CHAPTER 5 Space Simulation Studies - NASA · a control problem considerably different from that of...

Monographs in Aerospace History Number 40—Journey Into Space Research 35

CHAPTER 5 Space Simulation Studies

In the period before the start of thespace program, the division to which Iwas assigned was called the FlightResearch Division. As described in ref-erence 1.1, most of the research workwas conducted on full-scale airplanes.Some simulation studies were made,however, using special simulatorsdesigned to study specific problems.These simulators included a devicecalled the yaw chair, that enabled astudy of the ability of a human pilot tocontrol lateral oscillations over a widerange of oscillation periods with bothstable and unstable damping. In addi-tion, a device called the NAP (NormalOscillation and Pitch) chair was builtthat simulated the vertical and pitchingmotion of an airplane over a range ofvertical motion of about 6 feet. Thesesimulators allowed covering a range ofconditions systematically, rather thanobtaining different conditions by consid-ering results on a number of differentairplanes. At the time the space pro-gram started, a three-axis rotationalsimulator was under construction inwhich a cockpit was mounted to provideangular motion in pitch, roll, and yaw.This simulator was intended to studycombined rolling and yawing oscillationsof an airplane, but it found use for otherpurposes during the space program. Incontrast with modern simulators that

use electronic displays and general-pur-pose motion bases, these simulatorsused mechanical systems to simulateaccurately the motion of the vehicle. Theoutput of the mechanical system wasamplified by a hydraulic servomecha-nism based on a variable displacementhydraulic pump. Servomechanisms ofthis type were available in Naval gun tur-rets. I have since learned that the devel-opment of these servomechanisms waslargely attributed to Charles Manley, thesame man who earlier perfected theexcellent radial motor used in SamuelLangley’s aerodromes.

Almost simultaneously with the start ofthe space program, all testing of high-speed airplanes was transferred to theHigh-Speed Flight Research Center atEdwards Air Force Base, now calledthe Dryden Flight Research Center.I was left with the problem of decidingon the best use for the engineers undermy supervision, who had been trainedin studying the stability and control ofairplanes.

Airplanes, of course, had been flying formany years, and there was little needfor studying the basic principles of sta-bility and control. Most simulation workon airplanes was devoted to studyingoptimal stability and control characteris-tics or to finding the characteristics that

36 Monographs in Aerospace History Number 40—Journey Into Space Research

Space Simulation Studies

would provide the most desirable han-dling qualities for the pilot. At the start ofthe space program, before the firstmanned orbital flight, there was muchless confidence in the ability of a humanpilot to perform the tasks required forspace operations. Many engineersexpressed the view that it would bebetter to design spacecraft with com-pletely automatic control. Test pilots, onthe other hand, who had at leastapproached orbital flight conditions intests of very high-altitude airplanes,usually felt confident that they couldcontrol the entire flight of a spacecraftjust as they had controlled high-altitudeairplanes.

To resolve some of these questions, Ifelt that the conditions encountered inthe various phases of a space vehicleflight should be simulated as accuratelyas possible to give the astronautsexperience with the new problems ofspace flight. The following discussiondescribes some of the work done in thisperiod.

After the successful completion of JohnGlenn’s first orbital flight, most doubtsconcerning the effects of weightless-ness were dispelled. Soon after thistime, however, definite space programs,such as the Gemini and Apollo mis-sions, were planned. The simulationwork then focused on specific problemsencountered in the launching, flight,entry, and landing of the spacecraftdesigned for these missions. These sim-ulations are described in reference 5.1.

Lunar Landing Research Facility

Landing on the surface of the Moon wasknown to be one of the most criticalphases of the Apollo program. Controlby an astronaut, at least during the finalphases of the descent, was considered

mandatory because the nature of thelunar surface was not known in suffi-cient detail to plan the exact spot fortouchdown. Several conditions presenta control problem considerably differentfrom that of landing an airplane onEarth. The lunar gravity is one-sixth thaton the Earth. All control of lift and atti-tude is provided by rockets, which oftenprovide a discontinuous, on-off controlrather than a linear variation of controlforce familiar on airplane controls. Thecomplete lack of an atmosphere on theMoon makes it impossible to use anytype of aerodynamic control.

When President Kennedy made hisannouncement of a major program tosend men to the Moon on May 25, 1961,I immediately started thinking abouthow this operation could be simulated.I wrote a memorandum on this sub-ject in May 1961 and discussed thesubject with the Associate Director,Lawrence K. Loftin, Jr. on June 26,1961.

The trajectory of the lunar vehicle wouldbe different from that on Earth because,as stated previously, the gravitationalattraction of the Moon is only one-sixththat of the Earth. To simulate thereduced gravity, I visualized a suspen-sion system for the simulated vehiclethat would exert a constant force in thevertical direction equal to five-sixth theweight of the vehicle. The force on thecable on which the vehicle was sus-pended could be measured by a strain-gauge balance at the vehicle and usedto control the output of a servomecha-nism that reels the cable in and out asrequired to apply the desired constantforce to the top of the cable. To providefor horizontal motions of the vehicle inthe fore-and-aft and lateral directions,sensors would measure the tilt of thecable from the vertical and would beused to control servomechanisms thatmoved the suspension point to keep itdirectly over the vehicle.



The motions of the vehicle in responseto pilot commands would be provided byrockets. As in an actual lunar vehicle, arocket sufficiently powerful to supportthe weight of the vehicle in the lunarenvironment, plus some extra power tomaneuver, is required. Smaller rocketsare used to provide pitching, rolling, andyawing moments. Previous studies hadfound that a system using a platinumcatalyst to decompose hydrogen perox-ide into steam and oxygen provided aconvenient and relatively safe means tomake a controllable rocket.

My first concern in designing the lunarlanding facility was to analyze the servo-mechanism used to maintain a constantforce in the suspension cable while thevehicle was going through the maneu-vers of landing. While the technicaldetails of this analysis are too involvedto present in this discussion, a briefreview of the problems involved in ser-vomechanism design may be of interest.An example of a simple type of servo-mechanism is an autopilot to hold anairplane on a desired constant heading.If the heading deviates from the desiredvalue, a compass or other headingdetector measures the error in heading.The error may be converted to an elec-tric voltage that is fed to an electronicamplifier. The output of the amplifierdrives an electric motor that moves therudder of the airplane in the direction toreduce the error. As the heading error isreduced to zero, the rudder is returnedto its neutral position. The ratio betweenthe rudder angle and the error in head-ing is called the gain of the servomech-anism. Increasing the gain increasesthe speed with which the heading erroris reduced, but it may cause the rudderto overshoot its neutral position andoscillate about zero. Beyond a certainvalue of the gain, the oscillationsincrease with time, a condition calleddynamic instability. The gain must be

kept to a value safely below that whichproduces instability.

In the case of the lunar landing researchfacility, a servomechanism maintains aconstant force in the suspension equalto five-sixth of the vehicle weight whilethe pilot controls the rocket that pro-vides an additional one-sixth of theweight plus whatever additional force isrequired to control the rate of descent orto slow the vehicle down for landing.These force variations supplied by therocket act as disturbances to the force inthe cable. The error in the cable force iscorrected by the servomechanism byvarying the speed at which the cable isreeled in or out.

If a steady disturbance is applied to asystem controlled by a servomecha-nism, a steady error may result. Thiscondition may be corrected by placingan integrator in the feedback loop. Theintegrator builds up a signal thatincreases with time to offset the steadyerror. An integrator also has a gain todetermine its speed of operation. If thegain is too large, dynamic instability willagain result. In general, the tendency toinstability is greater with an integrator inthe circuit.

In the case of the lunar landing facility, atypical mode of operation is to let thevehicle fall freely under a steady accel-eration of one-sixth g until the descentrate reaches the desired value to startthe landing run. In this case, the cabletension should remain at its desiredvalue during this period of constantacceleration. To maintain this steadyvalue during conditions of constantacceleration, servomechanism theoryshows that a double integration of theerror is required. Such an arrangementis called a type 2 servomechanism. Itherefore realized that a double integra-tion should be included in the computerof the lunar landing facility. With suchan arrangement, careful design of the



gains is required because an evengreater tendency to instability exists.

Another effect that influences the stabil-ity of the response of the system is thetime lag between the output of the ser-vomechanism and the resulting forceapplied to the vehicle. This lag resultsfrom the time required for tension andcompression waves to travel down thesuspension cable. This speed is equiva-lent to the speed of sound in the cable.If a braided steel cable is used, thespeed of sound in the cable is about1000 feet per second. For a cable200 feet long, the lag is two-tenths of asecond.

Despite the effort to make an accuratesimulation of the trajectory and controlcharacteristics of the lunar module inthe lunar environment, one factor thatcannot truly simulate conditions on theMoon is the gravitational force of theEarth acting on the pilot’s body. Theeffect of this force can be minimized bystrapping the pilot in and otherwise sup-porting his body so that this force doesnot interfere with his control activities.

Other factors that complicate the analy-sis of the servomechanism are thespringiness of the cable, which changeswith the altitude of the vehicle, thedamping of the drive mechanism, andthe change in weight with fuel usage. Iconsidered as many of these factors aspossible in an analytical study of thesystem. All these factors made the anal-ysis difficult either with available theoryor with the computers available at thatdate. I therefore considered an experi-mental study of the system necessary.

To test the feasibility of the contem-plated system, a simplified system,called the pilot model, was built in whicha pilot’s chair was suspended by avertical cable. A photograph of the pilotmodel in operation is shown in fig-ure 5.1-1 with the engineering test pilotJack Reeder at the controls. A servo-

mechanism that reeled the cable in orout was mounted in the girders in theroof of the NACA Full Scale Tunnel,60 feet above the vehicle. This facilitywas chosen because the wind tunnelhad a compressed air supply that couldbe used to provide thrust for the rocketthat was used to overcome the simu-lated lunar gravity of one-sixth of thevehicle weight. An available analogcomputer of fairly early design was usedto calculate the signals driving the ser-vomechanism. With this apparatus, thegains of the system were varied experi-mentally until a reasonably constantcable tension could be maintained asthe pilot performed a simulated landing.

An important reason for using the over-head suspension is that if any loss ofcontrol occurs, either due to pilot erroror some malfunction in the system,the vehicle can be locked in place orlowered slowly to the ground like anelevator.

The successful operation of the pilotmodel, despite some claims that itwould not be feasible, gave me confi-dence to propose the design and con-struction of a full-scale system. I wrote amemorandum on the proposed systemand made a trip with Lawrence K. Loftinto NASA Headquarters in Washington,DC. There we discussed the project withIra H. Abbot, who at that time was anassistant director for space programs.He quickly agreed to support the projectand to provide the necessary funds.

At the same time that my project wasproposed, engineers at the AmesResearch Center proposed a flight vehi-cle that worked on the same principle.This device, called the Lunar LandingTraining Vehicle (LLTV), used a turbojetengine to support five-sixth of the vehi-cle weight. The engine was mounted ongimbals and controlled by servos con-nected to an inertial measurement unitso that it would always point vertically.



The pilot’s cockpit was mounted in alarge frame that surrounded the jetengine and contained the necessaryrocket engines to exert force to offsetthe simulated lunar gravity and to pro-vide pitch, yaw, and roll control. Ithought that this device would be very

dangerous to fly because it had nomeans to recover in case of loss of con-trol. Eventually, three of these vehicleswere built. The cost of this project wasundoubtedly many times that of theLunar Landing Research Facility(LLRF), but the funding for the Apollo

FIGURE 5.1-1. Pilot model for LLRF being flown by NASA test pilot Jack Reeder.

L-63-9783



program was so generous that bothprojects were supported.

The firm of Jackson and Moreland wasselected to build the gantry and operat-ing system for the LLRF. The estimatedcost was $4,997,700, which was closeto the actual cost. The contract for a

piloted vehicle to be lifted on the LLRFwas let to Jered Industries, a small com-pany with shops located under an oldfootball stadium in New England. Thiscompany built the piloted vehicle for theremarkably low cost of $250,000. Thisvehicle differed from the LM used in the

FIGURE 5.1-2. Lunar Landing Research Vehicle showing origi-nal cab for pilot.



Apollo mission because the LM had notbeen designed at the time the LLRFwas constructed. The test vehicle dif-fered from the LM in being smaller andinitially having a bulbous helicopter typecockpit (fig. 5.1-2). Later the LLRF testvehicle was equipped with a cockpit thatplaced the astronaut in a standing posi-tion like that in the LM, with widewindows to see the lunar landscape(fig. 5.1-3). Both vehicles had out-stretched legs with pads on the ends togive a stable support on the lunarsurface.

A photograph of the completed LLRF isshown in figure 5.1-4. The gantry, orlarge crane structure used to suspendthe vehicle, is 300 feet long, 250 feethigh, and 100 feet wide. The overallheight of 250 feet was set by the limitson the height of buildings on Langley AirForce Base to avoid interference with airtraffic. The tracks on which the suspen-sion system rode allowed a lateral travelof ±15 feet and a movement down thetrack of 250 feet. With allowance for thespace taken up by the structure andsuspension system and the overallheight of the suspended vehicle, thechange in altitude during a test run wasabout 185 feet. Part of this amount wastaken up by the distance required to fallat 1/6 g to the initial descent velocity.The system was finished in 1966 andcompleted initial testing in 1967.

After the initial planning of the LLRF andthe construction had been completed, Itook little part in actual operations.Donald E. Hewes was placed in chargeof running the facility, with ThomasO’Bryan as assistant. Among the otherNASA engineers assigned to the facilitywere Eric Stewart, Maxwell Goode,Randall Harris, Max Kurbjun, AmosSpady, Marna Mayo, and Frank Read.

As the name implies, research work wasdone on the facility to study the controllaws relating the pilot’s control inputs to

the vehicle response, the ability of thepilot to make a sufficiently soft landing,and so forth. NASA test pilot LeePerson did much of this work. In laterstages of the work, Don Hewesarranged to have piles of cinders placedon the ground under the facility thatwere shaped to simulate lunar cratersand terrain. In addition, tests were madeat night with lighting simulating the lowposition of the Sun that would bepresent in the actual lunar landing. Theshadow of the vehicle on the terrain,shown in figure 5.1-5, gave the astro-nauts a good impression of the height ofthe lander as it approached the surface.All the astronauts who were scheduledto make lunar landings in the Apollo pro-gram took part in test runs made underthese conditions. The astronauts laterstated that the landings were very goodsimulations of the actual landings on theMoon, and that they were very helpful intraining for these operations (ref. 5.1).

While the tests at Langley were inprogress, the Ames flying vehicle, theLLTV, was placed in operation at theManned Space Center at Houston andwas flown by the astronauts. In one ofthe early tests, the vehicle went out ofcontrol. The astronaut, Neil Armstrong,ejected barely in time to save his life,and the vehicle was destroyed. I wascalled to Houston to serve on the reviewboard that studied the causes of thecrash. The cause was found to be thataerodynamic forces on the framework ofthe vehicle in forward flight were largeenough to overpower the jets used forcontrol. One of the remaining vehicleswas then mounted in the Langley Full-Scale Tunnel to study means for reduc-ing the aerodynamic forces on theframework.

Later, another of the LLTVs went out ofcontrol and crashed while being flownby Joe Algranti, a former Langley testpilot who was then in charge of flightoperations at Houston. The crash was



L-67-9108

FIGURE 5.1-3. Lunar Landing Research Vehicle hovering with stand-up cab for pilot.

L-69-6324

FIGURE 5.1-4. Lunar Landing Research Facility at NASA Langley Research Center, Hampton, Virginia.



again found to be caused by a windgust, this time from the side, that over-powered the controls. Both Armstrongand Algranti were saved by the Martin-Baker ejection seat, a so-called zeroaltitude ejection seat that shot the pilot

up from ground level with a rocket suffi-ciently powerful to put him at an altitudewhere he could be saved by his para-chute. This seat, an English invention,had become available just in time to beused in the program.

FIGURE 5.1-5. Simulated craters on terrain in landing area of LLRF. Shadow of vehicle with lighting simulates low Sun angle.

L-69-4850



L-06088

L-06092

FIGURE 5.1-6. Lunar Landing Research Facility in use at crash test facility.

(a) Test aircraft about to impact ground in crash test facility.

(b) Results of crash test.

Lunar Orbit and Landing Approach (LOLA) Simulator


The astronauts felt that both simulatorswere of value in training for the lunarlanding. The LLTV had the advantage ofproviding for a greater altitude at thestart of the maneuver, 300 or 400 feetinstead of about 185 feet in the LLRF.On the other hand, the LLRF had theadvantage of more realistic simulationof the terrain and lighting conditions.The astronauts, like typical test pilots,ignored the dangers inherent in theflight vehicle and continued making sim-ulated landings with the remaining vehi-cle, but great care was taken to restricttesting to calm days.

The Apollo program was stoppedabruptly after six successful lunar land-ings, even though two additional Saturnlaunch vehicles had been constructed tocontinue the program through two moreflights. Astronaut training flights weretherefore stopped. Don Hewes contin-ued research with a small one-manvehicle to study the feasibility of a flightvehicle for lunar exploration in place ofthe lunar rover that had been used onthe last three Apollo flights. The flightvehicle showed promise for muchgreater range than the rover becausethe low gravity on the Moon allows arocket-supported vehicle to fly for longdistances. This project was alsodropped, however, when there was noprospect for future lunar landings.

The LLRF stood idle for a number ofyears and the servos and other equip-ment on top of the gantry wereremoved. A new use was found for thegantry as a crash test research facility.In this application, used airplanes or air-plane structures were hauled up to thetop of the gantry and allowed to fall in acircular path, like a pendulum, to impactthe ground. For some tests, a rocketboost was used to increase the speed ofthe impact. These tests have beenfound to be very useful in designingthe airplane and the cockpit structureto protect the pilot in case of a crash.

Pictures of such a test are shown in fig-ures 5.1-6 (a) and (b).

The most novel feature of the LLRFfrom a technical standpoint was the ser-vomechanism that maintained a con-stant tension in the suspension cableduring the landing maneuver. Otheruses for this feature are rather scarce,but I did learn of one practical applica-tion. A group of engineers from Canadacame to Langley to discuss the use ofthe constant-tension servo to haul ahelicopter down to the deck of a ship ina rough sea. In this application, thecable would be lowered from the heli-copter and hooked onto the deck of theship by a crewman on the ship. Then thecable would aid the pilots of the helicop-ter in descending to the deck withoutexperiencing large impact loads due tothe motions of the deck.

The LLRF and the Research Vehiclehave been declared State Historic Land-marks. The gantry is still the most prom-inent feature on the NASA Langleylandscape, and the vehicle is on displayin the Virginia Air and Space Center inHampton.

Lunar Orbit and Landing Approach (LOLA) Simulator

The Lunar Orbit and Landing Approach(LOLA) simulator was, as the nameimplies, intended to study the ability of apilot to control the LM following its sepa-ration from the Command and ServiceModules during the lunar orbiting phaseof the flight until it reached the desiredlanding area. This simulator was placedunder the direction of my division,although I was not involved in its con-cept or design. Most of the planning forthe Apollo mission had been based onthe idea of computer-controlled naviga-tion based on either astronaut measure-ments of angles to the Sun, Moon, and



stars, or on Earth-based radar mea-surements. This idea was opposed byCharles H. Zimmerman, an earlyLangley employee who had made nota-ble contributions to stability and controlresearch by originating the free-flighttunnel and the spin tunnel and by pub-lishing reports on airplane stability the-ory. He later left to join the ChanceVought company where he supervisedthe design of the Vought F5U-1 “FlyingTurtle” VTOL fighter airplane, but laterreturned to Langley after this projectwas cancelled. As a result of his longexperience, his opinions carried consid-erable weight among the center supervi-sors. For a time, he was assigned tothe Director’s office but later wasappointed as assistant head of my divi-sion, the Aerospace Mechanics Divi-sion. Zimmerman felt strongly that allthe complex computers and instrumen-tation planned for navigation of theApollo were unnecessary and that anastronaut could direct a space vehicle toits destination based solely on his visualcues, just as a pilot (at that time) oftendid in flying an airplane cross-country.This opinion was supported by ClintonE. Brown, an Assistant Director atLangley. As a result, they placed muchemphasis on building a simulator thatcould investigate the Apollo pilot’s navi-gational ability and could be used totrain astronauts to fly to the Moon in thismanner.

The simulator involved several originalfeatures. A camera was moved inresponse to the pilot’s inputs overa large map covered with three-dimensional images of the lunar terrain.The lunar craters and other featureswere based on photographs of theMoon and machined into Styrofoam®

blocks with a computer-controlled mill-ing machine. The terrain modeled cov-ered a path around the Moon in thevicinity of the orbit. The images photo-graphed by the camera were projected

on the interior of a sphere about 10 feetin diameter, where the pilot’s cockpitwas located. As the pilot flew the simu-lated vehicle, the view of the Moonslowly moved under the pilot andappeared as it would in an actual lunarorbit. The altitude had to be limited to aminimum value equivalent to a few hun-dred feet on the actual Moon to preventthe camera from bumping into some ofthe higher features on the map.

The construction of the simulator wasvery expensive and time-consumingbecause a simulator based on theseideas had not been constructed before.The terrain map was constructed oflarge blocks of Styrofoam®, which werecarved by digital milling machines ormolded to the correct shapes. A specialTV camera with multiple lensesprovided a wide field of view. Thepilot’s control inputs were fed into acomputer that calculated the attitudesand orbital motions of the vehicle.The camera moved along a track, withits height and attitude controlled byservomechanisms (ref. 5.2).

Because the project was supported byhigh officials at Langley, funds and per-sonnel were provided to keep theproject moving. The engineers said:“what LOLA wants, LOLA gets,” thewords of a popular song of that period.When the simulator was finally tested, itworked as planned, but the resultsproved rather disappointing. In typicaltest runs, the pilot sat in his cockpit forup to an hour with little to do but watchthe landscape go by. When the pilotstarted his descent to land, the simula-tor cut off to protect the camera. Themain benefit gained from this projectwas the later development of simulationequipment suitable for aircraft landingstudies. The lunar terrain was replacedby a picture of the airport. The samecamera and projection sphere wereemployed to study the ability of a pilot toland an airplane under various condi-

Docking Simulator


tions. Some years later, electronic dis-plays were developed and the largeterrain map was no longer required.

I wrote a memorandum to Charles J.Donlan, then the Director of Researchat Langley, pointing out the fallacy in theuse of piloted control for navigation of aspace vehicle. Because of the largeamount of fuel required to put eachpound of weight into orbit, missionsmust be planned to use an optimal tra-jectory to minimize the amount of fuelrequired. This system is possible on aspacecraft because the trajectory, influ-enced mainly by the accurately knowngravitational attraction of the Earth,Moon, Sun and other heavenly bodies,can be calculated with extreme preci-sion. Disturbance from other sources,such as atmospheric drag, solar wind,radiation pressure, and magnetic effectsare ordinarily very small, and can beallowed for by small corrections to theideal trajectory. By contrast, an airplaneflying across the country is subject tolarge unpredictable disturbances due towinds and gusts, weather variations,and so on. Sufficient excess fuel isalways carried to allow flight to an alter-nate airport, usually at least 500 milesfrom the scheduled landing field. Thisexcess fuel allows the errors involved innavigation by a human pilot to betolerated. In space, on the other hand,automatic guidance equipment andcomputer-controlled corrections for anyerrors in the trajectory must be used. Asan example of the margin of errorthat can be tolerated, Astronaut NeilArmstrong, in making the first mannedlunar landing, had to make a slightmanual correction during his finallanding approach to avoid a crater. Thelanding was made with just 3 secondsof fuel remaining, yet no one seemedconcerned about this small tolerancefor error. I have heard estimates ofthe fuel remaining ranging from 3 to20 seconds.

Docking Simulator

Rendezvous and docking with a targetvehicle was one of the main objectivesof the Gemini program because it wasrequired in the Apollo program when theupper part of the LM was propelled upfrom the Moon and docked with thecommand and service modules in lunarorbit for the return trip to the Earth. Thefinal stage of the docking involved man-ual control by an astronaut becauseprecise control was required for theclose alignment and gentle impactrequired for a successful docking.

The docking simulator, suspended fromthe roof of the large flight research han-gar in the west area of the LangleyResearch Center, was built to explorethe problems of this operation. Arthur W.Vogeley proposed the concept for thisdevice and supervised the personnelwho made most of the tests with it. Atarget vehicle was suspended at oneend of a track attached to girders in thehangar roof. A movable vehicle wasdriven along the track to approach thetarget during a docking maneuver. Agimbal system was suspended from themovable vehicle by an ingenious systemof cables originated by Vogeley. Thecables were arranged in V-shaped pairsthat maintained the outer gimbal in ahorizontal position while keeping itdirectly under the movable vehicle. Thegimbal system could be moved up anddown by reeling in each pair of cablessimultaneously on a large pulley in theoverhead vehicle. Inside the gimbal sys-tem was the approach vehicle, which formost tests was a replica of the Geminicapsule.

The gimbal system with the capsulemounted inside was an existing pieceof equipment that had been built tostudy the pitch, yaw, and roll oscilla-tions of airplanes in lateral maneuvers.The gimbal angles were controlled by



electrohydraulic servomechanisms thatwere operated through an analog com-puter system with inputs from the pilot’scontrol stick. The main application of thedocking simulator was to study sightingand target devices that would allowthe astronaut to sense his rate ofapproach and alignment with the targetvehicle. The most successful devicesincorporated a “two-layer” device in thetarget vehicle, such as a cross mountedahead of a small circle that would allowthe astronaut to judge when he wasproperly aligned. In recent years, thedocking simulator was honored as aState Historical Landmark. A photo-graph of the docking simulator is shownin figure 5.3-1.

Aids for Extravehicular Maneuvering

Two devices were proposed for extrave-hicular maneuvering, one by Harold I.Johnson and one by John D. Bird. Indiscussing the reasoning behind thesedevices, the approaches proposed bythe Johnson Space Center and theircontractors should be kept in mind. Ingeneral, the devices considered by thespace center consisted essentially of astrap-on vehicle that could hold theastronaut in his space suit. The vehicleitself had a stabilization system withgyroscopes and automatic controlequipment, a set of rockets to allow pro-pulsion and attitude control, and twohand controllers, one for attitude control

L-64-4307

FIGURE 5.3-1. Docking simulator.

Aids for Extravehicular Maneuvering


and one for translation. Essentially, theproposed vehicles were completespacecraft that could fit around anastronaut. No one doubted that such adevice could be made to work, but thecost of manufacturing such a devicewould run into millions of dollars. Thefrugal NASA engineers at Langleythought that much simpler and lessexpensive devices could be made toaccomplish the same purpose, while atthe same time allowing the astronaut towork with a less encumbering piece ofapparatus.

Harold I. Johnson called his device a“Space Gun.” It consisted simply of atank of compressed gas and a hand-held tube with a nozzle at the end. Theastronaut was intended to hold the tubeand control the flow of gas with a trig-ger-type controller. Harold Johnsonbelieved that the astronaut could readilylearn to control his attitude and motionthrough space instinctively by directingthe thrust of the device properly withrespect to his center-of-gravity position.

John D. Bird’s device (called “jet shoes”)was inspired by swimmers using swimfins. He proposed putting a pair of noz-zles on the astronaut’s shoes that wouldallow control of the thrust by valvesoperated by the astronaut’s toes anddirected as required by motions of hislegs and ankles. This method had theadvantage of allowing complete free-dom of the astronaut’s arms and hands.

These devices seemed to me sounusual that simulation of their opera-tion would be required before they couldbe considered for use in space. To illus-trate the varied opinions of the value ofsimulation studies, Hartley A. Soule, anearly expert on aircraft control and han-dling qualities, said that he had no doubtwhatever that such devices could beoperated by an astronaut, and did notsee the need for any simulation tests.

For a time this opinion put an end totests of these ideas at Langley.

Harold Johnson, perhaps because ofhis desire to try out his concept, trans-ferred to the Johnson Space Center. Intime, the “space gun” was assigned aposition as an experiment on one of thespace flights involving EVA, and a brieftest in space was made. The astronautinvolved thought that the device per-formed successfully. So far as I know,no further use was made of the device.The Johnson Space Center personnelcontracted for the construction of the“space vehicle” type of device, and itwas tested in a large simulation facilityconstructed by the Martin Company.This facility, somewhat on the sameprinciple as the Langley Lunar LandingFacility, suspended the astronaut inhis chair and simulated its motions in azero g environment. The simulationfacility allowed perfecting the character-istics of the device. This system hassince been used for all tests involvingEVA. The astronauts expressed prefer-ence for the vehicle-type device, proba-bly because of a greater feeling ofsecurity obtained in a vehicle that wasautomatically stabilized.

Don Hewes undertook the tests of JohnBird’s idea of control through footmotions. He made a very flat andsmooth floor by pouring slow-dryingepoxy cement on a flat floor area thathad boundaries at its edges. Donthought Bird’s idea of thrust control bytoe switches was rather far-fetched.Instead he used a hand controller, andfitted the astronaut’s shoes with nozzlesthat could fire fore or aft. The astronautlay on his side on a framework sup-ported by three air bearings. Thesebearings were flat disks that had smallholes on the bottom through which com-pressed air was admitted. These bear-ings supported the weight of theastronaut and his equipment so that thesystem was floating on a cushion of air.



Similar air bearings were used bynumerous people at Langley for spacesimulations. They reduced the friction tozero, so that the motion on the floor pro-vided a two-dimensional representationof the motion in space. Hewes foundthat motion over the flat floor with thefoot-operated device could be controlledvery easily.

To avoid the limitation to two-dimen-sional motion, several organizationsbuilt large water tanks in which theastronauts in their space suits, loadedto a condition of neutral buoyancy, couldpractice operations required in space.This method eliminated any steadyforces on the astronaut in a motionlesscondition, but of course it introducedrather large drag forces opposing anymotion. Nevertheless, the methodproved effective because most motionsused in EVA were slow. Art Vogeley builta water tank at Langley 30 feet deepand 30 feet in diameter, in which the

Langley test pilots could don spacesuits and experience the feeling of zerogravity. These tests were soon stoppedbecause the space centers had largerfacilities, more trained personnel, andbecause of safety considerations.

A fixed-base, 6 degree-of-freedomvisual simulator utilizing a projectionsphere, to be described later, was alsoused to test Hewe’s foot-controlledmaneuvering device. These tests weresuccessful in demonstrating the abilityof a pilot to control tumbling motion inspace. Tests of the foot-controlledmaneuvering unit were later made in theSkylab by astronauts Alan Bean andGerald Carr during space missions. Thehand-controlled unit proposed by theJohnson Space Center was also testedin these missions. As mentioned previ-ously, the astronauts preferred the morecomplex vehicle made under contract bythe Martin Company.


CHAPTER 6 My Work With Johnson Space Center

When the Space Task Group moved tothe Johnson Space Center, they soonwere in a period of rapid expansion andwere involved in contract work to pro-duce ground facilities and space vehi-cles for the space program. Closecontact was maintained with Langleyresearch engineers to provide assis-tance with research and developmentproblems that arose. Axel Mattson, aLangley engineer with long experiencein aerodynamics and in facility develop-ment, was appointed as the liaisonman stationed at Johnson Space Cen-ter to maintain contact with Langleypersonnel.

Dr. Christopher C. Kraft, an engineerwho worked under me in the Stabilityand Control Branch of the FlightResearch Division, joined the SpaceTask Group and moved to the JohnsonSpace Center. He first held a position asFlight Controller and was instrumental indesigning the Flight Control Centers atCape Kennedy and later at Houston aswell as in directing many early flights.Kraft was an excellent administrator andheld successively higher positions at theJohnson Space Center until he becamethe director of the Center shortly afterthe start of the Shuttle program. Krafthas written a book describing his workas a flight controller (ref. 6.1). Kraft

always very generously gave me creditfor his early training in technical work.When problems arose at Johnson, hefrequently called on me to assist withtheir solution. I did not always have theknowledge to provide the solutions tothese problems, but I was usually able,through my contacts at Langley, to findthe necessary information or at least togive a reasonable opinion as to the workrequired. I cannot present in detail allthe problems on which I worked, but Iwill summarize some of those that seemmost important.

Design of Shuttle Control System

My first important work connected withthe Johnson Space Center concernedthe Shuttle control system, although Ihad made previous visits to that centerto discuss simulators for the Apollo. Abrief review of the background of theShuttle development is desirable to givea basis for the various problems thatarose.

The Johnson Space Center was animpressive place during the Apollo pro-gram. It had a beautiful campus withlarge office buildings and laboratories.No expense had been spared to make


My Work With Johnson Space Center

Apollo a success. As the Apollo pro-gram neared a close, however, theJohnson engineers started to give seri-ous thought to a follow-on program. Per-haps the most popular proposal was tobuild a space station, which in turnrequired a shuttle vehicle to supply it.The Apollo program was a great suc-cess. It beat the Russians to the Moon,completed its program on schedule, andprovided much important scientific data.After the sixth Apollo flight, however, theprogram had lost its political popularity,and Congress failed to appropriatefunds for a seventh flight, even though aSaturn launch vehicle had already beenbuilt to carry the Apollo vehicle to theMoon on its seventh mission.

With the abrupt halt of the Apollo pro-gram, funding for a space station wasnot available. As a possible alternative,the management proposed a shuttlevehicle as a first step toward a spacestation. Such an interim project wouldkeep the Space Center in operation. Lit-tle study had been made, however, onthe design of such a vehicle.

A young engineer in the Space TaskGroup, Max Faget, had proposed thedesign of the Mercury capsule, and laterextended the capsule concept to theApollo vehicle. His proposal for a spaceshuttle also provided the basic conceptfor the Space Shuttle that later wasactually built. Faget, a model airplaneenthusiast, brought in a balsa and tis-sue paper model of a shuttle that incor-porated an unswept, low aspect ratiowing, a rather fat fuselage, and a tailwith a wide-chord elevator that providedsufficient control power to trim the vehi-cle to a very high angle of attack, proba-bly 45°. The concept was to enter theatmosphere at a very high angle ofattack to reduce the heating on theleading edges and lower surface of thevehicle, and later to pitch down for aconventional landing at an airport.Because of the capability of an unswept

wing to provide a relatively high maxi-mum lift coefficient at subsonic speeds,the vehicle would have had a landingspeed low enough to land at practicallyany airport.

I am not sure of the sequence of eventsthat followed, but wind-tunnel tests ofFaget’s concept showed that at hyper-sonic speeds, very strong shock wavesformed at the intersection of the wingand fuselage that would have caused aheating problem on the side of the fuse-lage. Later, the proposal for a shuttledesign was sent to three contractors.The winner, Rockwell, Inc., proposed adelta-wing configuration with a largefairing between the wing root and thefuselage and included a large elevatorcapable of pitching the vehicle to highangles of attack at hypersonic speeds.This configuration avoided the exces-sive heating on the sides of the fuse-lage, but it was capable of only relativelylow values of lift coefficient for landing.As a result, long runways would berequired for landing. Later, the highlanding speeds resulted in problemswith brakes overheating and tiresexploding, but these problems wereovercome with new designs for brakesand tires.

I have often thought that more studyshould have been made of the problemsinvolved in Faget’s original design. Pos-sibly wing fillets and root fairings couldhave overcome some of the fuselageheating problems. The capability to landon shorter runways would certainly havebeen a safety feature in later operations.

My first involvement in the design of theShuttle involved the control system. TheShuttle was one of the first airplanes toincorporate a complete digital fly-by-wire control system. The term fly-by-wire means a system in which all controlsurfaces or other control componentsare operated by electrical signals sentthrough wires. A digital system means

Studies of Shuttle Control System


that control signals are generated by adigital computer that contains all thenecessary control laws and signal trans-mission devices. Previous airplanes hadused fly-by-wire control systems usinganalog components and usually with amechanical back-up system. An exam-ple is the highly successful control sys-tem on the Concorde. In the case of theShuttle, however, a digital system wasconsidered necessary because of thewide changes in the control lawsthroughout the flight required by thewide range of Mach numbers and flightconditions encountered. Mechanicalsystems consisting of control cablesand pushrods would have had to handleexcessive forces and would haveencountered problems due to heating.The development of digital computershad reached a state that was consid-ered to have adequate capability andreliability to perform the control task,though much development was requiredto overcome the new problems encoun-tered.

To verify the design of the control sys-tem, a large working mock-up of thesystem was built at the Johnson SpaceCenter. One of the lead engineers inthis project was Robert G. Chilton, whohad worked under me in the FlightResearch Division at Langley. He gaveme a tour of the facility and a briefingduring one of my visits to the Center.The system used three digital comput-ers designed and programmed by theHoneywell Corporation. Though eachcomputer was highly reliable, the reli-ability requirements are such that theycannot be met by a single computer.The consequences of a failure are sosevere that the system is required toperform safely for millions of flight hours,representing the lifetime of not just onevehicle but of a whole fleet of vehicles.This degree of reliability can be met byusing the principle of redundancy: thatis, three or more computers perform the

control task simultaneously. If any onecomputer disagrees with the other two,it is immediately shut down and repairedafter the vehicle lands. If a secondcomputer fails, a comparison of the out-put of the two computers is made and acheck based on expected output orother means picks the best remainingcomputer.

I was aware of an experimental triplyredundant digital control system thathad been installed in a helicopter atMIT. Despite the redundancy, thissystem had failed in flight. The problemwas that a programming error hadoccurred in the software for the comput-ers. The computer program was identi-cal for the three computers. When thiserror was encountered, all three com-puters shut down simultaneously, leav-ing the helicopter without a controlsystem. I discussed the problem withDr. Raymond C. Montgomery, an engi-neer in my division who knew moreabout computers than anyone else inthe division. We concluded that the onlyway to avoid such a problem on theShuttle was to install a fourth computerprogrammed by an independent group.The problem was discussed with per-sonnel at the Johnson Space Center. Asa result, a fourth computer was pro-grammed by personnel at the DraperLab. It was ruled that the fourth com-puter, like the other three, was a safetyof flight item. Thus, the Shuttle could notbe launched unless all four computerswere working properly. Once, while pre-paring for a Shuttle launch, the fourthcomputer malfunctioned. The launchwas delayed until this computer hadbeen fixed.

Studies of Shuttle Control System

The Shuttle system was designed to befully automatic except for the final stage



of the approach and landing, when theastronauts took over. The initial designof the system software, when tried on asimulator, resulted in the Shuttle diverg-ing to large angles when subject to cer-tain disturbances. Such a divergencewould be catastrophic, and thereforecaused considerable concern amongthe designers. At this point I was calleddown to Houston by Dr. Kraft to studythe system and work with the designersto obtain a satisfactory system.

I worked with Ken Alder, a contractorfrom Lockheed who obviously had anexcellent knowledge of stability andcontrol. The work was mainly educa-tional for me because considerable workhad already been done to provide a sat-isfactory system. Later, I observed testruns on a simulator that traced the re-entry trajectory of the Shuttle andallowed the study of the effects of gusts,cross winds varying with altitude, and soon to provide a sufficient margin ofsafety for all conceivable disturbances.

As mentioned previously, the vertical tailof the Shuttle, mounted on top of thefuselage, became ineffective at highangles of attack. A method was workedout to provide directional control usingonly the elevons during the high angleof attack part of the entry, which lastedfrom the start or the descent at aboutMach 23 to Mach 2. By this method, theelevons moved differentially in theopposite direction from what would berequired for roll control at subsonicspeeds. To yaw to the right, for example,the right elevon would move down (or toa smaller upward deflection), increasingthe drag of the right wing. The leftelevon would have the opposite move-ment. This right yaw (or left sideslip),because of the large dihedral effectof the delta wing, would cause theShuttle to roll to the right despite the leftrolling moment caused by the elevondeflection.

At low supersonic Mach numbers, whenthe aerodynamic heating was reduced,the Shuttle would be pitched to a lowerangle of attack, so that the rudder wasunshielded and became effective foryaw control. In this regime, the elevonswere moved differentially in the normaldirection for roll control, opposite fromwhat had been done at higher Machnumbers. The rudder then served to off-set the yawing moments from the differ-entially deflected ailerons as well as toprovide yaw control. The controls werethen in a normal configuration so thatthe astronauts could take control whenthey made the flare and landing.

The variation of control laws with Machnumber is an example of a change thatcan be made readily with a digital con-trol system but would require somecomplex device with a mechanical sys-tem. In fact, the digital system provideda smooth ramp-like reversal in theelevon control when the shift was made,and in addition provided control gainvariations from various sensors such asrate gyros and angle of attack and yawsensors throughout the descent. Thestudies being made to improve thesafety and control effectiveness of thesystem were mainly concerned with theadjustments of these gain values for thevarious phases of the descent.

After the changes in the software hadbeen worked out, I spent some timewatching the runs on a simulator, alongwith one of the engineers. This simula-tor just plotted a trace of the trajectoryon a screen. On one run, I observedthat the trajectory near the end ofthe run diverged. The engineer, whohad not been watching, was quite sur-prised. After analyzing the data, he con-cluded that this fault had been observedbefore and had been corrected, butthe change had not been made in thelatest version of the program. This expe-rience illustrates the vigilance required

Problem With Pilot-Induced Oscillations


in preparing software for computers in aflight vehicle.

Problem With Pilot-Induced Oscillations

To give the astronauts practice landingthe Shuttle before going through thedangers of a complete orbital flight, anumber of flights were made in whichthe Shuttle was mounted on a trapezeon top of a Boeing 747 airplane andwas released from an altitude of about10,000 feet to glide to a landing on thelong runway at Edwards Air Force Base,now the Dryden Flight Research Center.In the fifth landing, the Shuttle made anormal flare, but just before touchdownit made two or three rather violent short-period oscillations. The Shuttle landedsafely after the co-pilot had told the pilotto stop trying to control the oscillationand let the Shuttle land itself.

Oscillations of the type experienced onthe Shuttle are called “Pilot InducedOscillations,” (PIO). for short. Suchoscillations have been experienced onmany airplanes, starting with the WrightBrothers’ first flights, and became morefrequent with the introduction of hydrau-lically operated control systems. Thecause of these oscillations can often beanalyzed, and their reoccurrence can beavoided after they have occurred, butpredicting the tendency for oscillationsbeforehand can be much more difficultbecause they can result from many dif-ferent causes. In the case of the Shuttle,a detailed analysis was made on theresponse characteristics of the digitalcontrol system, and the system wasfound to have a response lag of 0.2 sec-onds or more following a pilot’s input.

Again I was called to the JohnsonSpace Center to help with the analysis.Dr. Robert Gilruth, who was the Directorof the Space Center, had run flying

qualities tests on 16 airplanes and, inthe early forties, had written a cele-brated report called Requirements forSatisfactory Flying Qualities of Air-planes. In this report he had stated thata lag in the control system response ofgreater than 0.1 second was unsatisfac-tory. Gilruth’s conclusion was based ontests of spoiler ailerons on the wing of alight plane. These spoilers were locatednear the leading edge of the wing.When they were abruptly deflected, theflow over the spoiler would initially havean effect similar to an increased wingcamber and would cause the lift toincrease. Then, as the boundary layerair collected behind the spoiler, the winglift would indeed be “spoiled” and theairplane would roll in the desired direc-tion. The pilots considered the resultinglag in response very undesirable.

Another problem with the Shuttle wasthe time required for motion of theelevons. These elevons were unusuallylarge and heavy and were operated byhydraulic motors. The time to make alarge deflection was appreciable and toreturn the elevon to neutral before mak-ing a movement in the opposite direc-tion also had to be considered.

A third characteristic of the Shuttle thatwas different from that of most airplaneswas the large, downward force appliedto the wing when the elevons weredeflected up. This force caused the cen-ter of gravity of the vehicle, as well asthe pilot’s cockpit, to move downwardfollowing application of nose-up control,and considerable lag occurred beforethe pilot could feel that he was rising.

Apparently the designers of the Shuttlehad not considered the importance ofthese effects. The 0.25-second lag inthe response was caused by the repeti-tion rate of calculations in the digitalcomputer. An obvious correction thatwas easy to apply was to approximatelydouble the repetition rate. This change



was made before subsequent glidetests, together with warnings to theastronauts to apply the controls gentlyand to allow sufficient time for flight pathchanges required in landing. Thesemeasures were apparently successful incorrecting the problem in subsequentglide tests, although one of the laterlandings showed some signs of anoscillation.

Flight Control Review Group

Because of the complication of theShuttle Control System and the numberof organizations involved in its design,the Flight Control Review Group wasorganized to coordinate the work of thevarious organizations. The Chairman ofthe group was Donald C. Cheatham, aformer Navy pilot and a member of theNaval Reserve, who had previouslyworked for me at Langley in the Stabilityand Control Branch of the FlightResearch Division and was now at theJohnson Space Center working on theShuttle Control System. I was a regularmember. Other members were S. Brayof the Ames Research Center, J. Weil ofthe Dryden Flight Research Center,R. G. Hoey of the Air Force Flight TestCenter at Wright-Patterson Air ForceBase, and about six members from vari-ous branches at the Johnson SpaceCenter. Meetings were held approxi-mately quarterly in the 1978–79 period.

The number of organizations working onthe Shuttle Control System and thenumber of independent simulators setup to study its problems far exceededthe numbers devoted to any otheraircraft. At least seven simulatorsdesigned to study the complete entry ofthe Shuttle from orbital flight to landingwere in existence. These included oneat Langley, one at the Dryden FlightResearch Center, one at the Air ForceFlight Test Center, and several at the

facilities of Johnson and its contractors.In addition, the Johnson Space Centerhad the Shuttle Training Aircraft (STA), aGulfstream with a modified control sys-tem to simulate as closely as possiblethe control characteristics of the Shuttle.The TIFS (Total In-Flight Simulator) air-plane at Calspan was also used to studyspecialized piloting problems. Just get-ting the various simulators to agree onthe same problem was a major task butwas solved with the aid of the largenumber of engineers involved.

The large number of independentgroups working on the flight controlproblem was an excellent method tocatch and eliminate errors in the pro-gramming as well as to take advantageof the knowledge of the experts workingfor the various contractors. The methodwas also very expensive, but the sup-port for the Shuttle at that time was suf-ficient to allow such expenditures. Laterairplanes, both military and commercial,that made the first attempts at using dig-ital fly-by-wire control systems withoutthe aid of such intensive design effortsalmost always encountered failures orcrashes.

The problem of pilot-induced oscilla-tions was a major concern of the FlightControl Group. The use of an increasedrepetition rate in the computersundoubtedly helped the situation. Still,the slow response of the Shuttle tolongitudinal control inputs was a matterof concern. This lag in response wascaused by the large inertia of the vehiclein pitch and the short moment armbetween the elevons and the center ofgravity. I later made an analysis thatcompared the longitudinal response ofthe Shuttle with that of several otherlarge airplanes. The response time ofthe Shuttle was over twice as long asthat of any other airplane (fig. 6.1). Asshown by this figure, the download onthe elevon at the start of a pull-upcaused the center of gravity of the

Flight Control Review Group


Shuttle to move down. Only after sometime did the pitch angle increase the liftand cause the Shuttle to rise. A time lagof 2.15 seconds passed before the cen-ter of gravity had returned to its originalaltitude. This type of motion occurred onother airplanes with delta wings but wasnever as severe as on the Shuttle.

The sensation that the pilot experiencedof accelerating downward when thecontrol was applied for a nose-upresponse had been analyzed previ-ously by some researchers and hadbeen blamed as a cause of pilot-induced oscillations. I did not agreewith this conclusion. I pointed out thatseveral other airplanes, such as theGee Bee racers, had been flown withthe pilot sitting near the tail withoutencountering any difficulty. I felt that the

pilot could become accustomed to hislocation in the airplane and could visual-ize the response of the nose of the air-plane or of its attitude in pitch.

As usual, accurate predictions of thetendency to pilot-induced oscillationswas difficult. R. G. Hoey of the Air ForceFlight Test Center felt that despiteinstructions and practice to teach theastronauts to control the Shuttle withslow and deliberate movements, theywould someday hit a disturbance thatwould require a fast control movementthat would start a pilot-induced oscilla-tion (PIO).

A device called the PIO Suppression Fil-ter was devised by an engineer at theJohnson Space Center to make theShuttle more resistant to PIO. This filter,inserted in the control system after the

FIGURE 6.1. Comparison of height response of several airplanes: F-104 and B-52—conventional aft tail, YF-12, B-58, and Shuttle—delta wing. Note that the Shuttle requires 2.15 s for the c.g. to return to its original altitude.

F-104, CL = 1.2

YF-12, CL = 0.6

B-52, CL = 1.0

Shuttle, CL = 0.6

B-58, CL = 0.6

160

120

80

40

0

0 1 2Time, s

3 4–40

200

Altitude of c.g., ft



pilot’s controller, sensed the frequencyand amplitude of the motions of the con-trol stick. If these motions became toolarge or frequent, the amount of motionof the elevons for a given controllerdeflection would be reduced, eventuallyreaching only one-third of its normalvalue. This device was tested on simula-tors and on the Shuttle Training Air-plane. The astronauts commented thatwith the device in operation, it was verydifficult to produce a pilot-induced oscil-lation. There was some objection that ina critical situation the pilot would beunable to use full control, but in such asituation the danger of a pilot-induced

oscillation would be greatest. Afterextensive testing, the device wasinstalled in the Shuttle control system.

The subsequent series of successfullandings of the Shuttle can be attributedto the astronaut training program as wellas to the suppression filter. Most of thesuccess can probably be attributed tothe training program. The suppressionfilter has, to my knowledge, never beenforced to come into operation. Thus, thefears expressed by R. G. Hoey andother engineers were perhaps neverrealized in practice.


CHAPTER 7 Continuation of Research Following Decline of Space Program

Following the completion of the designand construction of the Space Shuttle,administrators of the program hadexpected a continuation of active spaceactivity involving manned space flight.Changes in the policies of the presi-dents who later came into office and thegeneral lack of public interest in spacedevelopments resulted in a rapid declinein funding for this work. The Shuttle wasa technical success, but the expense ofits operation and the lack of a major pro-gram that required its use caused areduction in the frequency of flights toless than four per year, instead of everytwo weeks, as first envisioned by somespace enthusiasts. A temporary spacestation, the Skylab, using one of theSaturn tanks as its major component,stayed in operation for a while, but eventhis vehicle was allowed to fall back intothe atmosphere and burn up rather thanbe sustained in orbit with a small expen-diture of fuel.

I had continued to study some aeronau-tical problems during the space programand following the decline in space activ-ity, I resumed work to clean up someproblems that had not been completedbefore the space program, and madestudies of a number of new aeronauticalproblems. The remainder of this volume

gives brief accounts of a number ofthese problems.

Variable Sweep Wing Supersonic Transport

A variable sweep wing appears desir-able for a supersonic airplane becausefor takeoff and landing an unswept wingof high aspect ratio provides lower take-off and landing speed, whereas atsupersonic speeds a highly swept winghas much less drag. The X-5 experi-mental airplane was perhaps the firstthat allowed us to study a variablesweep wing in flight.

A problem with the variable sweep wingis that as the wing is swept back, theaerodynamic center of the wing movesback with respect to the center of gravityof the airplane. To balance the airplane,a large download on the stabilizer isrequired in the swept condition, result-ing in high drag. Alternatively, if theairplane is balanced in the swept condi-tion, it becomes longitudinally unstablein the unswept condition. This problemis particularly critical on a supersonicairplane because the aerodynamiccenter of the wing also moves backby about 25 percent of the chord at


Continuation of Research Following Decline of Space Program

supersonic speed because of theeffects of high Mach number.

To avoid this problem, the X-5 had thewing mounted on a movable cradle so itcould be moved back or forward to suitthe flight condition. Such a system,however, occupies an undesirably largevolume in the fuselage.

In 1948, I made a model glider simulat-ing a supersonic transport with avariable sweep wing, a picture of whichis shown in figure 7.1. To keep the

model in trim with the wing swept back,I added a retractable delta surfacenear the nose. Suitable hills to testmodel gliders are not common in thearea near Hampton, Virginia, where Ilive, but I found a suitable small hill onthe Yorktown Battlefield and spent anafternoon gliding the model. The modelglided well either with the wing sweptand the delta surface extended or withthe wing unswept and the delta surfaceretracted. In this glider, the wing hingewas at the centerline of the fuselage.

NACA LAL-57436

NACA LAL-57437

FIGURE 7.1. Figure 7.1. Glider model of super-sonic transport with variable-sweep wing and retractable canards. Length, 30 in., span (swept) 15 in., span (unswept) 26 in.

(a) Wing swept, canards extended.

(b) Wing unswept, canards retracted.

Differential Maneuvering Simulator (DMS)


Later, airplane companies discoveredthe idea that if the wing panels werehinged at a point some distance out onthe wing, the sweeping action wouldmove the inboard section of the wingfarther forward. This principle was usedon such airplanes as the Air ForceF-111 and F-14 Navy fighters and onthe B-1 Bomber. This system has beensuccessful, but it poses complicatedstructural problems for the wing mount-ing and for fairings at the wing root.


After graduating from MIT, I kept intouch with another MIT student namedHerbert K. Weiss. He and I both came toHampton, Virginia, I with the NACA andhe with the Coast Artillery Board at FortMonroe. He was a brilliant engineer andmathematician, and I often consultedhim on my problems. Later he went towork at the Aberdeen Proving Ground inMaryland. His work was primarily inmissile design, vulnerability of militaryaircraft, and air combat problems. Onone occasion, probably in the fortiesafter WWII, he asked me to come withsome of my engineers to discuss mutualproblems. During the meeting, he sug-gested that it would be possible to buildan air combat simulator with two cock-pits, two pilots, and displays showingthe image and motion of the opponent’sairplane to each pilot. I thought aboutthis idea but concluded that the NACAat Langley did not then have the facili-ties or funds to build such a simulatorand that the state of simulator designhad probably not advanced far enoughto undertake such a project.

In 1965, when the LOLA simulator waswell underway, an engineer namedDr. John D. (Jay) Bird, who was abranch head in my division, made a sim-ilar suggestion for an air combat simula-

tor. At this point, there was muchinterest in air combat as a result of thewars in Korea and Vietnam. Jay Birdwas an ingenious research man. I con-cluded that with his enthusiasm it wouldbe possible to undertake an air combatsimulator. To gain experience, a simula-tor with just one projection sphere,called the Tactical Effectiveness Simula-tor (TES), was built. It had a projectionsphere about 20 feet in diameter con-taining the pilot’s cockpit and a projectorto produce an image of the target air-plane. The target airplane was con-trolled by a pilot in an external cockpit.All the motions were controlled by alarge, digital simulation installationrecently installed at Langley.

A contract for building the TES was wonby the Rheem Corporation, a companybetter noted for its work on heaters andair conditioning systems, that had previ-ously worked on simulators for the AirForce. I was surprised that they did anexcellent job in designing the TES, andespecially the detailed solution of themathematics of the motions to besolved by the digital computer. Themotions in three dimensions weresolved by the use of quarternions, asystem that avoided any gimbal lock inthe projectors.

The TES was highly successful and wasused for several research studies. Theplans for the Differential ManeuveringSimulator (DMS) went ahead rapidly. Acontract for its construction was let tothe Northrop Corporation. An artist’sdrawing of the simulator is shown in fig-ure 7.2-1. A later drawing that is per-haps more accurate in certain details isshown in figure 7.2-2.

The simulator had two large projectionspheres, each 40 feet in diameter, onthe inside of which was projected thetarget airplane image, view of the Earthand sky, and a Sun image. The pilot satin a cockpit which did not rotate but was



capable of oscillating to simulate buffet-ing. In addition, the pilot could wear a“g suit” that could be inflated to simulateg loads. The image of the target air-plane was simulated by a detailed scalemodel about 15 inches long that was

suspended in a box by a system ofwires moved by servos to rotate the tar-get airplane in roll, pitch, and yaw andwas photographed by a camera topresent a correct image of the target air-plane in the projection sphere. Motion of

FIGURE 7.2-1. Pictorial view of Differential Maneuvering Simula-tor (DMS). The large spheres are projec-tion screens, and they contain the pilots’ cockpits, projectors, and servomecha-nisms to move the projected displays. Each pilot sees a view of the opposing airplane.

FIGURE 7.2-2. More detailed sketch of Differential Maneuvering Simulator.

Vehicle“A”

simulator

Vehicle“B”

simulatorProjection system

Cockpit

Buffetsystem

Target imagegeneration system(physically locatedbehind spheres)

Computer complex(in adjacent building)

Computer console



the entire scene through an angle of270° was produced by a large hydraulicservomechanism rotating the previ-ously mentioned projectors to producethe image as seen from a maneuveringairplane. All these effects were dupli-cated in the other sphere so that eitherpilot could be the attacker as the com-bat progressed.

The simulator was completed andplaced in operation in July 1969. By thetime the simulator was finished, Jay Birdhad lost interest in the project, and I wasfaced with selecting a group of engi-neers to run the programs. Continualhelp in operating the mechanics of thesimulator and modifying the digital com-puter programs was provided by theAnalysis and Computation Division(ACD). The simulator cost $5.5 million.The successful completion of the DMSwas valuable to both the Air Force andNavy, who started construction of theirown combat simulators. These simula-tors were more complicated, with suchfacilities as moving cockpits and thecapability of displaying four airplanes.Both were finished after the DMS andcost many times as much.

I once thought that a simulator such asthe DMS would be useful in developinga theory of air combat. Consultation withsome noted mathematicians who hadconsidered this problem, however,showed that it was an extremely difficultproblem. Certainly it was beyond thecapabilities of any of the engineers atLangley. The simulator was very useful,however, in developing empirical theo-ries and in checking combat strategiesthat had been developed by militarypilots. After some initial runs, pilots fromthe Air Force and Navy were invited tofly the simulator on a regular schedule.These pilots all considered the simulatorruns extremely beneficial in improvingtheir flying techniques. A series of rulesfor air combat developed by Al Meintelwas useful in training these pilots.

Perhaps the most useful analyticalstudy to come out of the simulationstudies was a program developed byGeorge H. Burgin, Lawrence J. Fogel,and J. Price Phelps of Decision Sci-ence, Inc., San Diego, California, work-ing under contract to NASA (ref. 7.1).This program acted as an artificial pilotto serve as an opponent for a humanpilot in the simulator. Later, Walter W.Hankins III, an engineer at the LangleyResearch Center, gave an AGARD talkon the program and wrote a LangleyWorking Paper summarizing the resultsof tests with the program (refs. 7.2and 7.3). This program was so good thatexperienced military pilots were oftenbeaten by it, and it was only after mak-ing a number of runs to uncover slightweaknesses or peculiarities in the pro-gram that they were able to beat itconsistently. This type of program, ofcourse, suggested its use as an auto-matic pilot to replace the human pilot inair combat. Such a program could savemany pilots’ lives in actual warfare. Thepilots, however, always opposed the useof such a method. They had trained tobe combat pilots and did not wish toleave the job to a machine, even if itmight have saved their lives. Only inrecent years when the use of unmannedvehicles for military missions hasreceived increased attention has theuse of automatic combat pilots beenseriously studied.

The DMS is still in operation (2004) andhas served many useful purposes inaddition to studying air combat. Forexample, the use of a single sphere tostudy spinning characteristics of air-planes has allowed investigation of amuch wider range of characteristicsthan could be done in flight tests. Sincemy retirement in 1979, I have been outof touch with simulation work, so I amnot familiar with much of the work thathas been done.



Circulation Control

In my own experiments on gust allevia-tion, as well as in most other studiesthat have been made, flaps on the wingsare moved up and down to offsetupward and downward gusts, respec-tively. This method works well in anexperimental study. On a practical air-plane, flaps must also be used for gen-erating high lift for landing. On transportairplanes, complex two-segment, dou-ble-slotted flaps are commonly used,which would complicate the problem ofusing flaps for gust alleviation. Possibly,the rear segment of the existing doublesegment flaps could be used for thispurpose, but no aircraft company, to myknowledge, has attempted to study suchan arrangement.

Another method to produce high lift,known mostly from wind-tunnel tests, iscirculation control produced by a jet ofair directed downward at the trailingedge. With this method, very high val-ues of maximum lift, around six orseven, can be produced. Another tech-nique, known as the Coanda effect, hasbeen used for this purpose. In thismethod, a thin jet of air directed along asharply curved surface at the trailingedge can deflect the jet downward andproduce high values of maximum lift.

An engineer and assistant professor atMIT named Joseph Bicknell was at MITwhen I was studying there. Later hepublished a paper describing a methodfor producing oscillating flow in a windtunnel. In this method, illustrated in fig-ure 7.3-1, he used a cylinder, with its

FIGURE 7.3-1. Use of Coanda effect to deflect airstream at trailing edge of wing. Illustration shows effect of oscillating cylinder that closes one side of the open-ing or the other. A slightly smaller rotat-ing cylinder could alternately close and open gaps at a high frequency, producing a rapidly oscillating airflow.

Compressedair supply

In practice, gaps would be only a few thousandths of an inch wide.

Cylinder turnedclockwise

Cylinder turnedcounterclockwise

Control Configured Vehicles (CCV)


axis slightly off center, rotating in anopening at the rear of a symmetrical air-foil that was mounted to span the windtunnel ahead of the test section. Com-pressed air was blown through the rearof the airfoil. The air jet, influenced bythe Coanda effect, followed first theupper and then the lower surface ofthe cylinder as a slot was opened atthe upper and lower surface of the air-foil. The cylinder could be rotated withvery little power and could produce anoscillating flow in the tunnel at very highfrequencies.

With slight modifications, this methodcould be used as a control for gust alle-viation. A source of compressed air,such as bleed air from a turbojet engineor a separate compressor, would berequired to produce the air jet. The cyl-inder would not rotate continuously, butwould oscillate about its axis just like theflap on a gust-alleviation system. Thisaction would open a slot up or down todeflect the flow as required to produce avarying circulation about the airfoil.

A possible advantage of this systemwould be that the deflected flow couldalso be used as a landing flap, takingadvantage of the very high values of liftcoefficient that can be obtained with ajet at the trailing edge.

Plans were underway at the LangleyResearch Center to try this system on aCessna 402B airplane, but the fundingfor the project was cancelled. A patentwas taken out on the system in thenames of Eric Stewart and myself. Inview of the continual disinterest of theairplane companies in gust alleviation, Ido not believe that this system is widelyknown.

Control Configured Vehicles (CCV)

The term CCV was introduced about1969 by personnel of the Air ForceFlight Dynamics Lab to describe thedesign of an airplane to obtain perfor-mance improvements by use of auto-matic control systems. The term wasmuch publicized about that period,resulting in formation in a special com-mittee at Langley and an intercentersymposium on the subject. In general,the term implied an airplane loaded withelectronic equipment with rather poorlydefined benefits in performance. At therequest of the associate director ofLangley, Lawrence K. Loftin, I wrote amemorandum describing my ideas onthe benefits obtainable with this system.

The most apparent benefit was longitu-dinal stability augmentation, allowingthe use of a smaller horizontal tail or amore rearward center-of-gravity loca-tion, thereby reducing drag. This benefitwas so obvious that some airplaneswere already being designed with thisfeature. Many other suggested benefits,however, required severe compromiseof other essential features of the air-plane and were rarely used.

After considerable thought, my overallconclusions were summarized in apair of charts shown as figures 7.4-1and 7.4-2. The main point was that thesizes of vertical and horizontal tails ofairplanes and their associated controlsurfaces were based mainly on emer-gency flight conditions such as spinrecovery, stall avoidance and recovery,and control with asymmetric power. Thereductions in tail sizes offered by CCVwere therefore not usable. As illustratedon the second chart, the control powerrequired for increased stability was usu-ally about one-third of that required foremergency flight conditions, and thecontrol power for improved damping of



oscillations was only one-ninth of thatrequired for emergency conditions. Thebenefits of CCV could be increased onlyby designing the airplane to reduce thecontrol power required for emergencyconditions. Methods for doing this weredeveloped, but in the ensuing years,such methods have not been generallyadopted.

Propulsive Effects Due to Flight Through Turbulence

In a wind-tunnel study made in 1922,R. Katzmayr showed that an airfoil sub-ject to periodic vertical motions of air ina wind tunnel would experience a pro-pulsive force. A more practical problem

FIGURE 7.4-1. Control-configured vehicle (CCV) chart.

FIGURE 7.4-2. Damping and stability control chart.

W H A T I S A C C V ?

A CONTROL– CONFIGURED VEHICLE IS ANAIRPLANE FOR WHICH THE MOMENTS

REQUIRED FOR TRIM THROUGHOUT THENORMAL AND EMERGENCY FLIGHT

ENVELOPE ARE MINIMIZED

C O N T R O L R E Q U I R E D F O R D A M P I N GA N D S T A B I L T Y

TOTAL CONTROL DEFLECTION

DAMPING

STABILITY

Decoupled Controls


is to determine the effect of flightthrough random turbulence. The Frenchaerodynamicist Breguet made an earlystudy of this problem. His analysis wascorrect in principle, but at that time themathematical tools for analyzing turbu-lence, such as power spectral densityand the spectra of atmospheric turbu-lence, were not available. As a result,quantitative results were not obtained.This problem, though it is considered tobe of some practical importance, hadapparently never been solved. I madean analysis of this problem and pre-sented the results in a note in the Jour-nal of Aircraft (ref. 7.4). In this analysis,the spectrum of atmospheric turbu-lence, which expresses the amplitude ofthe gust intensity as a function of fre-quency, is represented by the familiarDryden spectrum, and the responsetime of the airplane in responding to avertical gust is placed in frequency-response form. These two quantities arethen combined to give a formula for thethrust coefficient.

The results show that a lightly loadedaircraft, such as a soaring glider, flyingat a high lift coefficient, experiences anincrease in lift and a forward tilt of the liftvector in upward gusts and a decreasein lift and a rearward tilt of the lift vectorin downward gusts. A net increase inthrust is produced, but such a lightlyloaded aircraft responds vertically to thegust very quickly, reducing the effects ofchange in angle of attack to a very shortduration. A heavily loaded aircraft, suchas a fighter or transport, is flying fastand in the same turbulence experiencesmuch smaller changes in angle ofattack, but the heavy airplane movesvertically much more slowly, so theduration of the change in lift is longer.The result is that both types of airplanesexperience about the same thrust effect,which is very small even in severe turbu-lence. Calculated results for typicalexamples show that a change in thrust

coefficient in severe turbulence wouldbe about 0.003 to 0.005.

Decoupled Controls

Though the early Wright Brothers’ air-planes used control systems that werenot consistent with the pilot’s normalreactions, the early pioneers Paulhanand Bleriot originated a system in whichfore and aft motion of the controller con-trolled the pitch control surface, side-to-side motion of the controller controlledthe roll control surfaces, and rudderpedal motion controlled the yaw controlsurface. A throttle was used to controlengine power or thrust. This system hasbeen used almost universally sincethen. With the development of automaticelectronic control systems, however,designers realized that other types ofairplane response could be obtained.One possibility was called decoupledcontrols, in which a given controllerwould control just one motion of the air-plane. For example, one controller mightcontrol only pitch angle, another mightcontrol airspeed, and another lateralvelocity. Such controls were sometimesthought to be easier for novice pilots tolearn, and in other cases, advantageswere claimed for gunnery or missileaccuracy. It should be realized thatmore than one airplane control surfacewould be used for each of the decou-pled motions.

Numerous analyses of controls with var-ious types of decoupling are available inthe literature, and flight tests of somearrangements have been tried, in partic-ular to study gunnery or missile accu-racy. No great advantage has beenfound for using a system different fromthe standard control system. In fact,some rather serious objections havebeen raised by pilots to these systems.For example, pilots use different controltechniques for upward or downward



control when flying near the ground.They object to any control which appliesdown elevator when flying near theground, even if the decoupled systemmight not result in a dangerous flightpath. Decoupled systems are usuallyunsuitable for emergency conditionssuch as spin recovery, where pilotshave learned techniques appropriateto a given airplane that might applycontrols opposite from those producedby a decoupled system. Most work ondecoupled controls appears to havedisappeared, although automatic sys-tems may be used for specific flightregimes on individual airplanes, such asstall-limiting devices or Mach numbercontrol.

Soaring Gliders

The Wright Brothers experimented withman-carrying gliders for three yearsbefore they made their successful flightin a powered airplane. Later, OrvilleWright, in 1911, returned to Kitty Hawkto try a glider with a tail-aft locationwhich, by that time, had been adoptedby most other aviation pioneers. Orville,slope soaring in the wind on the largedune near the site of their first flight,made a soaring flight of 9 minutes45 seconds, which stood as a worldrecord for soaring endurance until afterWWI.

After the war, construction of airplanesin Germany was forbidden. Many of theGerman manufacturers went in for soar-ing glider activities, both for sport andexperimentation. The design and perfor-mance of soaring gliders improved rap-idly. In this country, the center of glideractivity became concentrated in Elmira,New York, where a level area at the topof a steep hill provided a good site forlaunching. This area became known asHarris Hill, where a Lieutenant Harris

had been killed when his car overturnedwhile towing a glider.

The performance of soaring glidersdepended to a great extent on aerody-namic efficiency. As a result, manyaeronautical engineers did analyticalstudies in this field. There had beensome glider activity at the LangleyResearch Center before I arrived in1940, but very little was done after that.When the first radio-control systems formodels were developed in the early 40s,I made some of the first radio-controlledsoaring glider models, which served tokeep up my interest in this field. Later,about 1955, radio-controlled glidersbecame a very popular branch of modelaviation. About 1968, Oran Nicks, whowas a soaring enthusiast, became Dep-uty Director of Langley. He organized anannual technical conference on soaring,held at MIT, and always requested that Itry to prepare a paper to be given at theconference.

The first paper that I presented wasentitled Analysis of Effect of AsymmetricLoading on Sailplane Performance inCircling Flight (ref. 7.5).

The second paper that I presented wasentitled Gyroscopic Moments on aGlider in Turning Flight (ref. 7.6). Soar-ing gliders, because of their high aspectratio, have a strong tendency to roll to alarger bank angle when in a turn. Oppo-site aileron and rudder controls must beapplied to offset this tendency. Thiseffect was discovered by the WrightBrothers and resulted in their develop-ment of the coupled rudder and wingwarping lateral control system that wasa major factor in their success in makingcontrollable flights.

The gyroscopic moment acting on aglider results from the tendency of allrotating bodies to align themselves withthe plane of rotation. This moment tendsto reduce the tendency to roll to a largerangle of bank. This effect is not very

Soaring Gliders


large, however. In a typical example, thegyroscopic rolling moment was only13 percent of the aerodynamic rollingmoment.

The third paper that I presented wasintended to study the effect of the trail-ing vortices shed from the wing tips of aglider in circling flight on the forces act-ing on the glider in subsequent turns. To

study this effect, I made a flow visualiza-tion study. At that time, a large watertank, 30 feet in diameter and 30 feetdeep, was being used at Langley tostudy zero gravity effects on the perfor-mance of astronauts in space. I made amodel glider of 40-inch span out of solidmahogany and weighted with about 4 lbof lead in the fuselage. The glider was

FIGURE 7.5. Heavily weighted glider (40-in. wingspan) gliding slowly (about 3 ft/s) underwater in large tank. Fluorescent dye is emitted from wing tips. When glider made complete circles, trails from previous circle were always well above glider.

L-69-4307



launched underwater with a rubber cat-apult and would slowly glide in circles inthe tank at a speed of about 3 to 4 feetper second (fig. 7.5). A fluorescent dyecarried in a tank in the model wasexpelled through nozzles on the wingtips. This dye left a clear trace of thetrailing vortices from the model as themodel slowly descended in the tank.The conclusion was that the modelwas always well below the vortex shedin a previous circle. As a result, thevortices would be expected to have anegligible effect on the efficiency of themodel. These results were demon-strated to the attendees at the sympo-sium with movies.

As a matter of interest, I took the leadweights out of the model, equipped itwith radio control, and tried flying it inair. The model weighed about a pound,a rather heavy wing loading for a model.Launched from a dune into a strongwind of about 35 miles per hour, it wouldglide with a speed exceeding 40 milesper hour. I claimed that I had made thefirst glider that flew both in air andunderwater, like a duck. The ratio ofdensities of water and air is about 1000.

Another fact that I learned from theseexperiments was that the buoyancy ofthe parts of the model that are lighterthan water must be taken into accountin trimming the model for underwaterflight. For example, the tail surfacesmade of mahogany were buoyantunderwater, allowing the model to be intrim with a center of gravity that wouldbe too far rearward to be stable in air. Asa result, I had to make a lighter set oftail surfaces for flight in air. Buoyancyeffects also exist on the structure of aconventional airplane, but the effectsare so small compared to the weightof the airplane that they are rarelymentioned.

Complementary Filters

When considering the subject of aircraftcontrol systems, a filter is considered tobe a device that modifies the pilot’sinput to improve the response of the air-plane. In many cases, the filter isdesigned so that its presence is not feltat the pilot’s controller but is simplyrecognized as part of the airplanedynamics. For example, a longitudinallyunstable airplane may be made longitu-dinally stable from the pilot’s standpoint.

The simplest types of filters may pro-duce undesired side effects. For exam-ple, a filter that improves the longitudinalstability may introduce an undesiredstructural oscillation. To avoid suchundesired effects, a device known as aComplementary Filter has been intro-duced. This system has relatively simpledesign, and does not require compli-cated mathematics for its analysis, yethas many applications in aeronauticalproblems.

During the Apollo and Shuttle programs,mathematicians introduced filters toobtain an optimal flight path or otherobjective based on some criterion suchas minimum time of flight, with provisionfor continually correcting the path basedon periodic observations of measuredreferences such as line of sight to navi-gational stars and planets. The best-known filter for this purpose is called theKalman-Bucy Filter. The design of sucha filter required advanced mathematicsthat had never been covered in my col-lege courses. Some of the brightyounger engineers had studied this sub-ject in college and were able to followthe development. Because of the impor-tance of this subject, some prominentmathematicians were hired to givecourses to the personnel of the LangleyDynamics and Control groups, includingKalman and Bucy themselves, as wellas other engineers prominent in this

Turbulence Problems


field. Because of my lack of back-ground, I was never able to get anythingbut a general understanding of suchsubjects. In later years, companieswere formed that specialized in theseanalyses and apparently made a goodliving by assisting the aerospace com-panies in work that required this type ofknowledge.

Most airplane control system applica-tions do not require the complication ofthe Kalman-Bucy filter. In these cases,the Complementary Filter provides areadily understandable device that cansolve many of the practical problemsencountered. I was first introduced tothis filter by John F. Garren, Jr., an engi-neer in the Helicopter Branch. He foundthat the conventional helicopter controlsystem that incorporated angle-of-attack sensors to improve the longitudi-nal stability and rate sensors to improvethe damping of oscillations caused anundesirable amplification of the one-per-revolution vibration of the rotors. TheComplementary Filter provided an ana-log model of the helicopter response.Then the output of a simplified responsemodel was passed through a high-passfilter, the output of which was added tothat of a low-pass filter on the rate gyro.In this way the high-frequency vibrationssensed by the rate gyro were elimi-nated, whereas other high-frequencymotions as well as the low-frequencyresponse of the helicopter were allowedto operate the controls. The deviceproved quite effective in reducing theone-per-revolution vibration of thehelicopter. This study is reported inreference 7.7.

I later made a brief study of the use of aComplementary Filter in reducing thestructural vibrations of a rocket duringlaunch, often called the pogo effect.

Turbulence Problems

In my previous book, Journey in Aero-nautical Research, I devoted a largeamount of space to my studies ofresponse of airplanes to gusts and thedesign of systems to produce asmoother ride in turbulence. These sys-tems have been called “gust alleviationsystems,” although it was the responseto gusts rather than the gusts them-selves that was alleviated. In most ofthis work, the gust disturbances wereassumed to be one-dimensional; that is,the gust velocities varied along the flightpath but were assumed constant acrossthe wing span at any instant. This typeof analysis was useful because most ofthe response of airplanes comes fromlong wavelength gusts that do not usu-ally have much variation across thespan.

After finishing this work, which includedflight tests of a gust-alleviation systemon a Beech Model 18 (Navy C-45)transport, I went to other subjects, but Ialways kept in mind some related prob-lems that had never been solved. Oneof these problems was the effect of two-dimensional turbulence, that is, gustswith variations across the span as wellas along the flight path. To make suchstudies mathematically tractable, theassumption of isotropic turbulence isusually made. This type of turbulenceshows the same statistical propertiesalong any flight path that penetrates theturbulent region. Experimental studieshave shown that turbulence of this typefrequently occurs in the atmosphere.

A Langley engineer in the StructuralDynamics Branch, Franklin W.Diederich, made some valuable turbu-lence studies in which he calculated thestatistical properties of the wingresponse of various planforms to flightthrough isotropic turbulence. (The turbu-lence might be called axisymmetric



because the airplane was assumed tobe in straight horizontal flight.) Althoughthe mathematics involved was complex,I considered Diederich’s approach eas-ier to use than studies made previouslyby other investigators.

I knew from previous studies that whenan airplane hit an abrupt gust, theresponse was not instantaneous. Therewas a lag in response called theunsteady lift effect, which had beeninvestigated by several noted aerody-namicists. If a high-aspect ratio winghits a step-shaped gust, it travels five orsix chord lengths before the lift builds upto a steady value. As the aspect ratio isdecreased, the lag is reduced until at anaspect ratio of 3, the response is practi-cally instantaneous. In the gust-allevia-tion systems that have been tried, thegusts are usually measured by a vanelocated ahead of the nose of theairplane. If an airplane flies through aturbulent region with high-frequency dis-turbances, the vane will respond accu-rately to the disturbance, but because ofthe unsteady lift effect, the averageresponse of the wing lift over a period oftime will be reduced. A plot of the ampli-tude of the lift as a function of gust fre-quency will therefore show a decreaseas a function of frequency.

Other investigators studying two-dimen-sional turbulence had concluded thatthe response to high- frequency distur-bances may be reduced because ofvariations of gust velocity across thewing span. In this case, the vane mea-sures the disturbances along the center-line of the airplane, but the gust velocityat other points along the wing will be dif-ferent, and if the amplitude of responseis averaged across the span, the valuewill be less than the amplitude at thecenterline. This problem was called the“spanwise averaging effect.” These phe-nomena had been studied by differentgroups of engineers and no effort hadbeen made to compare the relative val-

ues of the decrease in the amplitude ofresponse due to unsteady lift with thedecrease due to spanwise averaging. Infact, in most studies, one or the otherof these two effects was investigatedand no account was taken of the otherone. I therefore undertook a study ofthis problem.

Without my knowledge, an engineer inthe Structural Dynamics Branch namedKermit G. (Cary) Pratt had been study-ing this same problem and published areport before mine was completed(ref. 7.8). It turned out that both analy-ses reached similar conclusions, but myanalysis was more complete from thestandpoint of logic and mathematics. Itherefore completed my report, whichwas published as a NASA TechnicalMemorandum (ref. 7.9).

In this presentation, no attempt is madeto give the mathematical details. Sev-eral interesting points were encounteredin the analysis. The decrease in ampli-tude of response with increasinggust frequency resulting from the span-wise averaging effect obtained fromDiederich’s report was given by a com-plex mathematical expression involvingBessel functions. A complete plot of thiscurve showed that a slight decrease inamplitude of lift occurred as the fre-quency approached zero at very lowfrequencies. This slight change in liftwas unexpected. I believe it may becaused by the rare occurrence of anoccasional isolated gust of large ampli-tude, which when averaged over thelong period of time between such gusts,shows up as a decrease in amplitude ofvery low frequency.

Another interesting discovery I madewas that Diederich’s complex expres-sion could be approximated very accu-rately by a very simple expressioncontaining just two terms and no Besselfunctions, which applied at reduced fre-quencies greater than about 4.

Turbulence Problems


The expressions for the reduction inamplitude of the response caused bythe unsteady lift effect was calculatedfor aspect ratios of 3 and 6 by methodsdescribed in reference 7.10. A com-parison of these results with thosedue to the spanwise averaging effectare given in figure 7.6. In this figure,

ω = frequency, rad/s, b = wing span, ft,and U = airspeed, ft/s. These curves areremarkably similar considering that theycame from completely different theories.In practice, the values of the two effectsshould be multiplied together to get thetotal attenuation.

FIGURE 7.6. Comparison of attenuation of lift due to spanwise aver-aging on an elliptical wing with that due to unsteady lift effects for wings of aspect ratios 3 and 6.

Spanwise averagingUnsteady lift, R = 6Unsteady lift, R = 3

1.0

.8

.6

.4

.2

0 4 8 12

Reduced frequency, k =

Attenuation of lift

16 20 24ωbU


CHAPTER 8 Some Nonaerodynamic Studies

Some Studies of Fireplaces

At one time, I was approached byWilliam J. Michael, then the Chief Scien-tist of the Langley Research Center, totry to find some work for the SpaceRadiation Effects Laboratory (SREL).This laboratory was under the combineddirection of Langley and the College ofWilliam and Mary and had been usedfor studying radiation effects on materi-als for use in space. The laboratory wasalso associated with the Virginia Associ-ated Research Center (VARC), a head-quarters building for SREL. The SRELbuilding contained a cyclotron to pro-duce the radiation. These laboratorieswere located in the northern end ofNewport News in an area separate fromthe main NASA center and now calledOyster Point. Unfortunately, funding forthe activities at SREL had been cut offwith the decline of the space program. Aproposal had been made to develop theSREL into a nuclear research center.This program was in the early develop-ment stages and required approval by aconsortium of Southern universities(SURA), as well as approval by Con-gress and other groups. The problemwas to keep VARC alive with somesmall contracts until the large nuclear

research program could be put intooperation.

I knew that VARC was a very usefulfacility with a fine technical library andseveral permanent employees, includ-ing mechanics. I found that the Depart-ment of Energy had a fund to providesmall grants to individuals or researchgroups to improve the efficiency of fire-places. The interest in fireplaces devel-oped because a serious oil shortageexisted in this period, and many peoplewere trying to burn firewood to heattheir houses. Although I knew very littleabout heat transfer or thermodynamics,I submitted a proposal to the Depart-ment of Energy to study the efficiency offireplaces.

The funding was sufficient to giveemployment to three summer students,two men and one woman, who werevery capable and benefited from theexposure to a technical problem. I wentto work at VARC two afternoons a weekto supervise the work.

I studied the works of Count Rumfordand Benjamin Franklin, both of whomwere noted for improving the fireplacesof their day. Count Rumford’s main con-tribution was to change the shape of thefireplace to a more narrow, taperedshape to direct more heat radiation into


Some Nonaerodynamic Studies

the room. I concluded that BenjaminFranklin knew more about the problemsof fireplaces than most people livingtoday. He developed a fireplace in whichthe firebox was airtight. Air from theroom was heated in a closed containerin the firebox and returned to the room,while air from the outside, coming underthe house, burned the logs and went upthe chimney. All of the circulation wasdone by convection as there were noelectrically driven blowers in those days.The fireplace greatly improved effi-ciency because cold air from the outsidewas not sucked into the room. Franklin,always a genius, ran the Post Office inPhiladelphia and installed one of hisfireplaces there, while selling piles ofiron castings to the customers to buildtheir own fireplaces. These fireplaceswere short-lived, however, becausesealing of the joints between the cast-ings, done with a mixture of mud andstraw, soon developed leaks andallowed smoke to escape into the room.Franklin’s fireplaces using welded steeljoints would be excellent for use todayand would not make the distractingnoise of the blowers used on presentday fireplace inserts.

Franklin’s knowledge of fireplaces soonbecame well-known, and many manu-facturers called their fireplaces Franklinstoves, though they did not incorporatethe advantages of Franklin’s invention.

When I started work on fireplaces, somemanufacturers of fireplace inserts hadalready produced inserts with glass win-dows that allowed a view of the fire.These glass windows quickly becamedirty with soot and pitch, which was verydifficult to clean off. I conceived the ideaof using a fine-mesh stainless steelscreen for the window, which wouldallow some view of the fire, but whichwould be kept clean by the small flow ofair through the screen. I found that fine-mesh stainless screen was produced inlarge quantities for the paper industry,

and as a result was not as expensive asI had imagined. One project that thegroup undertook was to build a smallwind tunnel, with a test section about4 by 8 inches, in which the pressuredrop through screens could be mea-sured. Although screens were used insome NASA wind tunnels to smooth theairflow, no tests of the pressure dropthrough screens as fine-meshed asdesired had ever been made.

Another project was to write a computerprogram for predicting the efficiency offireplaces. This program accounted forthe heat transfer to the air in the chim-ney and to the room, but accounting forthe radiation between all the heatedparts was probably not done very accu-rately. The young lady who worked withthe group proved very skillful at com-puter programming.

Finally, two commercially available fire-place inserts were obtained to measuretheir efficiency. Special instrumentswere made to measure the airflow upthe chimney and the pressures andtemperatures in various parts of the sys-tem. One of the inserts, which used aset of stainless steel tubes behind thefire to take in air at floor level and blow itinto the room above the fire proved tobe about 60 percent efficient. The otherinsert, which had a metal shell com-pletely surrounding the fireplace to heatair from the room, was found to beabout 80 percent efficient.

I soon found that numerous manufactur-ers wanted to improve fireplaces at thesame time that I did, and many of themcame out with efficient units. The airabove New Hampshire became so pol-luted by wood smoke that fireplaceinserts there were required to have acatalytic converter, similar to those usedin automobile exhausts, to remove thecarbon monoxide and nitric oxide fromthe combustion products.

Control System for a Small Hydrofoil Boat


I wrote a report on the studies that wassubmitted to the Department of Energy.Interest in the project soon declinedwhen the fuel shortage ended, andthe report was never published. TheVARC and SREL stayed alive, however,and subsequently developed into theJefferson Laboratory, one of the maincenters in the country for nuclearresearch.

Control System for a Small Hydrofoil Boat

My first boss, Dr. Robert R. Gilruth, whowas head of the Stability and ControlBranch of the Flight Research Divisionwhen I came on duty, had shown aninterest in hydrofoil boats when he firstcame to Langley. He had already built asmall hydrofoil sailboat and later built asmall outboard motor-propelled run-about and a larger sailboat, both liftedfrom the water on hydrofoils. Theseboats used surface-piercing foils, onwhich the tips of the foils project abovethe water when cruising. This featureprovides lateral stability, much like thedihedral on an airplane wing. Longitudi-nal stability was provided by a sub-merged rear foil in conjunction with thecorrect center of gravity. Later Gilruthbecame a consultant to the GrummanCompany on the design of some largehydrofoil boats for use by the Navy inWWII. All this work was done as ahobby, outside of working hours. Natu-rally, I also became interested in boatsof unusual design, although I did notbuild any full-scale hydrofoil boats. I didbuild a small rubber-motor propelledhydrofoil boat with submerged foils,which was stabilized laterally by a gyro-scope linked to the front foil. The gyrowheel was spun up by pulling a stringwound around the shaft. This boat con-clusively proved to be stable in its shortruns.

In later years, hydrofoil passenger boatswere made in several European coun-tries. The Boeing Company made onein this country. These boats, whichused submerged foils for greater effi-ciency, required complex gyroscopicsystems and electronic autopilots toprovide stability.

I had a conventional 15-foot outboardboat that I sailed in Hampton Roads. Irealized that a hydrofoil boat of thesame size and power would sail muchfaster and more smoothly. I did not,however, have the time or facilities tobuild a full-scale boat. I did considerhow a simple mechanical autopilotmight be built to stabilize a hydrofoilboat of this size. I made an analysis ofsuch a system at work with John D.Shaughnessy. The analysis was doneby using a high-speed NASA digitalcomputer. I felt that this work wasappropriate in my position because Iwas head of the Stability and ControlDivision and because the AIAA atthat time published a Journal ofHydronautics along with its other techni-cal journals. A report on this work waslater published in the AIAA Journal ofHydronautics (ref. 8.1).

The primary sensor for the control sys-tem was visualized as a long, stream-lined stick pivoted at the hull to allow itto swing fore and aft. The drag force ofthe water on the stick would increasewith the depth of submersion. This stickwas linked to a trailing-edge flap on thefront foil to produce more downward flapdeflection when the drag on the stickincreased. Some damping of the slickmotion results from the variation of dragon the stick with stick motion fore andaft, and from the variation of flap hingemoment with rate of change of flapdeflection. It was not known whetherthese sources of drag would be suffi-cient to damp oscillations of the stick. Tofurther increase the damping, a bob-weight was linked to the stick so that as


Some Nonaerodynamic Studies

the stick moved rearward, the bob-weight would move up. The bobweight isa weight mounted on a pivoted arm andrestrained by a spring. The bobweightsenses vertical acceleration at the loca-tion of its pivot. Locating the bobweightnear the front of the boat would presum-ably produce some lead in the accelera-tion of the weight, causing the flap to bedeflected up by an upward pitching orby an upward vertical motion of the hull.

The equations of motion of the systemsare too complex to present in this reportbut may be found in reference 8.1. Thestability of the system was studied byroot locus plots, by transient responsesto disturbances, and by frequency-response plots of the various variablesin the system. The root locus studiesshow that adequate stability of allmodes of motion may be obtained by asystem of the type analyzed. The pre-dominant low-frequency mode of the

boat, however, appears to have a fre-quency too low to interact with the bob-weight system on the flap. The originalpremise that the bobweight would con-tribute to the damping of the low-frequency modes of the boat, therefore,was found to be incorrect.

The system studied with a relatively lowvalue of the variation of restoring forcewith vertical displacement providesexcellent attenuation of the verticalmotions of the boat due to head wavesthrough a large range of frequencies,whereas in stern waves the motion isattenuated to a value less than the waveamplitude at frequencies above 0.8 Hzbut amplified at frequencies below thisvalue. Stabilization of the boat in sternwaves of low frequencies would proba-bly require a more sophisticated controlsystem involving an attitude gyro. Inrestricted bodies of water, long wave-lengths are probably rare.


CHAPTER 9 Concluding Remarks

This volume has contained most of thework that I conducted at Langley duringthe space program. After this workdeclined, I worked on many differentprojects, mainly to provide interestingsubjects for my employees and to solvesome problems that arose in workbefore the space program. I have pre-sented four chapters that contain twelveexamples of projects or research con-ducted during this period. These exam-ples represent only a small fraction ofthe number of different studies that Imade. A bibliography of my reports thatwere published during this period is pre-sented at the end of this volume.

In 1979, after 39 years of service, Ifound that much of my time as Chief ofthe Stability and Control Division wastaken up by administrative matters. Mybackground and much of my earlierwork had been devoted to research.Most of the other division chiefs at Lan-gley had for some time been administra-tors, leaving the conduct of theirresearch to the personnel assigned totheir divisions. I found that I did not haveenough time to do personal researchand to adequately perform the adminis-trative duties. I discussed this problemwith Oran Nicks, an assistant Director ofthe Langley Research Center. Headvised me to resign and accept the

position of Distinguished ResearchAssociate (DRA). In this position, some-what like that of a Professor Emeritus ina college, I would be free to use thefacilities at Langley and to do researchas I desired. My retirement annuitywould be about two-thirds of my maxi-mum salary as a federal employee. Iconsidered this offer as a good opportu-nity, and started what turned out to be along career as a DRA.

One objective of my career as a DRAwas to conduct a wind-tunnel test. Manyof the engineers at Langley wereengaged in operating the numerouswind tunnels at this center. I had neverhad this opportunity because when Icame on duty, I was assigned to theFlight Research Division, in which theduty of a flight engineer was to analyzerecorded data obtained by the test pilotsin flights of full-scale airplanes. Theclosest I came to a wind-tunnel test wasshortly after I came on duty, when I wasassigned to bundle up in my overcoatand climb into the cockpit of a FairchildF-22 airplane mounted in the Full-ScaleTunnel. The airplane had beenequipped with a bob-weight in the con-trol system. My duty was to applyimpulses to the control stick to measurethe damping of the elevator motion asaffected by the bob-weight.


Concluding Remarks

After I started work as a DRA, the Shut-tle had started to fly, and it was soonfound that the very high landing speedresulted in damage to tires and over-heating of brakes. I believed that thelanding speed could be lowered byinstalling a canard surface near thenose of the Shuttle that would help to liftthe nose so that the elevons at the rearof the delta wing could be trimmed downand further increase the lift prior totouchdown. I did not contemplate thatthe test would be difficult, but I soonfound that all the major wind tunnelshad schedules that were booked up forat least three years. I was able to getsome test time in the old 12-Foot Tun-nel, a wind tunnel that was built in thethirties and had originally been used tocreate an airstream in which freely flyingsmall models would be tested. Later, thetunnel was equipped with balances andused in the conventional manner. Themodel that I acquired from Rockwellwas an old balsa-wood Shuttle fluttermodel, about 5 feet long that had beenpartially crushed. I rebuilt the model andran tests on a number of canarddesigns. I found that the flow in the tun-nel was variable along its length so thatthe results were not quantitative butcould be used for comparative pur-poses. The tests took over three yearsto make and analyze. As a result, Ibecame more impressed by the workrequired in wind-tunnel testing. On com-pletion of the study, the results werepresented in a talk at the Johnson FlightResearch Center. The canard surfaceswere never used on the Shuttle becauseof the difficulty in changing the design ofthe Shuttle, and because some reduc-tion in landing speed could be obtainedby rearward positioning of the center ofgravity combined with some downwarddeflection of the elevons. This arrange-ment reduced the longitudinal stability,but the stability could be restored byadjusting the electronic control system.

A second interest that I had afterbecoming a DRA was to learn moreabout airfoils. I had studied airfoil theoryat MIT some 40 years earlier, but manyadvances had been made in this fieldwith which I was not familiar. A notabletheory had been developed by The-odore Theodorsen at Langley to calcu-late the pressure distribution on anairfoil given its contour. Later, Lighthill inEngland, Trockenbrot in Germany, andothers had solved the inverse problemof determining the contour required toproduce a given pressure distribution.Robert T. Jones and Eastman Jacobs atLangley also had developed an iterativetechnique to solve this problem, but asfar as I know it was never published.

Other groups in this country and atLangley developed boundary-layer the-ory that allowed the development of theboundary layer and the resulting frictiondrag to be calculated. These investiga-tors at Langley were in a different groupfrom the airfoil theorists. I had sug-gested to Oran Nicks that the theoriesbe combined to allow the boundary-layer distribution on an airfoil to be cal-culated given the airfoil shape. Beforethis analysis was done at Langley, how-ever, Richard Eppler in Germany haddeveloped a computer program for atheory that combined these parts of theproblem. Dan M. Somers, an engineerat the Low-Turbulence Tunnel at Langleyarranged for Eppler to come to Langleyand explain his program to him and tothe people in his group. (refs. 9.1and 9.2). I consulted with Somers tolearn the details of this program, and Imade many runs to study airfoils of dif-ferent types. Later, I modified the pro-gram to allow the calculation of bothfriction drag and pressure drag on anairfoil as a function of angle of attack.This work was presented in a Society ofAutomotive Engineers (SAE) meeting atAnaheim, California in October, 1988. 1believe that experts at the large airplane


companies had developed similar pro-grams, but the results had not beenpublished previously.

In most wind-tunnel tests to determinedrag of airfoils, the total drag (pressuredrag plus profile or skin-friction, drag) ismeasured by means of a rake survey.Until recent years, very few attemptswere made to separate the two sourcesof drag or to determine the distributionof these components of drag over thesurface of the airfoil. The only study ofthis type with which I was familiar was aremarkable analytical study by H. B.Squire and A. D. Young made in 1937 inEngland (ref. 9.3). Without modern com-puting facilities, calculations of this typewere extremely tedious. My papershowed these characteristics for threeairfoils, each at three values of Rey-nolds number. The effects of uniformsuction through the surface of the airfoilwas also studied, and the possibility ofnegative pressure drag, or thrust, on theairfoil was demonstrated.

At that time, there was interest in veryhigh-altitude, unmanned aircraft tomake possible long endurance for sur-veillance or for communication pur-poses. I made recommendations forairplane airfoils of this type. The person-nel at the companies doing this workcontacted me because they had heardof my experience with model airplanes.They learned of my theoretical studieslater.

I was also interested in learning aboutthe theory of propellers. At this time,most airplanes of interest were jet-pro-pelled, and very little work on propellerswas being done at Langley. Earlier, inthe thirties and forties, there was muchinterest in this work, and Theodorsenhad published a propeller theory thatwas based on accurate aerodynamictheory. Even as early as 1919, Betz andPrandtl in Germany had developed atheory based on some good approxima-

tions to the flow characteristics, andFred Weick at Langley in the thirtiesbuilt the Propeller Research Tunneland made empirical studies of full-scalepropellers. By this time, propellersgood enough for all practical purposescould be designed. Incorporating thefine points of aerodynamic theory didnot change the efficiency more than 2 or3 percent.

A paper summarizing practically all thatwas known about propeller design waswritten by H. Glauert in England andpublished in Volume IV of the DurandSeries in 1934. By the seventies, fewpeople paid any attention to this work,but E. E. Larrabee at MIT studied thearticle, made a few important correc-tions, and programmed the theory on apocket calculator. Later, I used this the-ory as an example of programming onthe HP 9820 and later on the HP 9830computers. Designers of commercialairplanes at that time could purchasepropellers from companies that special-ized in this work, but homebuilders oftenwished to design and build their ownpropellers. I received over 50 requestsfrom all around the world for copies ofmy programs.

The preceding paragraphs are exam-ples of my work as a DRA. Many otherstudies, probably of lesser importance,were also conducted, some of whichare mentioned in the bibliography. Inaddition I started writing a history of mywork at Langley. This book, entitledJourney in Aeronautical Research waspublished in November 1998, as theNASA publication Monographs in Aero-space History, Number 12 (ref. 1.1).This document covered the work to thestart of the space program in 1958. Thepresent volume contains my work dur-ing the space program and later work to2004. This work is necessarily abbrevi-ated because of the large number ofsubjects that were encountered in thework after the space program.


Concluding Remarks

The long duration of my work at theLangley Research Center has involvedmany changes. Perhaps the greatestchange, from the engineering stand-point, is the development and wide-spread use of computers. Thesemarvelous devices allow analyses to bemade in seconds that previouslyrequired days or weeks. Most of the

older techniques are now obsoletebecause they are incorporated in com-puter programs. Younger engineers whogrew up with these methods produceresults with computers with a facility thatis beyond my abilities. The changingresearch emphasis, as well as the newbody of personnel involved, makes thisa good time to end this volume.


Appendix

Abbreviated Chronology of Space Flight Accomplishments

List of major accomplishments and total launch attempts in eachyear through 1962, showing number of failures.

Major milestones in U.S. manned space programthrough first lunar landing and return.

Year Date Name Notes

1957

10/04/57 Sputnik I First artificial satellite (Russian)

11/03/57 Sputnik II Second artificial satellite (Russian), carried dog, Laika

1958

2/01/58 Explorer I, Jupiter C

First successful U.S. artificial satellite, discovered Van Allen belts

3/26/58 Explorer III, Jupiter C

Radiation, micrometeoroid data

22 total launch attempts, Russian and U.S.; 7 achieved orbit

1959

U.S. made launch attempts of several artificial satellites, including Vanguard, Discoverer and Transit. Russia launched Luna II

25 total launch attempts, 13 achieved orbit

1960

U.S. launched Atlas Able (Pioneer 1960)


1961

4/12/61 Vostok 1 Orbit, Yagi Gogarin (Russian, first man in space)

5/05/61 Mercury, Redstone 3

Suborbital Flight, Alan Shepard (first U.S. man in space)


Appendix

9/13/61 Mercury, Atlas D

One orbit, unmanned

10/29/61 Mercury, Atlas 5

Two orbits, Chimpanzee Enos


1962


Three orbits, John Glenn


Three orbits, Scott Carpenter


Three orbits, Wally Schirra


1963, 1964: at this point, total launch attempts are omitted, skip to major U.S. space pro-grams in 1965

1965

7/28/65 Ranger 7 First successful close-up photos of Moon

12/04/65 Gemini 7 Borman and Lovell

12/15/65 Gemini 6 Schirra, Stafford; rendezvous with Gemini 7

1966

3/16/66 Gemini 8 target Atlas Agena D

3/16/66 Gemini 8 Armstrong, Scott docked with Gemini 8 target; stuck thruster caused emergency.

7/05/66 Apollo 2, Saturn 1B

AS-203, unmanned test

7/18/66 Gemini 10, Titan II

Young, Collins; first EVA, docked with Gemini 10 target; raised its apogee to 755 km

8/10/66 Lunar Orbiter I Atlas Agena D

Returned lunar photos, crashed on Moon

11/08/66 Gemini 12, Titan II

Lovell, Aldrin docked with target, successful EVA tests

11/08/66 Atlas Agena D Target

1967

2/05/67 Lunar Orbiter 3 Atlas Agena D

Returned lunar photos, crashed on Moon

1968

4/04/68 Apollo 6, Saturn V

Command module test

10/11/68 Apollo 7, Saturn 1B

Schirra, Cunningham, Eisele, Earth orbital test of Apollo

12/21/68 Apollo 8, Saturn V

Borman, Lovell, Anders; first circumlunar mission


1969

3/3/69 Apollo 8, LM Saturn V

McDivitt, Scott, Schweickert; Earth orbital test.

5/18/69 Apollo 10, LM, Saturn V

Stafford, Young, Cerman; LM undocking and docking in Lunar orbit

7/16/69 Apollo 11, LM, Saturn V

Armstrong, Aldrin, Collins; first lunar landing and return


References

1.1. Phillips, William H.: Journey in Aeronautical Research, A Career at NASA Langley Research Center—Monographs in Aerospace History Number 12, 1998.

1.2. Hansen, James R.: Spaceflight Revolution, NASA Langley Research Center From Sputnik to Apollo. NASASP 4308, 1995.

2.1. Anon: Space Log, Volume 32. Compiled by Editor, TRW Space Log, TRW Space and Electronics Group,Redondo Beach, California 90278.

2.2. Swenson, Lloyd S., Jr.; Grimwood, James M.; and Alexander, Charles C.: This New Ocean, a History ofProject Mercury. NASA SP 4201, 1966.

2.3. Grimwood, James M.; Hacker, Barton C.; with Vorzimmer, Peter J.: Project Gemini, a Chronology. NASASP 4002, 1969.

2.4. Brooks, Courtney G.; Grimwood, James M.; and Swenson, Loyd S.: Chariots for Apollo. NASA SP 4205,1979.

2.5. Anon: Notes on Space Technology. Compiled by the Flight Research Division, Langley Aeronautical Labo-ratory, Langley Field, Virginia, Feb.–May 1958.

2.6. Vyssotsky, Alexander N.: Exploration of Space From the Observatory of the University of Virginia. U. ofVirginia News Letter, Vol. XXXV, No.8, Jan. 15, 1959.

2.7. Allen, H. Julian; and Eggers, A. J., Jr.: A Study of the Motion and Aerodynamic Heating of Missiles Enteringthe Earth’s Atmosphere at High Supersonic Speeds. NACA TN 4047, 1957.

2.8. Staff of the Langley Flight Research Division, Compiled by Donald C. Cheatham: A Concept of a MannedSatellite Reentry Which Is Completed With a Glide Landing. NASA TM X-226, 1959.

2.9. Spencer, Bernard, Jr.: An Investigation at Subsonic Speeds of the Longitudinal Aerodynamic Characteris-tics at Angles of Attack From –4° to 100° of Delta-Wing Reentry Configurations Having Vertically Displacedand Cambered Wing-Tip Panels. NASA TM X-440, 1961.

2.10. Olstad, Walter B.: Static Longitudinal Aerodynamic Characteristics at Transonic Speeds and Angles ofAttack up to 99° of a Reentry Glider Having Folding Wing-Tip Panels. NASA TM X-610, 1961.

2.11. Hondros, James G.; and Goldberg, Theodore J.: Aerodynamic Characteristics of a Group of Winged Reen-try Vehicles at Mach Number 6.01 at Angles of Attack From 60° to 120° and –10° to 30° Roll at 90° Angleof Attack. NASA TM X-511, 1961.


References

2.12. Popma, Dan C.: Spacecraft Life Support Systems. Presented at the Institute of Electrical and ElectronicEngineers. Region 3 meeting, Richmond, Virginia, April 24–26, 1963. Paper number CP 63-693.

2.13. Von Doenhoff, Albert E.; and Hallissy, Joseph M., Jr.: Systems Using Solar Energy for Auxiliary SpaceVehicle Power. I.A.S. Rept. 59-40, Jan. 1959.

3.1. Braün, Gerhard W.: An Analytic Treatment of the Unsteady Spinning of Airplanes of High Relative Densi-ties. WADC TN WCLS 53-79, Nov. 1953.

3.2. Goldstein, Herbert: Classical Mechanics, Second Edition. Addison-Wesley Publ. Co., Reading, Massachu-setts, 1980.

3.3. Chobotov, Vladimir A.: Spacecraft Attitude Dynamics and Control. Krieger Publ. Co., Malabar, Florida,1991.

3.4. Hooper, John W.: The Damping of Metallic Cylindrical and Spherical Bodies Rotating in a Uniform Mag-netic Field. ABMA Tech Report DG-R-15, Nov. 1, 1957.

4.1. Clohessy, W. H.; and Wiltshire, R. S.: Terminal Guidance System for Satellite Rendezvous. Preprint Mo.59-93, S.M.F. Fund Paper, I.A.S., June 1959.

4.2. Battin, Richard H.: Astronomical Guidance. McGraw-Hill Book Co. Inc., New York, 1964.

5.1. O’Bryan, Thomas C.; and Hewes, Donald E.: Operational Features of the Langley Lunar Landing ResearchFacility. NASA TM D-3828, 1957.

5.2. Phillips, William H.; Queijo, M. J.; and Adams, James J.: Langley Research Center Facilities for MannedSpace Missions. ASME paper No. 63-AHGT-91, Jan. 1, 1964.

6.1. Kraft, C. C.: Flight—My Life in Mission Control. DUTTON, Penguin Putnam, Inc., New York, Mar. 2001.

7.1. Burgin, George H.; Fogel, Lawrence J.; and Phelps, J. Price: An Adaptive Maneuvering Logic ComputerProgram for the Simulation of One-on-One Air-to-Air Combat, Vol. 1, General Description.

7.2. Hankins, Walter W., III: First DMS Engagements Between Human Pilots and a Computer-Driven Opponent.Langley Working Paper LWP 1127, Aug. 13, 1973.

7.3. Hankins, Walter W., III: Interactive Computerized Air Combat Opponent. Presented at the AGARD FMP/GCP Joint Symposium on Flight Simulation/Guidance Systems Simulation. The Hague, Netherlands, Oct.20–24, 1975.

7.4. Phillips, William H.: Propulsive Effects Due to Flight Through Turbulence. J. Aircr., vol. 12, no. 7, July 1975,pp. 624–626.

7.5. Phillips, William H.: Analysis of Effect of Asymmetric Loading on Sailplane Performance in Circling Flight.Presented at the Symposium on the Science and Technology of Motorless Flight, Cambridge, Massachu-setts, Oct. 18–22, 1972.

7.6. Phillips, William H.: Gyroscopic Moments on a Glider in Turning Flight. “Technical Soaring,” Nov. 21, 1972.

7.7. Garren, John F., Jr.; Niessen, Frank R.; Abbott, Terence S.; and Yenni, Kenneth R.: Application of a Modi-fied Complementary Filtering Technique for Increased Aircraft Control System Frequency Bandwidth inHigh Vibration Environment. NASA TM X-74004, 1977.

7.8. Pratt, Kermit G.: Effect of Spanwise Variation of Turbulence on the Normal Acceleration of Airplanes WithSmall Span Relative to Turbulence Scale. NASA TM X-72748, 1975.

7.9. Phillips, William H.; and Lichtenstein, J. H.: Comparison of Effects of Unsteady Lift and Spanwise Averag-ing in Flight Through Turbulence. NASA TM X-74028, 1977.

References


7.10. Jones, R. T.: The Unsteady Lift of a Wing of Finite Aspect Ratio. NACA Report No. 681, 1940.

8.1. Phillips, William H.: and Shaughnessy, John D.: Investigation of Longitudinal Control System for a SmallHydrofoil Boat. J. of Hydronautics, vol. 10, no. 2, April 1976, pp. 39–42.

9.1. Eppler, R.; and Somers, D. M.: A Computer Program for the Design and Analysis of Low Speed Airfoils.NASA TM-80210, 1980.

9.2. Eppler, R.; and Somers, N. M.: Supplement to a Computer Program for the Design and Analysis of LowSpeed Airfoils. NASA TM-81862, 1980.

9.3. Squire, H. B.; and Young, A.D.: The Calculation of the Profile Drag of Airfoils. RAE R.&M. No. 1838, 18 Nov.1937.


Index

A

Adamson, David 6Algranti, Joe 41, 43Allen, H. Julian 7, 87Armstrong, Neil 41, 43, 47, 84, 85

B

Battin, H. Richard 32, 88beryllium 9Bicknell, Joseph 64Bird, John D. 48, 49, 61, 63Braün, G. W. 18, 88Burgin, George H. 63, 88

C

Calspan 56Cheatham, Donald C. 6, 11, 56, 87Chilton, Robert G. 53Clohessy, W. H. 31, 88closed-form solution 19, 31Collins, Robert 33

D

Diederich, Franklin W. 71, 72Donlan, Charles J. 14, 47Draper Lab 32, 53

E

Eggers, A. J., Jr. 7, 87Eggleston, John M. 33ellipsoid of inertia 17, 19Eppler, Richard 80, 89Euler’s dynamical equations 18, 19

F

Faget, Maxime A. 7, 52Fogel, Lawrence J. 63, 88

G

Garren, John F. 71, 88Gilruth, Robert R. 6, 26, 55, 77Goldstein, Herbert 18, 19, 22, 88

H

Halissy, Joseph M. 15Hankins III, Walter W. 63, 88Hewes, Donald E. 41, 45, 49, 50, 88Hill, Paul 14

J

Jackson and Moreland 40Jered Industries 40Johnson, Harold J. 49


Index

K

Kalman-Bucy Filter 70, 71Katzmayr, R. 66Kennedy, John F. (President) 15, 29, 36Kraft, Christopher C. (Dr.) 51, 54, 88

L

Loftin, Lawrence K. 36, 38, 65Long Duration Exposure Facility (LDEF)15Lunar Landing Research Facility (LLRF)39, 40, 41, 43, 45Lunar Landing Training Vehicle (LLTV)38, 41, 45Lunar Module (LM) 32, 40, 41, 45, 47,85

M

Mathews, Charles W. 6Mattson, Axel 51Mayer, J. P. 6, 33Mercury Project 6, 7, 13, 87MISS program (Man in Space Soonest)6Moment of inertia 25moment of inertia 17, 20, 23

N

NACA (National Advisory Committee forAeronautics) 1, 2, 4, 6, 14, 15, 38, 60,61, 87, 3NASA (National Aeronautics and SpaceAdministration) 3, 4, 6, 7, 11, 13, 14, 20,38, 39, 41, 42, 45, 49, 63, 72, 75, 76,77, 81, 87, 88, 89Newtonian flow 25Nickel-cadmium (NiCad) batteries 15Nicks, Oran 68, 79, 80

O

Ohlson, Roland 15

P

Pearson, Henry A. 4, 5, 22, 33Person, Lee 41Phelps, J. Price (Decision Science, Inc.)63, 88Pilotless Aircraft Research Division(PARD) 6, 7, 14, 26Poinsot, Louis 18, 19Popma, Dan C. 14, 88Pratt, Kermit G. (Cary) 72, 88

R

Reeder, John P. (Jack) 38, 39regenerative space power systems 15Rene 41 9retro-rocket 9, 26Russell, Walter J. 33

S

Shaughnessy, John D. 77, 89Shuttle Training Aircraft (STA) 56Somers, Dan M. 80, 89Space Radiation Effects Lab (SREL) 14,75, 77Spaceflight Revolution 1, 87Sputnik 1, 4, 83, 87

T

Tactical Effectiveness Simulator (TES)61Theodorsen, Theodore 80, 81TIFS (Total In-Flight Simulator) airplane56TRW Space Log 3, 87

V

Verne, Jules 1Vogeley, Arthur W. 47, 50Von Doenhoff, Albert E. 15, 88

W


W

Wiltshire, R. S. 31, 88

Z

Zalovcik, John A. 8

Zimmerman, Charles H. 46

Phillips in 2002 at the age of 84. From 1979 to 2005, he was aDistinguished Research Associate at the NASA Langley Research Center.


About the Author

William Hewitt Phillips was born in Port Sunlight, Merseyside, England, and came tothe United States with his parents at the age of 2. He was educated in the Belmont,Massachusetts public schools and studied aeronautical engineering at MIT where heobtained his S.B. degree in 1939 and his S.M. degree in 1940. His entire professionalcareer has been spent with the NACA, later NASA, at Langley Research Center inHampton, Virginia. Langley Research Center is the original government center foraeronautical research in the United States. On entering duty in July 1940, he wasassigned to the Flight Research Division. He specialized in the study of flying qualitiesand stability and control of airplanes. His duties included studies to improve the flyingqualities of many World War II military airplanes. After the war, he was involved inresearch on the development of jet-powered fighter airplanes, supersonic airplanes,stability augmentation and its effect on human pilot control, automatic control, gustalleviation, and aeroelastic effects. His previous book, Journey in AeronauticalResearch, ends with the advent of the nation’s space program. After the start of thespace program, he became Chief of the Space Mechanics Division and supervised80 to 90 people in the areas of space rendezvous, navigation, and lunar landing. As apart of his responsibility to the space program, this division developed simulators forthe Gemini and Apollo programs. He developed the Lunar Landing Facility that wasused for training astronauts in landing on the Moon. His work also included consulta-tion and analysis in the development of the Space Shuttle. Later work included super-vising studies of effects of turbulence and of application of control theory andcontributing to the development of the Differential Maneuvering Simulator, a facilityused for air combat studies. He retired from government service in February 1979 butcontinued until 2004 in the position of Distinguished Research Associate, duringwhich time he performed original research on solar-powered aircraft, propellers, airfoildesign, and wind-tunnel studies of canard surfaces use for the Space Shuttle. Heserved as a consultant on studies of flight dynamics and control. He has receivednumerous awards throughout his career, including the IAS Lawrence Sperry Award foraeronautics in 1944 and the President’s Award for Distinguished Federal Civilian Ser-vice in 1979. In 2005, at the age of 86, he continues to design and fly model airplanesand still has a keen interest in aeronautics.

Phillips married Viola Ohler in 1947 when she was head of the Editorial Office atLangley. They had three children, Frederick H., Robert O., and Alice B. Phillips. All are


About the Author

now married. Frederick, whose wife is Joanne, is a financial consultant. Robert andwife Cheryl have three children: Tyler, 25; Ross, 22; and Jocelyn, 20. Robert works atThe Volpe Center in Cambridge, Massachusetts. Alice and husband Thomas Checkhave three children: Candace, 18; Nolan, 16; and Aubree, 14. Alice formerly workedfor robotics firms and is now a homemaker.

www.nasa.gov

CHAPTER 5 Space Simulation Studies - NASA · a control problem considerably different from that of...

Documents

Transcript of CHAPTER 5 Space Simulation Studies - NASA · a control problem considerably different from that of...