aerodynamic and reac­tion controller motions be made similar, that the reaction controls be made operable by the same controller used for the aerodynam­ic controls, or that a third controller combin­ing the functions of the aerodynamic and reaction controllers be added to the right console, were all rejected on the grounds that actual flight experience was needed with the controllers already selected before a decision could be made on worthwhile improvements or combinations. As two of the three sugges­tions on the controllers came from potential pilots of the X-l5 (Joseph A. Walker and

X-15 Design and DevelopmentThe vertical stabilizer was one of the most obvious changes between the industry conference configura­tion and the final vehi­cle. The first design did not use the exag­gerated wedge-shape of the final unit. It was also more traditional, using a fixed forward portion and a conven­tional appearing rud­der. The final version used an all-moving design. Note the rud­der splits to become speed brakes, much like the shuttle design 25 years later. (NASA)

Iven C. Kincheloe, Jr.24), it would appear that the planned controllers were not all that might have been desired.

A request that the pilot be provided with continuous information on the nose-wheel door position (loss of the door could produce severe structural damage) was rejected because the committee felt that the previous­ly approved suggestion for gear-up inspec­tion panels would make such information unnecessary. This particular item would come back to haunt the program during the flight research phase.

After the completion of the development engineering inspection, the X-15 airframe design changed only in relatively minor details. North American essentially built the X-15 described at the industry conference in October and inspected in mockup in December 1956. Continued wind tunnel test­ing resulted in some external modifications, particularly of the vertical stabilizer, and some weight changes occurred as plans became more definite. But while work on the airframe progressed smoothly, with few unexpected problems, the project as a whole did encounter difficulties, some of them seri­ous enough to threaten long delays. In fact, North American’s rapid preparation of draw­ings and production planning served to high­light the lack of progress on some of the components and subsystems that were essen­tial to the success of the program.

Papers Published

Not the least of the technological legacies of the X-15 consisted of the more than 765 technical documents produced in association with the program, including some 200 papers reporting on general research that the X-15 inspired. John Becker saw them as “confirmation of the massive stimulus and the focus provided by the [X-15] program.”51

Other Views

William Dana took time in 1987 to write a paper for the Society of Experimental Test Pilots pointing out some of the lessons learned from the X-15 program.52 Dana should know—he was the last pilot to fly the X-15. Two he cited were particularly appro­priate to the designers of the X-30 and X-33, although neither heeded the lessons. They are included here in their entirety:

The first lesson from the X-15 is: Make

it robust. As you have already seen, the

X-15 was able to survive some severe mistreatment during landings and still came back to fly another day. The X-15 that broke up after a spinning re-entry had self-recovered from the spin prior to break up, and might well have survived the entire episode had fixed, rather than self-adaptive, damper gains been used during re-entry. Another example exists of where the X-15 did survive a major stress in spite of operating with a major malfunction. This flight occurred in June 1967, when Pete Knight launched in X-15 No. 1 on a planned flight to

250,0 feet. At Mach 4 and at an alti­tude of 100,000 feet during the boost, the X-15 experienced a complete electri­cal failure that resulted in shutdown of both auxiliary power units and, there­fore loss of both hydraulic systems. Pete was eventually able to restart one of the auxiliary power units, but not its gener­ator. By skillful use of the one remain­ing hydraulic system and the ballistic controls, Pete was able to ride the X-15 to its peak altitude of 170 or 180,000 feet, reenter, make a 180 degree turn back to the dry lake at Tonopah, and dead-stick the X-15 onto the lakebed. All of these activities occurred without ever flowing another electron through the airplane from the time of the original failure.

There will be a hue and cry from some that the aerospace plane [X-30—NASP] cannot afford the luxury of robustness; that the aerospace plane, in order to be able to get to orbit, will have to be highly weight-efficient and will have to forego the strength and redundancy margins which allowed the X-15 to survive during adversity. And my answer to these people is: build your first aerospace plane with X-15 margins, even at the expense of per­formance; these margins will serve well while you are learning how to make your propulsion system operate and learning how to survive in the heating thicket of hypersonic flight. Someday, with this

knowledge in hand, it will be time to build a no-margins aerospace plane, but for now I suggest that you seize all the margins that you can because you will need them, as did the X-15.

The other lesson from the X-15 is: con­duct envelope expansion incrementally. The typical increment of speed increase for the original X-15 was about half a Mach number. With this increment it was easy to handle the heating damage that occurred in the original speed expansion phase. Again, I would expect to hear protest from the aerospace plane community, because when using one – half Mach number increments it is a long flight test program to Mach 25. Indeed, I cannot specify what size bite to take during the aerospace plane enve­lope expansion, but I can offer you the X-15A experience, in which two con­secutive flights carrying the dummy ramjet were flown to Mach numbers of 4.94 and 6.70. The former flight exhibit­ed no heat damage because of the wake of the dummy ramjet; the latter flight almost resulted in the loss of the aircraft due to heat damage.

Looking at the X-33 program in particular, another lesson jumps out. There will only be a single X-33. The building of three X-15s allowed the flight test program to proceed even after accidents. In fact, each of the X-15s was severely damaged at some time or another requiring it to be rebuilt. Plus, with multiple aircraft, it is possible to have one aircraft down for modification while the oth­ers continue to fly. And should one aircraft be lost, as sometimes happens in flight research, the program can continue. In today’s environment it is highly unlikely that the X-33 program would continue if it exploded during an engine test like the X-15-3 did while ground testing the XLR99. Hopefully the X-33 will not experience such a failure, but is that not part of the reason we conduct flight research—to learn from the failures as well as the successes?

The Engine

Those concerned with the success of the X-15 had to monitor the development of the aircraft itself, the XLR99 rocket engine, the auxiliary power units, an inertial system, a tracking range, a pressure suit, and an ejection seat. They had to make arrangements for support and B-36 carrier aircraft, ground equipment, the selection of pilots, and the development of simulators for pilot training. It was necessary to secure time on centrifuges, in wind tunnels, and on sled tracks. The ball-nose had to be developed, studies made of the compatibility of the X-15 and the carrier aircraft, and other studies on the possibility of extending the X-15 program beyond the goals originally contemplated. In addition to such tasks, funds to cover ever increasing costs had to be secured if the project were to have any chance of ultimate success, and at certain stages the effects of possibly harmful publicity had to be considered. With such multiplicity of tasks, it could be expected that several serious prob­lems would arise; not surprisingly, probably the most serious arose during the develop­ment of the XLR99.

Finding a suitable engine for the X-15 had been somewhat problematic from the earliest stages of the project, when the WADC Power Plant Laboratory had pointed out that the lack of an acceptable rocket engine was the major shortcoming of the NACA’s original propos­al. The laboratory did not believe that any available engine was entirely suitable for the X-15 and held that no matter what engine was accepted, a considerable amount of development work could be anticipated. Most of the possible engines were either too small or would need too long a development peri­od. In spite of these reservations, the labora­tory listed a number of engines worth consid­ering and drew up a statement of the require­ments for an engine that would be suitable for the proposed X-15 design. The laboratory also made clear its stand that the government should “… accept responsibility for develop­ment of the selected engine and… provide this engine to the airplane contractor as Government Furnished Equipment.”2′

The primary requirement for an X-15 engine, as outlined in 1954, was that it be capable of operating safely under all condi­tions. Service life would not have to be as long as for a production engine, but engi­neers hoped that the selected engine would not depart too far from production standards. The same attitude was taken toward reliabil­ity; the engine need not be as reliable as a production article, but it should approach such reliability as nearly as possible. There could be no altitude limitations for starting

or operating the engine, and the power plant would have to be entirely safe during start, operation, and shutdown, no matter what the altitude. The laboratory made it quite clear that a variable thrust engine capable of repeated restarts was essential.

The engine ultimately selected was not one of the four originally presented as possibil­ities by the Power Plant Laboratory. The ultimate selection was foreshadowed, how­ever, in discussions with Reaction Motors concerning the XLR10, during which atten­tion was drawn to what was termed "… a larger version of [the] Viking engine [XLR30].” In light of subsequent events, it was interesting to note that the laboratory thought26 the XLR30 could be developed into a suitable X-15 engine for less than $5,000,000 …” and with “ … approximate­ly two years’ work.”27

After North American had been selected as the winner of the X-15 competition, plans were instituted to procure the modified XLR30 engine that had been incorporated in the winning design. Late in October, Reaction Motors was notified that North American had won the X-15 competition and
that the winner had based his proposals upon the XLR30 engine.28

On 1 December 1955 a $1,000,000 letter con­tract was initiated with Reaction Motors for the development of a rocket engine for the X-15.-"’ Soon afterwards, a controversy devel­oped over the assignment of cognizance for the development of the engine. It began with a letter from Rear Adm. W. A. Schoech of the Bureau of Aeronautics. Adm. Schoech con­tended that since the XLR30-RM-2 rocket engine was the basis for the X-15 power plant, and the BuAer had already devoted three years to the development of that engine, it would be logical to assign the responsibili­ty for further development to the Navy. The admiral felt that retention of the program by the BuAer would expedite development, especially as the Navy could direct the devel­opment toward an X-15 engine by making specification changes rather than by negotiat­ing a new contract.30

The Navy’s bid for control of the engine development was rejected on 3 January 1956 on the grounds that the management respon­sibility should be vested in a single agency, that conflict of interest might generate delay,

The EngineThe XLR99 was an extremely compact engine, considering it was able to produce over 57,000 pounds – thrust. This was the first throttleable and restartable man-rated rocket engine. Many of the lessons-learned from this engine were incorporated into the Space Shuttle Main Engine developed 20 years later. (NASA)

and that BuAer was underestimating the time and effort necessary to make the XLR30 a satisfactory engine for piloted flight.

The Final Reaction Motors technical propos­al was received by the Power Plant Laboratory on 24 January, with the cost pro­posal following on 8 February.31 The cover letter from Reaction Motors promised deliv­ery of the First complete system “… within thirty (30) months after we are authorized to proceed.”32 Reaction Motors also estimated that the entire cost of the program would total $10,480,718.33 On 21 February the new engine was designated XLR99-RM-1.34

The 1956 industry conference heard two papers on the proposed engine and propul­sion system for the X-15. The XLR99-RM-1 would be able to vary its thrust from 19,200 to 57,200 pounds at 40,000 feet using anhy­drous ammonia and liquid oxygen (LOX)35 as propellants. Specific impulse was to vary from a minimum of 256 seconds to a maxi­mum of 276 seconds. The engine was to Fit into a space 71.7 inches long and 43.2 inch­es in diameter, have a dry weight of 618 pounds, and a wet weight of 748 pounds. A single thrust chamber was supplied by a
hydrogen-peroxide-driven turbopump, with the turbopump’s exhaust being recovered in the thrust chamber. Thrust control was by regulation of the turbopump speed.36

The use of ammonia as a propellant present­ed some potential problems; in addition to being toxic in high concentrations, ammonia is also corrosive to all copper-based metals. There were discussions early in the program between the Air Force, Reaction Motors, and the Lewis Research Center37 about the possi­bility of switching to a hydrocarbon fuel. It was finally concluded that changing fuel would add six months to the development schedule; it would be easier to learn to live with the ammonia.38 There is no documenta­tion that the ammonia ultimately presented any significant problems to the program.

The decision to control thrust by regulating the speed of the turbopump was made because the other possibilities (regulation by measurement of the pressure in the thrust chamber or of the pressure of the discharge) would cause the turbopump to speed up as pressure dropped. As the most likely cause of pressure drop would be cavitation in the pro­pellant system, an increase in turbopump



This 1956 sketch shows the controls and indicators for the XLR99. A different set of controls were used for the XLR11 flights, although they fit into the same space allo­cation. Notice the sim­ple throttle on the left console, underneath the reaction control side-stick (not shown).

The jettison controls took on particular sig­nificance on missions that had to be aborted prior to engine burn­out. (NASA)






















The Engine

speed would aggravate rather than correct the situation. Reaction Motors had also decided that varying the injection area was too complicated a method for attaining a variable thrust engine and had chosen to vary the injection pressure instead.

The regenerative cooling of the thrust cham­ber created another problem since the vari­able fuel flow of a throttleable engine meant that the system’s cooling capacity would also vary and that adequate cooling throughout the engine’s operating range would produce excessive cooling under some conditions. Engine compartment temperatures also had to be given more consideration than in previ­ous rocket engine designs because of the higher radiant heat transfer from the struc­ture of the X-15. Reaction Motors’ spokesman at the 1956 industry conference concluded that the development of the XLR99 was going to be a difficult task. Subsequent events were certainly to prove the validity of that prediction.

A second paper dealt with engine and acces­sory installation, the location of the propel­lant system components, and the engine con­trols and instruments. The main propellant tanks were to contain the LOX, ammonia, and the hydrogen peroxide. The LOX tank,
with a capacity of approximately 1,000 gal­lons, was located just ahead of the aircraft’s center of gravity; the 1,400 gallon ammonia tank was just aft of the same point. A center core tube within the LOX tank would pro­vide a location for a supply of helium under a pressure of 3,600 psi. Helium was used to pressurize both the LOX and ammonia tanks. A 75-gallon hydrogen peroxide tank behind the ammonia tank provided the monopropel­lant for the turbopump.

Provision was also made to top-off the LOX tank from a supply carried aboard the carrier aircraft; this was considered to be beneficial in two ways. The LOX supply in the carrier aircraft could be kept cooler than the oxygen already aboard the X-15, and the added LOX would permit cooling of the X-15’s own sup­ply by boil-off, without reduction of the quantity available for flight. The ammonia tank was not to be provided with a top-off arrangement, as the slight increase in fuel temperature during flight was not considered significant enough to justify the complica­tions such a system would have entailed.

On 10 July 1957, Reaction Motors advised the Air Force that an engine satisfying the contract specifications could not be devel­oped unless the government agreed to a nine-

The Engine

The XLR99 on a main­tenance stand. The engine used ammonia (NH3) as fuel and liq­uid oxygen (LOX) as the oxidizer. The XLR99 required a sep­arate propellant, hydro­gen peroxide, to drive its high-speed turbop­ump—the Space Shuttle Main Engine uses the propellant itself (LH2 or L02, as appropriate) to drive the turbopumps. (AFFTC via the Tony Landis Collection)

month schedule extension and a cost increase from $15,000,000 to $21,800,000. At the same time, Reaction Motors indicated that it could provide an engine that met the per­formance specification within the established schedule if permitted to increase the weight from 618 pounds to 836 pounds. The compa­ny estimated that this overweight engine could be provided for $17,100,000. The Air Force elected to pursue the heavier engine since it would be available sooner and have less impact on the overall X-l5 program.

Those who hoped that the overall perform­ance of the X-l5 would be maintained were encouraged by a report that the turbopump was more efficient than anticipated and would allow a 197 pound reduction in the amount of hydrogen peroxide necessary for its operation. This decrease, a lighter than expected airframe, and the increase in launch speeds and altitudes provided by a recent substitution of a B-52 as the carrier aircraft, offered some hope that the original X-l5 per­formance goals might still be achieved.39

Despite the relaxation of the weight require­ments, the engine program failed to proceed at a satisfactory pace. On 11 December 1957 Reaction Motors reported a new six-month slip. The threat to the entire X-l5 program posed by these new delays was a matter of serious concern, and on 7 January 1958, Reaction Motors was asked to furnish a detailed schedule and to propose means for solving the difficulties. The new schedule, which reached WADC in mid-January, indi­cated that the program would be delayed another five and one-half months and that costs would rise to $34,400,000—double the cost estimate of the previous July.10

In reaction, the Air Force recommended increasing the resources available to Reaction Motors and proposed the use of two 11 XLR11 rocket engines as an interim installation for the initial X-l5 flights. Additional funds to cover the increased effort were also approved, as was the estab­lishment of an advisory group.42

The threat that engine delays would serious­ly impair the value of the X-l5 program had generated a whole series of actions during the first half of 1958: personal visits by gen­eral officers to Reaction Motors, numerous conferences between the contractor and representatives of government agencies, increased support from the Propulsion Laboratory43 and the NACA, an increase in funds, and letters containing severe censure of the company’s conduct of the program. An emergency situation had been encountered, emergency remedies were used, and by mid­summer improvements began to be noted.

Engine progress continued to be reasonably satisfactory during the remainder of 1958. A destructive failure that occurred on 24 October was traced to components that had already been recognized as inadequate and were in the process of being redesigned. The failure, therefore, was not considered of major importance.41 A long-sought goal was finally reached on 18 April 1959 with completion of the Preliminary Flight Rating Test (PFRT). The flight rating program began at once.13

At the end of April, a “realistic” schedule for the remainder of the program showed that the Flight Rating Test would be completed by 1 September 1959. The first ground test engine was delivered to Edwards AFB at the end of May, and the first flight engine was delivered at the end of July.4*

A total of 10 flight engines were initially procured, along with six spare injector – chamber assemblies; one additional flight engine was subsequently procured. In January 1961, shortly after the first XLR99 test flight, only eight of these engines were available to the flight test program. There was still a number of problems with the engines that Reaction Motors was continuing to work on; the most serious being a vibra­tion at certain power levels, and a shorter than expected chamber life. There were four engines being used for continued ground tests, including two flight engines.47 Three of the engines were involved in tests to isolate

and eliminate the vibrations, while the fourth engine was being used to investigate extend­ing the life of the chamber.48

It is interesting to note that early in the pro­posal stage, North American determined that aerodynamic drag was not as important a design factor as was normally the case with jet-powered fighters. This was largely due to the amount of excess thrust available from the XLR99. Weight was considered a larger driver in the overall airplane design. Only about 10 percent of the total engine thrust was necessary to overcome drag, and anoth­er 20 percent to overcome weight. The remaining 70 percent of engine thrust was available to accelerate the X-l5.44

The New Millennium

As we enter the new millennium, it is inter­esting to note how the X-15 has shaped aero­nautics and astronautics. Indeed, when the X-33 program began during 1996, it was sur­prising to find that many of the younger con­tractor engineers were totally unaware of the X-15, and that most thought the SR-71 was the fastest aircraft that had ever flown, dis­counting the Space Shuttle. Interestingly, the young engineers at Dryden remembered the program, and when it came to setting up the instrumentation range (which extends all the way to the Dakotas), lessons learned from the X-15 High Range were used.53

The most obvious difference today has absolutely nothing to do with the technology of hypersonic flight. It is the political climate that surrounds any large project. The NASA Administrator, Daniel Goldin, told an X-33 all-hands meeting that it was “okay to fail”—a reference that many times in order to succeed, you first have to experience prob­lems that appear to be failures. But this is not the climate that actually exists. Any failure is often used as an excuse to cut back or cancel a project. In most cases the only way to total­ly avoid failure is to completely understand what you are doing; but if you completely understood something, there would be no point in building an X-plane!

The X-15 is usually regarded as the most suc­cessful flight research program ever undertak­en. But the program had its share of failures. The XLR99 destroyed the X-15-3 before it had even flown; but the aircraft was rebuilt and the XLR99 became a very successful research engine. On several occasions the X-15s made hard landings, sometimes hard enough to sig­nificantly damage the aircraft; each time they were rebuilt and flew again. Mike Adams was killed in a tragic accident; but less than four months later William Dana flew the next research flight. Yes, the X-15 failed often; but its successes were vastly greater.

Perhaps we have not learned well enough.

Other Systems

In early 1958, at the very height of the furor over the problems with the XLR99, a note of warning sounded for the General Electric auxiliary power unit (APU). On 26 March 1958 and again on 11 April 1958, General Electric notified North American of its inability to meet the original specifications in the time available, and requested approval
of new specifications. North American, with the concurrence of the Air Force, agreed to modify the requirements. The major changes involved an increase in weight from 40 to 48 pounds, an increase in start time from five to seven seconds, and a revision of the specific fuel consumption curves.50

By the end of the summer 1958, the APU seemed to have reached a more satisfactory state of development, and production units were ready for shipment.51 The early captive flights beginning in 1959 would reveal some additional problems, but investigation showed that the in-flight failures had occurred partial­ly because captive testing subjected the units to an abnormal operational sequence that would not be encountered during glide and powered flight. Some components were redesigned, but the APU would continue to be relatively troublesome in actual service.

During the course of the X-l 5 program, many concerns were voiced over the development of a pressure suit and an escape system. Although full-pressure suits had been studied during World War II, attempts to fabricate a practical garment had met with failure. The

Other Systems

Soule to Storms: “You have a little airplane and a big engine with a large thrust margin."

And indeed they did. The XLR99 provided

57.0 pounds-thrust to propel an aircraft that only weighed

30.0 pounds. Consider that the con­temporary F-104 Starfighter, considered something of a hot rod, weighed 20,000 pounds and its J79 only produced 15,000 pounds-thrust in full afterburner. (NASA)











20 30 40 50 60


80 90


Other Systems

Air Force took renewed interest in pressure suits in 1954 when it had become obvious that the increasing performance of aircraft was going to necessitate such a garment. The first result of the renewed interest was the creation of a suit that was heavy, bulky, and unwieldy; the garment had only limited mobility and various joints created painful pressure points. However, in 1955 the David Clark Company succeeded in producing a garment using a distorted-angle fabric that held some promise of ultimate success.51

Despite the early state-of-development of full-pressure suits, Scott Crossfield was con­vinced they were the way to go for X-l5. North American’s detail specifications of 2 March 1956 called for just such a garment— to be furnished by North American through a subcontract with the David Clark Company.55 A positive step toward Air Force acceptance of the idea occurred during a conference held at the North American plant on 20-22 June 1956. A full-pressure suit developed by the Navy was demonstrated during an inspection of the preliminary cockpit mockup, and although the suit still had a number of defi­ciencies, it was concluded that “… the state – of-the-art on full pressure suits should permit the development of such a suit satisfactory for use in the X-15.”54

After an extremely difficult and prolonged development process, Scott Crossfield received the first new MC-2 full-pressure suit on 17 December 1958 and, two days later, the suit successfully passed nitrogen contamina­tion tests at the Air Force Aero Medical Laboratory. The X-15 project officer attrib­uted much of the credit for the successful and timely qualification of the full-pressure suit to the intensive efforts of Crossfield.55

Fortunately, development did not stop there. On 27 July 1959, the Aero Medical Laboratory brought the first of the new A/P22S-2 pressure suits to Edwards. The consensus amongst the pilots was that it rep­resented a large improvement over the earli­er MC-2. It was more comfortable and pro­vided greater mobility; and it took only 5 minutes to put on, compared to 30 minutes for the MC-2. However, it would take anoth­er year before fully-qualified versions of the suit were delivered to the X-15 program.56

While not directly related to the pressure suit difficulties, the type of escape system to be used in the X-15 had been the subject of debate at an early stage of the program; the decision to use the stable-seat, full-pressure – suit combination had been a compromise based largely on the fact that the ejection seat was lighter and offered fewer complications than the other alternatives.

As early as 8 February 1955, the Aero Medical Laboratory had recommended a cap­sular escape system, but the laboratory had also admitted that such a system would prob­ably require extensive development. The sec­ond choice was a stable seat that incorporated limb retention features and one that would produce a minimum of deceleration.51 During meetings held in October and November 1955, it was agreed that North American would design an ejection seat for the X-15 and would also prepare a report justifying the use of such a system in preference to a capsule. North American was to incorporate head and limb restraints in the proposed seat.58

Despite the report, the Air Force was not completely convinced. At a meeting held at Wright Field on 2-3 May 1956, the Air Force again pointed out the limitations of ejection seats. In the opinion of one NACA engineer who attended the meeting, the Air Force was still strongly in favor of a capsule—partly because of the additional safety a capsule system would offer, and partly because the use of such a system in the X-15 would pro­vide an opportunity for further developmen­tal research. Primarily due to the efforts of Scott Crossfield, the participants finally agreed that because of the “time factor, weight, ignorance about proper capsule design, and the safety features being built into the airplane structure itself, the X-15 was probably its own best capsule.” About

the only result of the reluctance of the Air Force to endorse an ejection seat was a request that North American yet again docu­ment the arguments for the seat.59

The death of Captain Milbum G. Apt in the crash of the Bell X-2, which had been equipped with an escape capsule, in September 1956 renewed apprehension as to the adequacy of the X-15’s escape system.® By this time, however, it was acknowledged that no substantive changes could be made to the X-15 design. Fortunately, North American’s seat development efforts were generally proceeding well.’’1

Sled tests of the ejection seat began early in 1958 at Edwards with the preliminary tests concluded on 22 April. Because of the high cost of sled runs, the X-15 project office advised North American to eliminate the planned incremental testing and to conduct the tests at just two pressure levels—125 pounds per square foot and 1,500 pounds per square foot. The X-15 office felt that suc­cessful tests at these two levels would fur­nish adequate proof of seat reliability at intermediate pressures.62

Between 4 June 1958 and 3 March 1959, the X-15 seat completed its series of sled tests. Various problems, with both the seat and the sled, had been encountered, but all had been worked through to the satisfaction of North American and the Air Force. The X-15 seat was cleared for flight use.62

Another item for which the Air Force retained direct responsibility was the all-attitude iner­tial flight data system. It was realized from the beginning of the X-15 program that the air­plane’s performance would necessitate a new means of determining altitude, speed, and air­craft attitude. This was because the traditional use of static pressure as a reference would be largely impossible at the speeds and altitudes the X-15 would achieve; moreover, the tem­peratures encountered would rule out the use of tradition pitot tube sensing devices. The NACA had proposed a “stable-platform iner­tial-integrating and attitude sensing unit” as the means of meeting these needs.64 A series of miscommunications resulted in the NACA assuming the Air Force had already developed a satisfactory unit and would provide it to the X-15 program.65 After it was discovered that a suitable unit did not exist, emergency efforts were undertaken to develop one without impacting the X-15 program. After a consid­erable amount of controversy, a sole-source contract was awarded to the Sperry Gyroscope Company on 5 June 1957 for the development and manufacture of the stable – platform.66 The cost-plus-fixed-fee contract, signed on 5 June 1957, was for $1,213,518.06 with a fixed fee of $85,000.67

In April 1958, the Air Force concluded that the scheduled delivery of the initial Sperry unit in December would not permit adequate testing to be performed prior to the first flights of the X-15. Consequently a less capa­ble interim gyroscopic system was installed in the first two aircraft and the final Sperry system was installed in the last X-15.68

By the end of 1958, the two major system components (the stabilizer and the computer) were completed and ready to be tested as a complete unit. The systems were shipped to Edwards in late January 1959, and during the spring of 1959 plans were made to use the NB-52 carrier aircraft as a test vehicle.69 In addition, arrangements were made to test the stable-platform in a KC-97 that was already in use as a test aircraft in connection with the B-58 program.™ The first test flights in the KC-97 were carried out in late April.71 By June, North American had successfully installed the Sperry system in the third X-15 22 In January 1961, wiring was installed in the NB-52B to allow the stable-platform to be installed in a pod carried on the pylon under the wing. The first complete stable – platform system installed in the B-52 pod was flown on 1 March 1961, Since the B-52 was capable of greater speeds and higher altitudes than the KC-97, it provided addi­tional data to assist Sperry in resolving prob­lems being encountered with the unit.7’

Resolution Adopted by NACA Committee on Aerodynamics, 5 October 1954 —

Подпись: This resolution was the official beginnings of the X-15 research airplane program.RESOLUTION ADOPTED HI NACA

WHEREAS, The necessity of maintaining supremacy In the «dr continues to place great urgency on solving the problems of flight with man-carrying aircraft at greater speeds and extreme altitudes, and

WHEREAS, Propulsion systems are now capable of propelling such aircraft to speeds and altitudes that impose entirely new and unexplored aircraft design problems, and

WHEREAS, It now appears feasible to construct a research airplane capable of initial exploration of these problems,

3E IT HEREBY RESOLVED, That the Committee on Aerodynamics endorses the proposal of the 1mediate initiation of a project to design and construct a research airplane capable of achieving speeds of the order of Mach Number 7 and altitudes of several hundred thousard feet for the exploration of the problems of stability and control of manned aircraft end aerodynamic heating in the severe form asaociated with flight at extreme speeds and altitudes.

The High Range

Previous rocket aircraft, such as the X-l and X-2, had been able to conduct the majority of their flight research in the skies directly over the Edwards test areas. The capabilities of the X-l5, however, would use vastly more air­space. The proposed trajectories required an essentially straight flight corridor equipped with multiple tracking, telemetry, and com­munications sites, as well as the need for suit­able emergency landing areas. This led to con­struction of the X-l5 High Range extending from Wendover, Utah, to Edwards AFB. Radar and telemetry stations were installed at Ely and Beatty, Nevada, as well as Edwards. Telemetry from the X-l 5, as well as voice communications, were received, recorded, and forwarded to Edwards by the stations at Ely and Beatty. Each of these stations was also manned by a person to back up the prime “communicator” (NASA 1) at Edwards in
case the communication links went down. Each ground station overlapped the next, and they were interconnected via microwave and land-line so that timing signals, voice com­munication, and radar data would be available to all. Provisions were made for recording the acquired data on tape and film, although some of the data was directly displayed on strip and plotting charts. The design and construction of the range was accomplished by Electronic Engineering Company of Los Angeles under contract with the Air Force.74 North American and the NACA also conducted numerous evaluations of various dry lakes to determine which were suitable for emergency landings along the route (see the summary included as an appendix to this monograph).

Carrier Aircraft

The group at Langley had sized their X-l5 proposal around the potential of using a

The use of a B-36 car­rier aircraft would have allowed the pilot to exit the aircraft while in transit to the drop area, or in case of emergency. However, personnel at the FRC worried that the B-36 would not be supportable since it was being phased out of active service. In the end, the B-52 pro­vided much better per­formance and was ultimately selected.

The High Range(AFFTC History Office)



Convair B-36 as the carrier aircraft. This was a natural extension of previous X-planes that had used a Boeing B-29 or B-50 as a carrier. The B-36 would be modified to carry the X-15 partially enclosed in its bomb bays, much like the X-l and X-2 had been in earlier projects. This arrangement had some advantages; the pilot could freely move between the X-15 and B-36 during climb-out and the cruise to the launch location. This was extremely advanta­geous if problems developed that required jet­tisoning the X-15 prior to launch. At the time of the first industry conference in 1956, it was expected that a B-36 would be modified begin­ning in the middle of 1957 and be ready for flight tests in October 1958.75

As the weight of the X-15 and its subsystems grew, however, the Air Force and NASA began to look for ways to recover some of the lost performance. One way was to launch the X-15 at a higher altitude and greater speed. In addition, the personnel at Edwards believed that the ten-engine B-36 would be difficult to maintain7" since it was being phased out of the Air Force inventory. Investigations showed that the X-15, as designed, would fit under the wing of one of the new Boeing B-52 Stratofortresses; the configuration of the B-52 . precluded carrying the X-15 in the bomb bay. This was not the ideal solution—the X-15 pilot would have to be locked in the research airplane prior to takeoff, and the large weight transition when the X-15 was released would provide some interesting control problems for the B-52. Further analysis concluded that the potential problems were solvable, and that the increase in speed and altitude capabilities were desirable. Fortunately, two early B-52s were completing their test duties, and the Air Force made them available to the program.

On 29 November 1957, the B-52A (52-003) arrived at Air Force Plant 42 in Palmdale, California, after a flight from the Boeing plant in Seattle. The aircraft was placed in storage pending modifications. On 4 February 1958, the B-52A was moved into the North American hanger at Plant 42 and modified with a large pylon under the wing, the capa­
bility to monitor to the X-15, and a system to replenish the X-15 LOX supply. The aircraft, now designated77 NB-52A, was flown to Edwards AFB on 14 November 1958; it was later named “The High and the Mighty One.” The Air Force also supplied a B-52B (52-008) that arrived in Palmdale for similar modifica­tions on 5 January 1959, and was flown, as an NB-52B, to Edwards on 8 June 1959.

Roll Out

As the first X-15 was being completed, the NACA held the second X-15 industry con­ference in Los Angeles on 28-29 July 1958. North American began the conference with a paper detailing the developmental status of the aircraft. Twenty-seven other papers cov­ered subjects such as stability and control, simulator testing, pilot considerations, mis­sion instrumentation, thermodynamics, structures, materials and fabrication. There were approximately 550 attendees,78

On 1 October 1958, High-Speed Flight Station employees Doll Matay and John Hedgepeth put up a ladder in front of the sta­tion building at the foot of Lilly Avenue and took down the winged-shield NACA emblem from over the entrance door. NASA had arrived in the desert, bringing with it a new era of space-consciousness, soaring budgets, and publicity. The old NACA days of concentra­tion on aeronautics, and especially aerody­namics, were gone forever, as was the agency itself. On this day, the National Aeronautics and Space Administration was created.79

The X-15 construction process eventually consumed just over two years, and on 15 October 1958, the first aircraft (56-6670) was rolled out. Following conclusion of the official ceremonies, it was moved back inside and prepared for shipment to Edwards. On the night of 16 October, cov­ered completely in protective heavy-duty wrapping paper, it was shipped by truck to Edwards for initial ground test work.


The first of three let­ters attached to the Memorandum of Understanding that created the X-15 research program. Since it was nominally an Air Force program, the Air Force began the signature process.






NOV 9 1954


The early 1950s was an era where carbon paper and onion-skin copies were kept. Forty-five years later they are not repro­ducible, so the three letters have been recreated.



SUBJECT: Principles for the Conduct of a Joint Project for a Hew

High Speed Research Airplane

1. The Air Force concurs in the establishment of a joint NACA- Navy-Air Force project to design and construct a research airplane capable of achieving speeds of the order of Mach Number 7 and altitudes of several hundred thousand feet.

2. Attached is a Memorandum of Understanding. signed in tripli­cate by the Air Force, setting forth the principles fox the conduct by the NACA, the Navy, and the Air Force of this joint project. It 1b reguested that the Navy sign this Memorandum, in triplicate, and forward the signed copies to the Director of the NACA for signature and distribution bach to the signatory agencies.

3. The Air Force is designating Brigadier General В, В. Кеївеу, Deputy Director of Research and Development, as the Air Force representa­tive on the ‘Research Airplane Oonmittee’ referred to in paragraph В

of the Memorandum of understanding.


The letters remained SECRET until 3 July 1963 when they were downgraded to CONFIDENTIAL.

It was not until 9 November 1966 that they were finally declassified.



Trevor Gardner
Special Assistant (Rid))



Memo of understanding

w/1 incl fin trip)


The Flight Research Program

During the ten years of flight operations, five major aircraft were involved in the X-15 flight research program. The three X-15s were des­ignated X-15-1 (56-6670), X-15-2 (56-6671), and X-15-3 (56-6672). Early in the test pro­gram the first two X-15s were essentially iden­tical in configuration; the third aircraft was completed with different electronic and flight control systems. When the second aircraft was extensively modified after an accident mid­way through the test program, it became the X-15A-2. The two carrier aircraft were an NB-52A (52-003) and an NB-52B (52-008); they were essentially interchangeable.

The program used a three-part designation for each flight. The first number represented the specific X-15; 1 was for X-15-1, etc. No differentiation was made between the origi­nal X-15-2 and the modified X-15A-2. The second position was the flight number for that specific X-15. This included free-flights only, not captive-carries or aborts; the first flight was 1, the second 2, etc. If the flight was a scheduled captive-carry, the second position in the designation was replaced with a C; if it was an aborted free-flight attempt, it was replaced with an A. The third position was the total number of times any X-15 had been carried aloft by either NB-52. This
number incremented for each captive-carry, abort, and actual release. The 24 May 1960 letter from FRC Director Paul Bikle estab­lishing this system is included as an appen­dix to this monograph.

Initial Flight Tests

The X-15-1 arrived at the Air Force Flight Test Center at Edwards AFB, California, on 17 October 1958; trucked over the hills from the North American plant in Los Angeles for testing at the NASA High-Speed Flight Station. It was joined by the second airplane in April 1959; the third would arrive later. In contrast to the relative secrecy that had attended flight tests with the XS-1 (X-l) a decade before, the X-15 program offered the spectacle of pure theater.1

As part of the X-15’s contractor program, North American had to demonstrate each air­craft’s general airworthiness during flights above Mach 2 before delivering it to the Air Force, which would then tum it over to NASA. Anything beyond Mach 3 was con­sidered a part of the government’s research obligation. The contractor program would last approximately two years, from mid – 1959 through mid-1960.

Two different mission profiles were flown— one for maximum speed; and one for maximum altitude.


The Flight Research ProgramThe first X-15 (56-6670) immediately prior to the official roll­out ceremonies at North American’s Los Angeles plant on 15 October 1958. The small size of the trapezoid-shaped wings and the extreme wedge sec­tion of the vertical sta­bilizer are noteworthy. (North American Aviation)

The task of flying the X-15 during the con­tractor program rested in the capable hands of Scott Crossfield. After various ground checks, the X-15-1 was mated to the NB-52A, then more ground tests were con­ducted. On 10 March 1959, the pair made a scheduled captive-carry flight (program flight number 1-C-l). They had a gross take­off weight of 258,000 pounds, lifting off at 168 knots after a ground roll of 6,200 feet. During the 1 hour and 8 minute flight it was found that the NB-52 was an excellent carri­er for the X-15, as had been expected from numerous wind tunnel and simulator tests. During the captive flight the X-15 flight con­trols were exercised and airspeed data from the flight test boom on the nose was obtained in order to calibrate the instrumentation. The penalties imposed by the X-15 on the NB-52 flight characteristics was found to be minimal in the gear-up configuration. The mated pair was flown up to Mach 0.83 at 45,000 feet.2

The next step was to release the X-15 from the NB-52 in order to ascertain its gliding and landing characteristics. The first glide flight was scheduled for 1 April 1959, but was aborted when the X-15 radio failed. The
pair spent 1 hour and 45 minutes airborne conducting further tests in the mated config­uration. A second attempt was aborted on 10 April 1959 by a combination of radio failure and APU problems. Yet a third attempt was aborted on 21 May 1959 when the X-15’s stability augmentation system failed, and a bearing in the No. 1 APU overheated after approximately 29 minutes of operation.

The problems with the APU were the most disturbing. Various valve malfunctions, leaks, and several APU speed-control prob­lems were encountered during these three flights, all of which would have been unac­ceptable during research flights. Tests con­ducted on the APU revealed that extremely high surge pressures were occurring at the pressure relief valve (actually a blow-out plug) during initial peroxide tank pressuriza­tion. The installation of an orifice in the heli­um pressurization line immediately down­stream of the shut-off valve reduced the surges to acceptable levels. Other problems were found to be unique to the captive-carry flights and the long-run times being imposed on the APUs; they were deemed to be of lit­tle consequence to the flight test program

Long before the NB-52 first carried the X-15 into the air, engineers had tested the separation charac­teristics in the wind tunnels at Langley and Ames. Here an X-15 model drops – away from a model of the NB-52. Note that the X-15 is mounted on the wrong wing. This was necessary because the viewing area of the wind tun­nel was on the left side of the aircraft.

The Flight Research Program(NASA photo EL-1996-00114)

since the operating scenario would be differ­ent. The APUs underwent a constant set of minor improvements early in the flight test program, but nevertheless continued to be a source of irritation until the end.

On 22 May the first ground run of the inter­im XLR11 engine installation was accom­plished using the X-15-2. Scott Crossfield was in the cockpit, and the test was consid­ered successful, clearing the way for the eventual first powered flight; if the first X-15 could ever make its scheduled glide flight.

Another attempt at the glide flight was made on 5 June 1959 but was aborted even before the NB-52 left the ground3 when Crossfield reported smoke in the X-15-1 cockpit. Investigation showed that a cockpit ventila­tion fan motor had overheated.

Finally, at 08:38 on 8 June 1959, Scott Crossfield separated the X-15-1 from the NB-52A at Mach 0.79 and 37,500 feet. Just prior to launch the pitch damper failed, but Crossfield elected to proceed with the flight, and switched the SAS pitch channel to stand­by. At launch, the X-15 separated cleanly and Crossfield rolled to the right with a bank
angle of about 30 degrees. The X-15 touched down on the dry lake at Edwards 4 minutes and 56 seconds later. Just prior to landing, the X-15 began a series of increasingly wild pitching motions; mostly as a result of Crossfteld’s instinctive corrective action, the airplane landed safely. Landing speed was 150 knots, and the X-15 rolled-out 3,900 feet while turning very slightly to the right. North American subsequently modified the control system boost to increase the control rate response, effectively solving the problem.

Although the impact at landing was not con­sidered to be particularly hard, later inspec­tion revealed that bell cranks in both main landing skids had bent slightly. The main skids were not instrumented on this flight, so no specific impact data could be ascertained, but it was generally believed that the shock struts had bottomed and remained bottomed as a result of higher than predicted landing loads. As a precaution against the main skid problem occurring again, the metering char­acteristics of the shock struts were improved, and lakebed drop tests at higher than previ­ous loads were made with the landing gear test trailer that had been used to qualify the landing gear design. All other airplane sys-

The Flight Research ProgramNorth American test pilot A. Scott Crossfield was responsible for demonstrating that the X-15 was airworthy.

His decision to leave NACA and join North American effectively locked him out of the high-speed and high – altitude test flights later in the program. (NASA photo EC-570-1

terns operated satisfactorily, clearing the way for the first powered flight.4

In preparation for the first powered flight, the X-15-2 was taken for a captive-carry flight with full propellant tanks on 24 July 1959. During August and early September, several attempts to make the first powered flight were cancelled before leaving the ground due to leaks in the APU peroxide system and hydraulic leaks. There were also several failures of propellant tank pressure regulators. Engineers and technicians worked to eliminate these problems, all of which were irritating, but none of which was considered critical.

The first powered flight was made by X-15-2 on 17 September 1959. The aircraft was released from the NB-52A at 08:08 in the morning while flying at Mach 0.80 and 37,600 feet. Crossfield piloted the X-15-2 to Mach 2.11 and 52,341 feet during 224.3 sec­onds of powered flight using the two XLR11 engines. He landed on the dry lakebed at Edwards 9 minutes and 11 seconds after launch. Following the landing, a fire was noticed in the area around the ventral stabiliz­er, and was quickly extinguished by ground
crews. A subsequent investigation revealed that the upper XLR11 fuel pump diffuser case had cracked after engine shutdown and had sprayed fuel throughout the engine compart­ment. Fuel collected in the ventral stabilizer and was ignited by unknown causes during landing. No appreciable damage was done, and the aircraft was quickly repaired.5

The third flight of X-15-2 took place on 5 November 1959 when the X-15 was dropped from the NB-52A at Mach 0.82 and 44,000 feet. During the engine start sequence, one chamber in the lower engine exploded. There was external damage around the engine and base plate, plus quite a bit of damage internal to the engine compartment. The resulting fire convinced Crossfield to make an emergency landing at Rosamond Dry Lake; he quickly shut off the engines, dumped the remaining fuel, and jettisoned the ventral6 rudder. Even so, within the 13.9 seconds of powered flight, the X-15 managed to accelerate to Mach 1. The aircraft touched down near the center of the lake at approximately 150 knots and an 11 degrees angle of attack. When the nose gear bottomed out, the fuselage literal­ly broke in half at station’ 226.8, with about 70 percent of the bolts at the manufacturing

Any landing you can walk away from…

The Flight Research ProgramThe X-15-2 made a hard landing on 5 November 1959, breaking its back as the nose settled on the lakebed. The dam­age looked worse than it was, and the aircraft was back in the air three months later. (NASA photo E-9543)

joint being sheared out. The fuselage contact­ed the ground and was dragged for approxi­mately 1,500 feet. Crossfield later stated that the damage was the result of a defect that should have broken on the first flight.8 The aircraft was sent to the North American plant for repairs, and was subsequently returned to Edwards in time for its fourth flight on 11 February I960.9

The X-15-1 made its first powered flight, using two XLRlls, on 23 January 1960. This was also the first flight using the stable platform, and the performance of the system was considered encouraging. Under the terms of the contract, the X-15 had still “belonged” to North American until they had demonstrated its basic airworthiness and operation. Following this flight, a pre-deliv­ery inspection was accomplished, and on 3 February 1960 the airplane was formally accepted by the Air Force and subsequently turned over to NASA.

The first government X-15 flight (1-3-8) was on 25 March 1960 with NASA test pilot Joseph A. Walker at the controls. The X-15-1 was launched at Mach 0.82 and 45,500 feet, although the stable platform had malfunc­tioned just prior to release. Two restarts were required on the top engine before all eight chambers were firing, and the flight lasted just over 9 minutes, reaching Mach 2.0 and 48,630 feet. For the next six months, Walker and Major Robert M. White alternated flying the X-15-1.10

It is interesting to note that the predictions regarding flutter made by Lawrence P. Greene at the first industry conference in 1956 did materialize, although fortunately they were not major and relatively easy to correct. During the early test flights, vibrations at 110 cycles had been noted and were the cause of some concern. Engineers at FRC added instrumentation to the X-15s from flight to flight in an attempt to isolate the vibrations and understand their origins, while wind tun­nel tests were conducted at Langley. It was finally determined that the vibrations were being caused by a flutter of the fuselage side tunnel panels. These had been constructed in removable sections with an unsupported length of over 6 feet in some cases." North American added longitudinal stiffeners along the underside of each panel, and this largely cured the problem.12

The X-15-1 flew three times in the two weeks between 4 August and 19 August 1960, with five aborted launches due to various problems (including persistent APU failures). Two of these flights were made by Joe Walker, and one by Bob White. The flight on 12 August was to an altitude of 136,500 feet, marking the highest flight of an XLR11 – powered X-15.

The Million Horsepower Engine13

The X-15-3 had arrived at Edwards on 29 June 1959 but had not yet flown when the first XLR99 flight engine (s/n 105) was installed in it during early 1960. It should be noted that the third X-15 was never equipped for the XLR11 engines. At the same time, the second X-15 was removed from flight status after its ninth flight (2-9-18) on 26 April 1960, in anticipation of replacing the XLR11 engines with the new XLR99. This left only the X-15-1 on active flight status.

The first ground run with the XLR99 in the X-15-3 was made on 2 June 1960. Inspection of the aircraft afterward revealed damage to the liquid oxygen inlet line brackets, the result of a water-hammer effect. After repairs were completed, another ground run was conducted on 8 June. A normal engine start and a short run at minimal power was made, followed by a normal shutdown, A restart was attempted, but was shutdown automati­cally by a malfunction indication. Almost immediately, a second restart was attempted, resulting in an explosion that effectively destroyed the aircraft aft of the wing. Crossfield was in the cockpit, which was thrown 30 feet forward, but he was not injured. Subsequent investigation revealed that the ammonia tank pressure regulator had failed open. Because of some ground han-

The top and bottom of the fuselage were usually covered in frost because the LOX tank was integral with the fuselage. Oxygen is liquid at -297 degrees Fahrenheit.

The Million Horsepower Engine13All three X-15s nor­mally carried a yellow NASA banner on their vertical stabilizers. (U. S. Air Force)

dling hoses attached to the fuel vent line, the fuel pressure-relief valve did not operate properly, thus allowing the fuel tank to over­pressurize and rupture. Tn the process, the peroxide tank was damaged by debris, and the mixing of the peroxide and ammonia caused an explosion.

Post-accident analysis indicated that there were no serious design flaws with either the XLR99 or the X-15. The accident had been caused by a simple failure of the pressure reg­ulator, exasperated by the unique configura­tion required for the ground test. Modification of the X-15-2 to accept the XLR99 continued, and several other modifications were incorpo­rated at the same time. These included a revised vent system in the fuel tanks as an additional precaution against another explo­sion; revised ballistic control system compo­nents; and provisions for the installation of the ball-nose instead of the flight test boom that had been used so far in the program. The remains of the X-15-3 were returned to North American, which received authorization to rebuild the aircraft in early August.14

The installation of the ball-nose presented its own challenges since it had no capability to determine airspeed. The X-15 was designed with an alternate airspeed probe just forward
of the cockpit, although two other locations, one well forward on the bottom centerline of the aircraft, and one somewhat aft near the centerline, had been considered alternate locations. Several early flights compared the data available from each location, while rely­ing on the data provided by the airspeed sen­sors on the flight test boom protruding from the extreme nose. This indicated that the data from all three locations were acceptable, so the original location was retained. After the ball-nose was installed, angle-of-attack data was compared to that from previous flights using the flight test boom; the data were gen­erally in good agreement, clearing the way for operational use of the ball-nose.

The first flight attempt of X-15-2 with the XLR99 was made on 13 October 1960, but was terminated prior to launch because of a peroxide leak in the No. 2 APU, Just to show haw many things could go wrong on a single flight, there was also propellant impingement on the aft fuselage during the prime cycle, manifold pressure fluctuations during engine turbopump operation, and fuel tank pressure fluctuations during the jettison cycle. Nevertheless, two weeks later, Crossfield again entered the cockpit with the goal of making the first XLR99 flight. Again, prob­lems with the No. 2 APU forced an abort.

On 15 November 1960, everything went right, and Crossfield made the first flight of X-15-2 powered by the XLR99. The primary flight objective was to demonstrate engine operation at 50 percent thrust. The launch was at 46,000 feet and Mach 0.83, and even with only half the available power, the X-15 managed to climb to 81,200 feet and Mach 2.97. The sec­ond XLR99 flight tested the engine’s restart and throttling capability. Crossfield made the flight on 22 November, again using the sec­ond X-15. The third and final XLR99 demonstration flight was accomplished using X-15-2 on 6 December 1960. The objectives of engine throttling, shutdown, and restart were successfully accomplished. This marked North American Aviation’s, and Scott Crossfield’s, last X-15 flight. The job of fly­ing the X-15 was now totally in the hands of the government test pilots.15

After this flight, a work schedule was estab­lished which would permit an early flight with a government pilot using North American maintenance personnel. The flight was tentatively scheduled for 21 December i960 with Bob White as the pilot. However, a considerable amount of work had to be accomplished before the flight, including the
removal and replacement of the engine (s/n 103) which had suffered excessive chamber coating loss, installation of redesigned canopy hooks, installation of an unrestricted upper vertical stabilizer, rearrangement of the alternate airspeed system, and the reloca­tion of the ammonia tank helium pressure regulator into the fixed portion of the upper vertical. During a preflight ground run, a pinhole leak was found in the chamber throat of the engine. Although the leak was found to be acceptable for an engine run, it became increasingly worse during the test until it was such that the engine could not be run again. Since there was no spare engine avail­able, the flight was cancelled and a schedule established to deliver the aircraft to the gov­ernment prior to another flight. The X-15-2 was formally delivered to the Air Force and turned over to NASA on 8 February 1961. On the same day, X-15-1 was returned to the North American plant for conversion to the XLR99, having completed the last XLR11 flight of the program the day before with White at the controls.16

From the beginning of the X-15 flight test program in 1959 until the end of 1960, a total of 31 flights had been made with the first two

The Million Horsepower Engine13Six of the X-15 pilots {from left to right): Lieutenant Colonel Robert A. Rushworth (USAF), John B.

“Jack” McKay (NASA), Lieutenant Commander Forrest S. Petersen (USN), Joseph A. Walker (NASA), Neil A. Armstrong (NASA), Major Robert M. White (USAF). (NASA via the San Diego Aerospace Museum Collection)

X-15s by seven pilots. But the X-15-1 was experiencing an odd problem. When the APU was started, hydraulic pressure was either slow in coming up, or dropped off out of limits when the control surfaces were moved. The solution to the problem came after additional instrumentation was placed on the hydraulic system. The boot-strap line which pressurized the hydraulic reservoir was freezing, causing a flow restriction or stoppage. Under these conditions, the hydraulic pump would cavitate, resulting in little or no pressure rise. The apparent cause of this problem was the addition of a liquid nitrogen line to cool the stable platform. Since the nitrogen line was installed adjacent to the hydraulic lines, it caused the Orinite hydraulic oil to freeze. The solution to the problem was to add electric heaters to the affected hydraulic lines.

Joe Walker’s flight on 30 March 1961 marked the first use of the new A/P-22S full-pressure suit instead of the earlier MC-2. Walker reported the suit was much more comfortable and afforded better vision. But the flight pointed out a potential problem with the stability augmentation system (SAS). As Walker descended through

100,0 feet, a heavy vibration occurred and continued for about 45 seconds until recovery was affected at 55,000 feet. Incremental acceleration of approximately 1-g was noted in the vertical and transverse axes at a frequency of 13 cycles. This cor­responded to the first bending mode of the horizontal stabilator. The center of gravity of the horizontal surfaces was located behind the hinge line; consequently rapid surface movement produced both rolling and pitching inertial moments. Subsequent analysis showed the vibration was sustained by the SAS at the natural frequency of the horizontal surfaces. Essentially, the oscilla­tions began because of the increased activi­ty of the controls on reentry which excited the oscillation and stopped after the pilot reduced the pitch-damper gain.’7

Two solutions to the problem were discussed between the FRC, North American, the Air Force, and the manufacturer of the SAS, Westinghouse; a notch filter for the SAS and a pressure-derivative feedback valve for the main stabilator hydraulic actuator. The notch filter eliminated SAS control surface input at 13 cycles, and the feedback valve damped the stabilator bending mode. In essence, the valve corrected the source of the problem, while the notch filter avoided the problem. Although it was felt that either solution would likely cure the problem, the final deci­sion was to use both.

NASA research pilot William Dana made a check flight in a specially-modified JF-100C (53-1709) at Ames on 1 November 1960, delivering the aircraft to the FRC the follow­ing day. The aircraft had been modified as a variable-stability trainer that could simulate the X-15’s flight profile. This made it possi­ble to investigate new piloting techniques and control-law modifications without using an X-15. Another 104 flights were made for pilot checkout, variable stability research, and X-15 support before the aircraft was returned to Ames on 11 March 1964.’®

The first government flight with the XLR99 engine took place on 7 March 1961 with Bob White at the controls. The X-15-2 reached Mach 4.43 and 77,450 feet, and the flight was generally satisfactory. The objectives of the flight were to obtain additional aerodynamic and structural heating data, as well as informa­tion on stability and control of the aircraft at high speeds. Post-flight examination showed a limited amount of buckling to the side-fuse­lage tunnels, attributed to thermal expansion. The temperature difference between the tunnel panels and the primary fuselage structure was close to 500 degrees Fahrenheit. The damage was not considered significant since the panels were not primary structure, but were only nec­essary to carry air loads. However, the buck­ling condnued to become more severe as Mach numbers increased in later flights, and eventually NASA elected to install additional expansion joints in the tunnel skin to minimize the buckling.141

By June 1961, government test pilots had been operating the X-15 on research flights for just over a year.20 The research phase of the X-15’s flight program involved four broad objectives: verification of predicted hyperson­ic aerodynamic behavior and heating rates, study of the X-15’s structural characteristics in an environment of high heating and high flight loads, investigation of hypersonic sta­bility and control problems during atmospher­ic exit and reentry, and investigation of pilot­ing tasks and pilot performance. By late 1961, these four areas had been generally examined, although detailed research continued to about 1964 using the first and third aircraft, and to 1967 with the second (as the X-15A-2). Before the end of 1961, the X-15 had attained its Mach 6 design goal and had flown well above 200,000 feet; by the end of 1962 the X – 15 was routinely flying above 300,000 feet. The X-15 had already extended the range of winged aircraft flight speeds from Mach 3.2’1 to Mach 6.04, the latter achieved by Bob White on 9 November 1961.

The X-15 flight research program revealed a number of interesting things. Physiologists discovered the heart rates of X-15 pilots var­ied between 145 and 185 beats per minute in flight, as compared to a normal of 70 to 80 beats per minute for test missions in other aircraft. Researchers eventually concluded that pre-launch anticipatory stress, rather than actual post launch physical stress, influ­enced the heart rate. They believed, correct­ly, that these rates could be considered as probable baselines for predicting the physio­logical behavior of future astronauts. Aerodynamic researchers found remarkable agreement between the wind tunnel tests of exceedingly small X-15 models and actual results, with the exception of drag measure­ments. Drag produced by the blunt aft end of the actual aircraft proved 15 percent higher than wind tunnel tests had predicted.

At Mach 6, the X-15 absorbed eight times the heating load it experienced at Mach 3, with the highest heating rates occurring in the frontal and lower surfaces of the aircraft, which received the brunt of airflow impact. During the first Mach 5+ flight, four expan­sion slots in the leading edge of the wing generated turbulent vortices that increased heating rates to the point that the external skin behind the joints buckled. It offered “… a classical example of the interaction among aerodynamic flow, thermodynamic proper­ties of air, and elastic characteristics of struc­ture.” As a solution, small Inconel X alloy strips were added over the slots and addi­tional fasteners on the skin.22

Heating and turbulent flow generated by the protruding cockpit enclosure posed other problems; on two occasions, the outer panels of the X-15’s glass windshields fractured because heating loads in the expanding frame overstressed the soda-lime glass. The difficulty was overcome by changing the cockpit frame from Inconel X to titanium, eliminating the rear support (allowing the windscreen to expand slightly), and replac­ing the outer glass panels with high temper­ature alumina silica glass. All this warned aerospace designers to proceed cautiously. During 1968 John Becker22 wrote: “The real­ly important lesson here is that what are minor and unimportant features of a subson­ic or supersonic aircraft must be dealt with as prime design problems in a hypersonic air­plane. This lesson was applied effectively in the precise design of a host of important details on the manned space vehicles.”

A serious roll instability predicted for the airplane under certain reentry conditions posed a dilemma to flight researchers. To accurately simulate the reentry profile of a returning winged spacecraft, the X-15 had to fly at angles of attack of at least 17 degrees. Yet the wedge-shaped vertical and ventral stabilizers, so necessary for stability and control in other portions of the flight regime, actually prevented the airplane from being flown safely at angles of attack greater than 20 degrees because of potential rolling prob­lems. By this time, FRC researchers had gained enough experience with the XLR99 engine to realize that fears of thrust mis-

A common sight dur­ing the 1960s over Edwards—an NB-52 carrying an X-15.This was a boy’s dream at the time; and the sub­ject of many fantasies.

The Million Horsepower Engine13Over the course of the program, the markings on the NB-52s changed significantly. Early on, they were natural metal with bright orange verti­cals; later they were overall gray. (NASA)

alignment—a major reason for the large sur­faces—were unwarranted. The obvious solu­tion was simply to remove the lower portion of the ventral, something that X-15 pilots had to jettison prior to landing anyway so that the aircraft could touch down on its landing skids. Removing part of the ventral produced an acceptable tradeoff; while it reduced stability by about 50 percent at high angles of attack, it greatly improved the pilot’s ability to control the airplane. With the ventral off, the X-15 could fly into the previously “uncontrollable” region above 20 degrees angle of attack with complete safety. Eventually the X-15 went on to reentry tra­jectories of up to 26 degrees, often with flight path angles of -38 degrees at speeds up to Mach 6.1J Its reentry characteristics were remarkably similar to those of the later Space Shuttle orbiter.

When Project Mercury began, it rapidly eclipsed the X-15 in the public’s imagina­tion. It also dominated some of the research areas that had first interested X-15 planners, such as “zero-g” weightlessness studies. The use of reaction controls to maintain attitude in space proved academic after Mercury flew, but the X-15 would furnish valuable information on the blending of reaction con­trols with conventional aerodynamic con­
trols during exit and reentry, a matter of con­cern to subsequent Shuttle development. The X-15 experience clearly demonstrated the ability of pilots to fly rocket-propelled air­craft out of the atmosphere and back in to precision landings. Paul Bikle saw the X-15 and Mercury as a “… parallel, two-pronged approach to solving some of the problems of manned space flight. While Mercury was demonstrating man’s capability to function effectively in space, the X-15 was demon­strating man’s ability to control a high per­formance vehicle in a near-space environ­ment… considerable new knowledge was obtained on the techniques and problems associated with lifting reentry.”25

Nearly all of the early XLR99 flights experi­enced malfunction shutdowns of the engine immediately after launch, and sometimes after normal engine shutdown or burnout. Since the only active engine system after shutdown was the lube-oil system, investiga­tions centered on it. Analyses of this condi­tion revealed very wide acceleration excur­sions during the engine-start phase. A rea­sonable simulation of this acceleration was accomplished by placing an engine on a work stand with the ability to rotate the engine about the Y-axis. Under certain con­ditions, the lube-oil pump could be made to

cavitate for about 2 seconds, tripping an automatic malfunction shutdown. To elimi­nate this problem, a delay timer was installed in the lube-oil malfunction circuit which allowed the pump to cavitate up to 6 seconds without actuating the malfunction shutdown system. After this delay timer was installed in early 1962, no further engine shutdowns of this type were experienced.26

But a potentially more serious XLR99 prob­lem was the unexpected loss of the Rokide coating from the combustion chamber during firing. A meeting was held at Wright Field on 13 June 1961 to discuss possible solutions. It was decided that the Wright Field Materials Laboratory would develop a new ceramic coating for the chambers, and that FRC would develop the technique and fixtures required to recoat chambers at Edwards. Originally, the Materials Laboratory award­ed a contract to Plasmakote Corp. to perform the coating of several chambers, but the results were unsatisfactory. By March 1962, the techniques and fixtures developed by the FRC allowed chambers to be successfully recoated at Edwards.

Early in the program, the X-15’s stability
augmentation and inertial guidance systems were two major problem areas. NASA even­tually replaced the Sperry inertial unit with a Honeywell system designed for the stillborn Dyna-Soar. The propellant system had its own weaknesses; pneumatic vent and relief valves and pressure regulators gave the greatest difficulties, followed by spring pres­sure switches in the APUs, the turbopump, and the gas generation system. NASA’s mechanics routinely had to reject 24-30 per­cent of spare parts as unusable, a clear indi­cation of the difficulties that would be expe­rienced later in the space programs in getting parts manufactured to exacting specifica­tions.27 Weather posed a critical factor. Many times Edwards enjoyed good weather while other locations on the High Range were cov­ered with clouds, alternate landing sites were flooded, or some other meteorological con­dition postponed a mission.