Airborne Simulators

In addition to ground simulators and the centrifuge, pilots and researchers used aircraft to simulate various aspects of the X-15. For instance, the Lockheed F-104 Starfighter closely approximately the wing loading of an X-15 during landing, and with the right combination of extended landing gear, flaps, and speed brakes, the F-104 at idle thrust did an excellent job of simulating the X-15. For the first 50 or so flights, the pilots dedicated an entire F-104 mission to practicing landing procedures. As new pilots entered the program, they conducted similar practices. Throughout the program, pilots used the F-104s to establish geographic checkpoints and important altitudes around the landing pattern at all the possible landing lakes.-1541

Scott Crossfield and Al White conducted similar work very early in the program using the North American YF-100A equipped with an eight-foot drag chute. Combined with extended gear and speed brakes, the F-100 at idle thrust did an adequate job of simulating the X-15 during landing, although not quite as well as the F-104. The entire process was a bit trickier since it required the in-flight deployment and release of the drag chute.1551

As Al White later remembered, "With gear down, speed brake extended, at idle power, and that drag chute deployed, the airplane was comparable to the X-15 on approach. I would start at about 25,000 feet, pick a spot on the lakebed, and see how close I could come to touching down on that spot. With all the room on the lakebed, it was not necessary to hit a spot, but it is always nice to have that much margin for error. I flew this trainer as much as I could, in preparation for that day that never came." Not flying the X-15 was one of the few disappointments during White’s significant career.-*56

Much of the X-15 flight planning took place prior to the first manned space flight. Since no one had ever left the atmosphere and returned in a winged vehicle (or anything else), there had been concern that the rapidly changing stability and control characteristics in the X-15 as it reentered the atmosphere might pose an unusually demanding piloting task. To address this question, engineers in the Flight Research Department of the Cornell Aeronautical Laboratory conceived the idea of simulating this brief (about 60 seconds duration) but unfamiliar X-15 piloting task in a NT-33A that was owned by the Air Force but operated by Cornell as a variable-stability trainer.-*57*

The NT-33A already had been equipped with a larger internal volume F-94 nose section that contained a three-axis (pitch, roll, and yaw) variable-stability and control system for in-flight simulation purposes. To support the X-15 program, Cornell modified the front cockpit to superficially resemble the X-15, with a side-stick controller on the right-hand console for atmospheric flight control and another side-stick on the left-hand console simulating the ballistic controls. An "instructor" pilot sat in the back cockpit with a normal set of T-33 controls. Jack Beilman at Cornell designed a programmable, non-linear function generator that changed the gains of 32 sensed aerodynamic and rigid-body-motion feedback variables. It also changed the flight-control sensitivities continuously during the

simulated reentry so that the NT-33A stability and control characteristics would match the predicted X-15 characteristics.-158!

The flight plan had the NT-33A entering a shallow dive at about 17,000 feet altitude and then pulling up to a ballistic trajectory that produced about 60 seconds of 0 g-about the same as the initial part of the X-15 reentry. At the same time, the variable-stability system on the NT-33A changed the flight-control sensitivities to simulate going from the vacuum of space to the rapidly increasing dynamic pressure of the atmosphere. Since the normal aerodynamic controls of the X – 15 would be ineffective outside the atmosphere, the pilot used the ballistic controller to establish the correct reentry pitch attitude.-*56

In the NT-33A simulation the "ballistic controller" produced no physical response whatsoever—it only changed the displayed pitch attitude on the instrument panel. (At this point in the simulation, the NT-33A was at 0 g.) In order to maintain the fidelity of the simulation, the X-15 pilot in the front cockpit wore a hood and had no view of the outside world, since there would be little view of the real world in the X-15 at the simulated altitudes. This deception was necessary for the high – angle-of-attack deceleration at the end of the simulated reentry because although the front cockpit instrumentation indicated the pilot was flying an unbanked steep descent (in the X-15), he was actually flying a steep 5-g turn in the NT-33A. The simulator achieved this deception by gradually biasing the attitude indicator to a bank angle of 75 degrees while the X-15 pilot used the ballistic controller to maintain wings-level flight at the proper airspeed, angle of attack, and descent rate on his cockpit instruments. It was a carefully choreographed ballet between the "student" in the front seat and the safety pilot in the back who was trying to keep the NT-33 from becoming a smoking crater in the high desert.*68!

Accordingly, a Cornell team headed by engineering test pilots Bob Harper and Nello Infanti arrived at Edwards in May 1960 to begin a series of flights in the NT-33A in order to provide reentry training for six X-15 pilots (Neil Armstrong, Jack McKay, Forrest Petersen, Bob Rushworth, Joe Walker, and Bob White). Each pilot was to receive six flights in the NT-33A that included a matrix of simulated Mach numbers, altitudes, and various control malfunctions (principally failed

dampers) both separately and simultaneously.1611 Infanti was the "instructor pilot" for each of the X-15 simulation flights in the NT-33A, and the rest of the Cornell team consisted of crew chief Howard Stevens, electronics technician Bud Stahl, and systems engineer Jack Beilman. As Beilman remembers:

During one of the flights, with Neil Armstrong in the front seat, we were simulating failed dampers at something like Mach 3.2 and 100,000 feet altitude. Neil had great difficulty with this simulated undamped X-15 configuration and lost control of the airplane repeatedly.

Nello had to recover from each one of these "lost-control" events using the controls in the back cockpit. [Infanti later recalled that some of these recoveries were "pretty sporty."] The ground crew was monitoring the test radio frequency as usual and followed these simulated flight control problems with great interest.

After landing, the NT-33A taxied to the ramp and Howard Stevens attached the ladder to the cockpits and climbed up to talk to Infanti about the airplane status. I climbed up the ladder front side to talk to Neil Armstrong. He handed me his helmet and knee-pad, got down from the cockpit and we talked about the flight and walked toward the operations building. As we arrived at the door Armstrong extended his right hand to grasp the door handle-but his hand still held the side-stick that he had broken during his last battle with the X-15 dampers-off simulation. I was unaware of any report of this incident during the flight and had not noticed the stick in Armstrong’s hand when he exited the cockpit. Addressing the matter for the first time, Armstrong said-without additional comment—"Here’s your stick!"

[It developed that Infanti had been aware of the broken side-stick after it happened because Armstrong had held it up over his head in the front cockpit for Nello to see.]

After the debriefing, we took the broken side-stick to the NASA workshop where Neil found the necessary metal tubing and repaired the stick while I mostly watched him work. The side-stick was reinstalled and ready for the first flight the next morning. Really good test pilots fix what they break!

In general, the pilots considered the NT-33 flights worthwhile, but there were some "obvious discrepancies or malfunctions" during the early flights. There were also a fair number of delays in the flights due to various system malfunctions caused by the high temperatures at Edwards. Eventually the Cornell crew corrected the malfunctions, but the X-15 pilots considered the first 10 flights unsatisfactory since they did not adequately simulate the X-15 flight profile. This was largely because the programmed trajectories required the NT-33 to fly close to its maximum capabilities: something that was not as easy as it sounds, especially in the heat over the high desert.-1621

The X-15 pilots considered the final six flights, flown during the first half of September 1960, reasonably satisfactory. In fact, the pilots discovered a novel control technique for the divergent closed-loop lateral-directional oscillation encountered at Mach 3.5 and 10 degrees angle of attack with the SAS off during these flights. By using the rudder in conjunction with the turn and bank indicator (which was, in effect, a yaw-rate meter) the pilot was able to damp the oscillations. With this technique, the ailerons were only a steady-state controller; in fact, any attempt to use the ailerons for control caused an immediate divergence. Researchers further investigated this technique on the North American fixed-base simulator with good results.1631

the X-15 flight profile somewhat more convincingly than the NT-33, making it possible to investigate new piloting techniques and control-law modifications without using an X-15. The most limiting factor was that the JF-100C was a single-seat aircraft, meaning that no safety pilot was available to lend assistance if things went wrong. To establish the X-15 flight characteristics on the JF-100C, technicians connected two portable analog computers to the airplane so that the combination became, essentially, a fixed-base simulator. One analog computer simulated the basic F-100C flight characteristics, and researchers manipulated the variable-stability gains until the motion traces matched those obtained from the North American X-15 simulator. Joe Walker and Bob White flew these pseudo fixed-base simulations until they were satisfied that the JF – 100C adequately represented the X-15.[64]

Much of the X-15 flight planning took place prior to the first manned space flight. There was concern that the rapidly changing stability and control characteristics in the X-15 as it reentered the atmosphere might pose an unusually demanding piloting task. To address this, the Cornell Aeronautical Laboratory developed a method of simulating this environment using an NT-33A operated by Cornell as a variable stability trainer. The simulations were hardly ideal, but provided much needed confidence to the original cadre of X-15 pilots. (U. S. Air Force)

The first actual flight of the JF-100C with the new mechanization was made on 24 March and was considered generally satisfactory. The major discrepancies were that the Dutch-roll and roll – subsidence modes appeared to be less stable than those of the actual X-15. Nevertheless, the JF – 100C was capable of performing some interesting simulations. For instance, six flights in late July 1961 simulated the X-15 at Mach 3.5, 84,000 feet, and 10 degrees angle of attack; later flights extended this to Mach 6 and angles of attack of 20 degrees. The aircraft returned to Ames on 11 March 1964 after making 104 flights for pilot checkout, variable-stability research, and X-15

[65]

support.

One of the tasks assigned to the JF-100C was investigating the effects of damper failure on the controllability of the X-15. Researchers had obtained the early wind-tunnel data on sideslip effects with the horizontal stabilizer at zero deflection, and used this data in the 1958 centrifuge program at Johnsville. Based on these data, reentries using an angle of attack of less than 15 degrees were possible even with the roll damper off. On the other hand, reentries at angles greater than 15 degrees (which were required for altitudes above 250,000 feet) with the roll damper off showed a distinct tendency to become uncontrollable because of a pilot-induced oscillation (PIO).[66]

As with a typical PIO, if the pilot released the control stick, the oscillations damped themselves. Nevertheless, researchers suspected that a large portion of the X-15 flight envelope was uncontrollable with the roll dampers off or failed. Investigations were initiated to find a way to alleviate the problem. The first method tried (perhaps because it would have been the easiest to implement) was pilot-display quickening. Sideslip and bank-angle presentations in the cockpit were quickened (i. e., presented with less delay) by including the yaw rate and roll rate, respectively. Researchers experimented with various quickening gains during investigation on the fixed-base simulator, but found no combination that significantly improved the pilot’s ability to handle the instability.-^

Shortly after the centrifuge program was completed, researchers conducted a wind-tunnel test to gather sideslip data with the horizontal stabilizer closer to the normal trim position (which was a large leading-edge-down deflection of -15 to -20 degrees). When researchers programmed the results of these tests into the fixed-base simulator at North American, it showed that the PIO boundary for reentry with the roll damper off had dropped from 15 degrees to only 8 degrees, adding new urgency to finding a solution.-681

To verify the magnitude of the problem in flight, several X-15 pilots explored the fringes of the expected uncontrollable region by setting the airplane up at the appropriate angle of attack and turning the roll and yaw dampers off. In each case, lateral motions began immediately. The pilots experimented with various combinations of angle of attack and control inputs in both the X-15 and the JF-100C to better define the problem.-691

Lawrence W. Taylor and Richard E. Day from the FRC, and Arthur F. Tweedie from North American independently investigated using the rolling tail to control sideslip angle during certain types of instability. An unconventional control technique, called "beta-dot," evolved from these investigations and showed considerable promise on the fixed-base simulator. This technique consisted of sharp lateral control inputs to the left as the nose swung left through zero sideslip (or vice versa to the right). The pilot kept his hands off the stick except when making the sharp lateral inputs, which eliminated the instability induced by inadvertent inputs associated with merely holding onto the center stick. However, when pilots used this technique in the JF-100C, it did not seem to work as well. Further investigations showed that it worked somewhat better in the X-15 when the pilot used the side-stick controller instead of the center stick.-701

It appeared that the beta-dot technique might allow reentries from high altitudes with the dampers failed, if anybody could figure out how to perform the maneuver successfully. As Bob Hoey, the flight planner who later discovered the ventral-off stability fix for the same problem, recalled, "the beta-dot technique is one of those things that is really difficult to explain. You could watch someone make 20 simulated reentries and still not understand what they were doing. The method was based on making a very sharp aileron pulse, timed exactly right, and totally foreign

to normal, intuitive piloting technique. Properly timed, this pulse would completely stop the rolling motion, although not necessarily at wings level. With a little finesse, you could herd the thing back to wings level flight, but, if at any time you reverted to a normal piloting technique, even for a second, you were in big trouble. Art Tweedie [who discovered this method] and Norm Cooper [a North American flight controls expert] could make successful simulator reentries with the dampers off while drinking a cup of coffee! This obviously became a big challenge for the rest of us." Hoey became pretty good at the technique himself, at least in the simulator.-171!

Dick Day later wrote that "Robert Hoey, lead Air Force engineer on the X-15 project, introduced the control technique to some of the X-15 pilots. Two pilots in particular, Major Robert White and Captain Joe Engle, became so adept at controlling ground and flight simulators that they considered the method would serve as a backup in case of roll damper failure. Fortunately, the beta-dot technique was not required because removing the ventral solved the dampers-off controllability problem. It is worth noting, however, that the complete beta-dot equation was later used in the yaw channel of the Space Shuttle control system to overcome unstable control coupling." It is another enduring legacy of the X-15 program.-721

All of the X-15 pilots trained using this technique, but the actual usefulness of the beta-dot maneuver was questionable. Furthermore, a lateral input in the wrong direction, which was conceivable considering other potential problems clamoring for the attention of the pilot, could be disastrous. One of the reasons the technique was so foreign to the pilot was that the aileron pulse had to be in the same direction as the roll, which is hardly intuitive for most pilots. Then the pilot had to remove the pulse just as the needle on the sideslip indicator hit the null mark. As Hoey remembers, "about half the pilots were dead-set against [the beta-dot maneuver] and essentially refused to consider it as an option. Others conquered the technique and actually became fairly proficient in its use on the fixed-base and in-flight simulations." Pilots flew the in-flight simulations using the NT-33 and JF-100C variable-stability airplanes, which somehow managed to survive the program.731

Researchers at Ames modified a North American JF-100C (53-1709) Super Sabre into a variable – stability trainer that could simulate the X-15 flight profile somewhat more convincingly than the NT-33, making it possible to investigate new piloting techniques and control-law modifications without using an X-15. The most limiting factor was that the JF-100C was a single-seat aircraft, meaning there was not a safety pilot to assist if things went wrong. (NASA)

There were two other answers to the PIO problem at high angles of attack. The first was to make the stability augmentation system truly redundant, at least in the roll axis, by installing the alternate stability augmentation system (ASAS); however, this took almost a year to accomplish. Another answer-discovered by Dick Day and Bob Hoey using the simulator-proved to be remarkably easy, and unexpected: remove the ventral rudder. With the lower rudder on, a considerable portion of the reentry from an altitude mission would be within the uncontrollable region should a damper fail. However, a similar reentry with the lower rudder removed would not enter the predicted uncontrollable region at all. The downside was that the pilots faced significantly reduced flying qualities at low angles of attack without the rudder. Despite a few gripes from the pilots, everybody eventually agreed to remove the lower rudder for almost all of the high-altitude missions. Only a few missions of the X-15A-2 used the ventral rudder, which in this case provided an adequate stand-in for the eventual dummy ramjet. In all, the program would make 73 flights with the ventral rudder on and 126 with it off.1741

By the time of the 1961 industry conference, researchers had determined that the fixed-base simulator and the F-104 in-flight landing pattern simulator were the two most valuable training tools available to the program. The centrifuge and variable-stability aircraft contributed to the overall pilot experience level, but were not necessary for use on a flight-by-flight basis. This mostly explains why only the first group of pilots got the thrills of "riding the wheel" at Johnsville and flying the NT-33 trainer.-1751