The MH-96 system did have some handling quality deficiencies during atmospheric flight. It lacked speed stability in the rate command mode. It also had a poor longitudinal trim system that would insidiously apply a continuous trim rate when you were not expecting it. The most serious handling quality problem of this system was a tendency to float or climb following the landing flare. This was very disturbing to the pilot because he only had a limited amount of time to get the aircraft on the ground before he ran out of airspeed. This problem was minimized by trimming in a nose-down pitching rate during landing approach. In general, the handling qualities of the aircraft were in agreement with the simulator predictions and the major discrepancies were a result of simulation deficiencies rather than real aircraft deficiencies.
The flight control systems performed as predicted with only two major exceptions. The standard SAS flight control system exhibited a structural resonance problem due to a poor choice of gyro package mounting location. Structural vibrations were transmitted into the flight control system command loop causing severe control system vibrations. This was corrected by a simple relocation of the gyro package. The MH-96 flight control system was susceptible to a self-sustaining limit cycle problem whenever abrupt control inputs were made while the control system gain was at its maximum. This subsequently resulted in a catastrophic accident.
Aerodynamically, the aircraft performed as predicted with base drag being the only significant discrepancy. Base drag was substantially higher than predicted, but this was attributed to the lack of proper compensation for the model sting which was mounted in the aft end of the model to support it in the wind tunnel. Thus the model drag was not simulated accurately. As a result of this increased base drag, the subsonic lift/drag ratio was lower than predicted.
In the structures area, there were no major discrepancies between predicted and measured loads and thus there were no structural problems except in the landing gear. The dynamic landing loads were not accurately predicted, particularly those due to aerodynamic forces, and as a result several structural failures occurred in various landing gear components. On one occasion the failure was catastrophic resulting in severe aircraft damage and injury to the pilot.
In the aerothermal structures disciplinary area, the thermal protection system consisting of high-temperature materials and a heat sink structural concept, worked as intended. The overall structural temperatures and stresses were within acceptable limits. Admittedly, the heating rates encountered at hypersonic speeds were lower than predicted, but it appears that there was enough design margin in the structure to accommodate the higher anticipated rates. The researchers were somewhat surprised that maximum-altitude flights subjected the aircraft to a larger total heat pulse or total Btu increment than a low-altitude heating flight. This may have been anticipated with a more thorough assessment of the two different missions. A typical heating flight was flown at relatively low altitudes: 80,000 to 90,000 feet. On this type of flight, the aircraft was subjected to high heating rates during the climb and acceleration phase, but the heating rate decreased abruptly after engine burnout due to the rapid deceleration of the aircraft in the denser air. Thus, although the heating rate was relatively high, the duration of the heat pulse was short.
A maximum-altitude flight subjected the aircraft to a somewhat lower heating rate during the climb, due to the higher average altitude, but it also exposed the aircraft to another long heat pulse during the lengthy entry, pullout, and deceleration phases of the mission. Little heat was dissipated while the aircraft was outside the atmosphere, so the end result was a hotter airplane at landing. This was not a major problem, but it was a surprise.
The thermal protection system concept used in the X-15 worked well for the short duration missions of the X-15. That TPS concept would not be a viable concept for vehicles such as the shuttle or the proposed National Aerospace Plane. A number of more advanced systems are being considered for future hypersonic vehicles.
For the most part, the X-15 was not probing the unknown. Rather it was validating existing design tools. In one area however, it was doing pioneer work, investigating the pilot’s ability to fly an aircraft outside the atmosphere and then make a successful reentry into the atmosphere. There was no way to conduct that investigation on the ground. Existing simulators were inadequate. Flight was the only way. The answer soon became obvious as more and higher altitude flights were successfully accomplished. The maximum-altitude flight of the X-15 to 67 miles above the earth provided dramatic proof of the pilot’s ability to accomplish such a mission.
Overall, the research results obtained during the flight program were very impressive. The data generally agreed with predictions. There were some surprises and some problems, but the results confirmed the validity of the existing design tools and provided the necessary confidence to proceed with the design of a lifting entry spacecraft, the space shuttle. Many of the discoveries encountered in flight were the kind that Dr. Dryden had described; the unanticipated or practical kinds of problems that should have been foreseen but were overlooked; the kind of problems that were quite obvious after the fact. Some of those are described in the following pages.
As mentioned previously, we did not have any major problems during the initial envelope expansion flights to the design speed and altitude conditions. We did not encounter major problems until we began exceeding those design conditions or until we really began stressing the airplane at the limits of the design conditions. We were courting trouble when we intentionally flew the airplane at high speed at low altitude to heat it up, but we wanted to get good heating data. We were also at risk routinely flying above 250,000 feet altitude since a damper or stability augmentation system failure could have been catastrophic but again the potential information return was considered to be worth the risk. This is not to say that we had no problems early on—we had problems beginning with the first flight.
One of the early problems we encountered was the unreliability of the inertial system. The inertial system provided aircraft attitude information, inertial velocities and inertial altitude. Inertial data were essential for flying an accurate flight profile, since barometric airspeed, altitude, and rate of climb were not measurable at the flight conditions where the X-15 normally operated. We could not use a standard flight test nose boom to obtain air data above Mach 4, because the nose boom would fail structurally due to aerodynamic heating. Instead, we utilized a cooled servo-driven ball nose to measure dynamic pressure and angles of attack and sideslip, but we could not readily convert dynamic pressure into airspeed. Thus, we were forced to rely on inertial velocity to ascertain our true speed. We were also dependent on inertial measurements for altitude and rate of climb above 80,000.
The original X-15 inertial system was very unreliable. The system would quite often fail completely or would accumulate errors of such magnitude as to render it useless. It was not uncommon to have the inertial system indicating an altitude of over 100,000 feet when you were back on the lakebed after landing.
Loss of inertial data was not a major safety of flight problem. We had radar data from the tracking radar which could be relayed to us for energy management purposes, but those data were not available quick enough to use for flight test purposes. We wasted many flights during the early part of the program because we could not get to the desired flight condition due to inertial system failures.
The inertial system failed on my second flight passing through Mach 5 at 80,000 feet. It is hard to imagine how helpless one feels climbing and accelerating at fantastic rates and not knowing how high and how fast he is going. It is very disconcerting, to say the least.
Some major modifications were made to the original inertial system to improve its accuracy and reliability. These mods improved the system significantly, but we later opted for the inertial system developed for the Dyna-Soar spacecraft and used that system for the remainder of the flight program. That system proved to be both accurate and reliable.
A
nalysis of wind tunnel data prior to flight predicted a serious stability and control problem at high angles of attack for the basic airplane without stability augmentation. The problem exhibited itself as a divergent lateral-directional oscillation which increased in severity with increasing angle of attack. Above 10 degrees angle of attack, the airplane became uncontrollable. The airplane was equipped with a stability augmentation system, but there was a concern about the reliability of that system. Electronic systems in aircraft were relatively new at that time. The implications of this problem were that we could not safely achieve the design altitude of 250,000 feet since a pullout from that altitude required an angle of attack greater than 10 degrees. If we attempted a design altitude mission and lost the stability augmentation system for any reason, we would lose the airplane.
This problem became a challenge to the stability and control engineers. They attacked the problem in a number of different ways. One straightforward proposed solution was to provide a redundant stability augmentation system. Another proposal was to provide stability augmentation in the reaction control system. This would not be as good a solution but it would help. A less attractive proposal was a special control technique to damp the divergent oscillations. This proposal was not highly regarded by the pilots. The most innovative proposal involved the removal of the lower segment of the lower ventral fin. Analysis and simulation indicated that this would eliminate the basic problem rather than just fix it as the other solutions would do.
An aggressive flight investigation was initiated on Flight 37 to confirm that the problem existed and to demonstrate a solution or solutions to the problem. On Flight 37, Joe Walker evaluated the controllability of the airplane at high angles of attack with the stability augmentation system deactivated. He did not like what he saw. He continued the controllability evaluation on Flights 40 and 44 and confirmed the seriousness of the problem. In this same series of flights, the special control technique to damp the divergent oscillations was evaluated and a flight was also made without the lower portion of the lower ventral. The special control technique did not appear to be a viable solution and it was not pursued any further. The results of the ventral-off flight appeared encouraging, but more data was needed to commit to removing the ventral permanently.
A redundant stability augmentation system was developed, installed in the airplane, and checked out on Flight 50. The results were encouraging enough to convince the engineers to proceed with a flight to the design altitude. On Flight 52, Walker flew to 246,700 feet, essentially achieving the 250,000 feet design altitude. Thus, in less than fifteen flights, three of the four proposed solutions had been evaluated and the flight envelope had been expanded to the design altitude.
The effort to achieve a final solution to the problem continued. A reaction augmentation sytem, the fourth solution, was installed and evaluated on Flight 66. It showed promise. On Flight 67, Joe Walker was scheduled to investigate a new reentry technique using a constant pitch attitude instead of a constant angle of attack, in a flight to 220,000 feet. During the reentry Joe inadvertently deactivated the roll stability augmentation system. The airplane began a wild roll oscillation to bank angles over 90 degrees. Joe grabbed the center stick with his left hand during this oscillation in an attempt to control the airplane. He almost lost control. He finally managed to recover by decreasing the aircraft’s angle of attack. It was a wild ride, and it vividly demonstrated the severity of the controllability problem at high angles of attack without stability augmentation.
This incident expedited the decision to permanently remove the lower portion of the lower ventral. Flight 69 was the last flight flown with the lower portion of the lower ventral. This effectively eliminated the catastrophic nature of the high angle of attack problem. A pilot could still lose control, but the odds were now in his favor.
Some gutsy decisions were required during the effort to solve this serious problem. The decision to attempt a design altitude mission with a newly developed backup stability augmentation system was a bold move. Removing the lower ventral was very risky since no one was certain whether we would retain an adequate level of directional stability. We could possibly solve one problem but create another. As it turned out, the overall effort was successful—a tribute to an aggressive flight research team.
Local heating effects were another problem encountered during envelope expansion. This involved localized structural failure of the aircraft’s skin or other structure due to aerodynamic heating. A good example of this problem was the buckling of the wing skin behind the wing leading edge expansion gaps. The wing leading edge was constructed in several segments with gaps to accommodate expansion of the leading edge due to aerodynamic heating.
These gaps, however, created small vortices which were like welding torch jets of hot air that overheated the skin behind the gap and caused it to buckle. This buckling resulted in popped rivets or torn wing skin. The solution involved adding covers over these gaps to prevent the creation of these vortices. These small vortices of air were extremely hot. The wing skin was made of Inconel-X material, which was capable of sustaining strength up to 1,800° F, and yet these vortices caused the material to fail.
This type of local heating caused buckling of other skin panels at high speed and, ultimately, cracking in the skin at various locations that had to be patched. This skin buckling was suspected to be the source of the popping and banging noise that the pilots noted as the airplane got hot. We referred to this phenomenon as the oil-canning effect. The airplane really popped and banged as it got hot and I noted that the airplane twitched as it popped and banged. We also observed smoke entering the cockpit. We subsequently learned that the smoke was a result of outside air entering the aircraft through small gaps in external doors or panels. The incoming air acted just like a torch at high speeds and burned electrical wiring, aluminum internal structure, metal tubing, or anything else that it impinged on.
The nose gear scoop door, a small door on the nose gear door, normally opened during the gear deployment sequence to assist the deployment of the nose gear. It opened unexpectedly on three different occasions at high speed due to structural deformation resulting from excessive heating. The air rushing into the nose wheel well through this scoop door burned the tires on the dual wheels and partially burned the wheels. The wheels stayed intact during the landing, but the rollout was extremely rough due to disintegration of the tires.
Aerodynamic heating also shattered cockpit windows on two different flights. Luckily, on each occasion only one windshield shattered. The shattered windshield was completely opaque. If both had shattered, the pilot would have had no choice other than to eject. He would not have been able to see through the shattered glass well enough to land the aircraft.
Shattered windows remind me of an incident that occurred several years after the conclusion of the X-15 program. One of our former X-15 flight planners, John Manke, had been selected as a research pilot. He happened to be flying in an F-111 on a research mission that involved a supersonic low level maneuver. These types of maneuvers were generally conducted in a designated portion of the supersonic corridor. The supersonic corridor, as the name implies, is a specified corridor running through the Edwards restricted area which is reserved for supersonic testing. The low-level portion of this corridor is relatively short and is located in an extremely remote uninhabited area to ensure that nothing will be damaged as a result of the strong sonic booms produced at low altitudes.
John entered the corridor to begin his supersonic research maneuver. It took longer than anticipated to set up the proper flight conditions for the maneuver, thus using up some precious distance in the corridor. Once he established the proper conditions, he had to maintain them for 30 seconds to obtain good stabilized data. During this stabilized period, ground control requested that he extend the maneuver for another 15 seconds. John knew that he was approaching the end of the low-level corridor and was coming up on the town of Mojave, but he felt that he could complet
e the extended maneuver before he boomed Mojave. He was wrong—dead wrong. He boomed hell out of Mojave. He shattered windows in a wide swath right through the heart of the city. The angry phone calls started coming in before John could get back in the landing pattern at Edwards. John was advised by radio before he landed that he was in trouble.
Some of the wags in the pilot’s office, including myself, quickly located a picture of our acting center director, De Beeler. When John walked into the pilot’s office, he was greeted by a picture over his desk of the center director with a stern look on his face, which we had captioned with the words, “Damn you, John Manke.” The crowning touch or coup de maître was the shattered picture frame glass covering the picture. It was a magnificent stroke of genius, if I do say so myself.
At the Edge of Space Page 28
