‘THE DESIGN OF TAILLESS AIRPLANES ‘Thesis by ‘Frank Dore In Partial Fulfillment of the Requirements for the Degree of Aeronautical Engineer California Institute of Technology Pasadena, Californie 1se7ACKNOWLEDGHENTS ‘The author wishes to thank Mr, Henry Negamateu of the GALCIT staff for his encouragement and assistance in the preperation of this thesis. Special thanks 1s to be given to Dr. E, E, Sechler, Mr. W. E. Bowen, and members of the GALCIT staff for their assistance and advice in constructing the model and making the wind tunnel tests.‘TABLE OF CONTENTS Acknowledgenents Table of Contents Table of Figures Summary Introduction Nomenclature Problems of Tailless Airplanes Wing Design Control Zlenents Fuselage ond Fillets Recommendations for Future Work Design of Supersonic Tetlless Airplene Preliminary Wind Tunnel Results Bibliography Page 1 26 a7 34 37Figs Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. ‘TABLE OF FIGURES Aspect Ratio vs, Angle of Sweepback, Spanwise Lift Distribution, Wing Loads in X-¥ Plane. Supersonic Teilless Airplane. Drag Coefficient vs. Mach Number. Performance Estimates, Power = 4000 # St. Th Performance Zstimates, Power = 6000 # St. Th. Wind Tunnel Test Results. Tuft Studies.SUMMARY This paper reviews the various sources of information pertaining to the design of tailless airplanes. The particular topic of any report or menorandun is illustrated by its application to the prob+ lems encountered in the design of a given airplane. The subjects mentioned are the determination of the sweepback angle, aspect ratio, taper ratio, twist, dihedral, airfoil section, spanwise loading, and aerodynamic center of the wing; the control elements; and the fuse- lage end fillet for high speed sircraft. o illustrate the application of the source material the design of a tailless supersonic airplane is considered, The slow speed wind tunnel tests of the 1/12-scale model are included in the report,2 INTRODUCTION The problem of basic airplane design has received considerable attention in the lest several years. This problem concerns the basic design of the airplane configuration as a whole, the relation of wing, fuselage, and empennage to one another, their purpose, practicability, and necessity. The focusing of attention on this problem has highlighted the work of several men, such as Alexander Lippisch of Germeny and John Northrop of this country, who have felt that the standard airplane configuration was not necessarily the best, These men and others have considered basic designs dissin Alar to the conventional airplane, The development of the modern airplane has been an evolutionary process, The first designers-aviators built a multitude of types of airplanes, only a few of which were successful. They continued on the path of design sugzested by the successful airplane, but after a few attempts, discarded the unsuccessful designs. Of necessity only the simplest and most stable craft would fly, due to the lack of knowledge of nearly all aircraft theory end practice by the de~ signers, The fact that the most stable airplane had a straight wing, and a tail at some distance from the wing, led to the improvement of airplanes of this type and eventually the subsidence of all other types. Thus,the basic design of the modem airplane was evolved, its existence being due to the fact that it wes originally more stable and could thus be flown more easily than any of its competitors in design.3. In the ensuing years more persons were considering the problen of most efficient airplane performance and their conclusions were that @ higher performance flying machine could be achieved if the body and empennage could be discarded, leaving only the wing for Lift and stowage space end the propulsive unit for flight. At first the advocates of this type of design thought only in terns of reducing the parasite dreg of the airplane by eliminating the fuselage. This was the basis for their efforts until World War II and the appearance of the very high speed pursuit airplanes, Theory had indicated, and experience was proving, that the drag of conventional airplanes in- creased tremendously at speeds near the specd of sound, Also, at very high Mach numbers the airplane was apt to becone unstable, Faced with the fact that the standard design was no longer clearly superior, a few designers considered modifications, In 1936, Busemann suggested sweeping the wing as a means of increasing the critical speed. In this country, Robert Jones made a similar suggestion in 1944, But since eweeping the wing made it possible to use the wing itself for the entire control of flight, the next step was to discard the tail. Hence, the satisfactory design of a tailless airplane be- came @ necessity, from the stanépoint of more efficient performance for large, relatively slow airplanes, and for transonic and super- sonic aircraft and missiles,oy e y A ‘NOMENCLATURE Wing span Wing area Aspect ratio Wind chord, parallel to axis of symmetry except where noted Taper ratio, ratio of tip chord to root chord Angle of sweep of wings, measured at the Line connecting the 25% chord points of the wing sections, except where noted Mach nunber, ratio of speed of airplane to the free stream speed of sound Aerodynamic center Center of pressurePROBLEMS OF TAILLESS AIRPLANES Consider the outward appearance of the conventional modern air- planes. Whether they be large or small, slow or fast, they all have ‘the same general shape, differing only in nunber of engines and vert~ ical tails and location of the wing on the fuselage. Mot so the teil- less airplane! In the latter type the aspect ratio, taper ratio, sweepback angle, and size of fuselage all depend on the primary pur~ pose of the airplane in which they are incorporated. G. H, Lee, in reference 1, gives a broad outline of the basic types of tailless aircraft, and the type of application for which each case is suited. He concludes that large, heavy airplanes should have the fuselage entirely submerged in the wing, while the sweepback angle would be determined by stability considerations, The high speed airplane should have a fuselage plus a wing large enough to avoid high lift coefficients at cruising speeds, while the sweepback angle of the wing is determined by both drag and stability considerations, From the brief description of the problems of the design of tail~ one fact becom less airplane: outstanding: at the present time there is no broad background of prectical experience to refer to in the design of this type of airplane. This will probsbly be the case for some time to come, because of the fact that the appearance of the airplane will depend so much on the application for which it ie intended, Therefore, in the following procedure for the design of tailless aircraft, with the references given for the various phases of the design, one mst be aware that the specific problens encountered im a given airplane will be so important as to dictate the method or procedure in the design of that airplans.6. Furthermore the references given will in many cas not apply to extreme types of design, This is especially true for very small high-sy@ed airplones with extremely large leading edge. sweepback angles, and relatively large bodies, compared to the wing span. In the following part of the report reference will be made to ‘the specific problems of two types of airplanes: the slow speed type with relatively little sweep to the wing, and the very fast subsonic, or supersonic, eircraft with large amounts of sweepbeck. In a majority of cases the airplanes to be designed will lie some~ where between these extremes, and the problems will be different, but the designer who is aware of the limiting conditions has a basis for an approximation to the solution.WING DESIGN It seons reasonable to expect that the wing of a tailless airplane Will entail more design work than the wing of a conventional aizplane, due to the larger number of variables and uses of the wing. ‘That this is ectually the case can be seen from the following pro~ cedure to establish the wing planform and acrodynamic characteristics, ‘The two basic parameters of airplane performance and design are ‘the gross weight and maximm velocity. Though the decision as to these two factors might be influenced by some other criterion, in general they control the design of the rest of the flying wing air- plane, reference 1. For a given airplane configuration, such as fighter, light Domber, or long range bomber, the gross weight can be approxinated from experience with conventional aizplenes that perform the same function, or else a weight breakdown can be attempted to estimate the weight. ‘The airplane top speed can be chosen, thus determining the power plant to overcome the estimated drag, or else the airplane will be built for e given engine, and the speed will then depend on the drag. In either case the dreg coefficient, Op, and the wing area, S, will have to be estimated in order to find the drag. The ving area can be found fairly closely by using wing load~ ings of comparable conventional airplanes. The wing loading itself must be given some thought because it can have an effect on lateral stability, reference 2, The wing area is also a function of the allowable lending speed, since swept-back wings inherently have lower maximum lift coefficients than straight wings. Thus the wingarea is a compromise between a desire for a high maximm speed and a low stalling speed. But sincé the sweepback angle is not timate of wing yet known, it is better to make a preliminary area, and then try to estimate the approxinate drag coefficient. At the present tine there is a sonovhat limited amount of experimentel data on low speed drag and moment characteristics of swept-back wings for nearly ell types of wing planforns, but the high speed subsonic, transonic, and supersonic data is extrencly scerce. This means that the estimate of the drag coefficient will depend to a great extent on theory for high speed flight. For low speed airplanes the dreg coefficient can de estimated from wing alone tests on conventional airplanes, for tailless air craft with low sweepback, or from the experimental data of refer- ences 8, 4, 5, 6, and 7, In many cases it is necessary to know the approximate angle of sweepback even for low speed tests and it is mandatory for high speed drag calculations, Robert Jones, in refer- ences 8 and 9, clearly shows that the theoretical condition for super sonic flight with straight swept-back wings is thet the component of velocity perpendicular to the wing panel axis be less than the speed of sound. In actual practice the component of velocity mst be less than the critical velocity of the wing section. At this point there are two possibilities, depending on the design procedure. If a definite top speed is desired the sweepback angle of the wing can quickly be estinated fron the above sources, using references 10 and 12 to find the critical Mach number of a wing at high subsonic flight, In the more usual case where the drag% will determine the top speed it must be assumed that the wing will be swept back to an angle sufficient to appreciably diminish the effect of wave drag, which can be approximated from the two-dimensional theory given in references 12 and 13, and from the three-dimensional theory of references 14, 18, 16, 17, snd 16, The fuselage dreg coef- ficient, which is discussed in the following pages, is added to the wing coefficient to get the total dreg coefficient. Knowing the overell preliminary drag coefficient, the maximum velocity can quickly be estimated, and from this preliminary velocity, the induced areg coefficient of the airplane can be approximated from the references already given for the wing drag. ‘The estimated maximum velocity can now be determined and thus the necessary sweepback angle of the wing can be found from the references given previously. It mst be remenbered that the dreg estimate given above was made solely to determine the sweepback angle of the wings, before the aspect ratio, teper ratio, or other wings characteristics could de determined, Thus, when the two other major items in the wing design, the aspect ratio and taper ratio, are found, the drag and velocity estimates mst be reviewed to see if they have been changed vecause of better knowledge of the airplene layout. In other words the design of ony airplene is a repetitive process in that certain arbitrary assumptions mst de made in order to determine factors that can be used to check the initial assumptions. For the airplene in question the three parameters found so far are gross weight, wing ares, and sweep angle of the wing. Up to this point mention of the sweep angle of the wings has deen in terms of sweepback, but this is not the only case since10. theoretically sweep forward will work as well as sweepback. The adventages of sweep forward over sweepbeck are given in references Zand 19, The principal advantage is that wing-root stelling occurs before the wing tips stell, while the opposite effect is observed for swept back wings. The disadvantage of forward swept wings is that the forwerd travel of the center of pressure requires the fu lage to be very far forward. Mot mich more can be said about this configuration becouse at the present tine nearly all the date applies to wept back wings, the Germans being the only ones who have made any tests at all on the other types. The other factor that has not been clearly defined in the above @iscussion of the necessary sweepback angle is the reference line to which the eveepback is measured. This again depends upon the type of airplene to be considered. For the relatively slow speed air- plane with little eweepito the wings the angle of sweep is usually referred to the line connecting the 25% points of the wing sections, This is done because the section aerodynamic centers are located approximately at the 25% point for small angles of sweep. The idea of sweep referred to the 25% line loses its meaning if carried to wings with fairly high sweepback angles, of the order of 20° or greater, since for larger angles the section aerodynamic center moves rearward, reference 20, Actually, for supersonic aeroplanes the critical sweep angle can be considered to be either the leading edge or the line connecting the points of maximun thickness, es shown by Puckett in reference 15, At the present time the test date refers to the angle of the 25% line in all cases except as noted.n. The aspect ratio is the next major item to be found in the determination of the wing characteristics, Systenatic stability tests have been carried out to determine the effect of aspect ratio and sweep angle. These tests are shown in references 2, 4, 5, 6, 19, and 21, The results from isolated cases can be seen in references 3, 7 and 22, The results of all these tests indicate that the aspect ratio mst decrease as the sweepback angle increases in order to maintain longitudinal stability at low speeds, ‘The curve of aspect ratio versus sweepbeck angle, figure 1, is representative of the data from references 4, 5, and 6. It can be seen from this curve that aizplenes with high sveepback angles must have a very low aspect ratio, which can be determined quite closely fron the curve of figure 1, As can be seen fron the data of reference 5 it is possible to use slightly higher aspect ratios for a given sweep angle Af measures are taken to control the tip stall characteristics. Yor stability reasons the taper ratio of highly swept wings should be from 0,6 to 0.3 since the data of reference 25 indicates that higher taper ratios greatly enhance the possibilities of tip stalling. Unfortunately this condition for the design of the wing ti does not egree too well with the theory for the design of supersonic airplanes given in the previous references, for theory indicates that a wing with pointed tips is best. The decision as to the proper taper ratio to use will depend on which condition is more important, high speed or stability at low speeds. Summing up the conclusions arrived at so far it is seen thatlz. the gross weight, power, wing area, and sweepback angle, aspect ratio, and teper ratio have been approximated for high-speed air- planes designed to fly at supersonic or very high subsonic speeds. From this date the complete wing plenfor can be found, nanely the span, root chord, tip chord, angle of sweepback of leading edge, etc. Having for the first time « clear picture of the wing shape, it is necessary to refer back to the references given on wing dreg and check the calculations in the light of better know- ledge. Thus, if it is known thet a delta wing, or any wing with zero taper ratio, is to be used, then references 8, 9, 15, 16, 17, 24, and especially reference 18, can be used to make a more intelli- gent estimate of the total wing dreg. If wings with cut-off tips are used, the above references should be used in conjunction with ref- erences 13 and 14, though the theory is not yet known for such wing shapes. In the preceeding pages the sweepback angle and aspect ratio of the wing was considered for high speed aircraft, with the primary emphasis on speed and secondly on stability. For the large relativel; slow-speed airplanes the situation is altered in that the drag con- siderations will not determine the sweep angle because stability and load carrying ability are the primary factors. One fact that can be seen in all the low speed tests is that the asic maximum lift coefficient decreases with sweep, while the drag increases slightly for large sweep angles. ‘Thus, the maximim lift over drag retios will decrease with increase in the sweepback angle. Another effect is that the increment in lift coefficient attainable13, with a given flep will decrease with the sweep angle, while the angle of attack required for a given value of the lift coefficient will increase vith the sweep angle, From these facts it would s that a very low eweepback angle, 1f any, should be used, but the opposite effect is indicated in reference 29 where it is shown that “eweepback gives the wing an effective tail length adaptable for permits the use of high lift flaps at the center tailless airplane: of the wing where their lift increments produce only minor changes in the pitching moment about the center of gravity of the airplane, allows flaps for longitudinal control to be located near the wing tips where only minor changes in lift are necessary to produce the requisite pitching monents for trim, and permits more leeway in locating the center of gravity inasmuch as the aerodynamic center of the wing can be controlled by the angle of sweepback." The above discussion has perhaps helped make it clear that for low speed airplanes the determination of the swoepback angle can be quite difficult, since it will have an effect upon the maximum lift coefficient attainable with and without flaps, the landing angle, pitching moment, rolling moment, end several other factors in dynamic stebility. The one bright light in the wilderness of indecision lies in the fact that in addition to the material on ewept back wings, the many studies on straight wings with varying taper ratios and aspect ratios can be used to indicate the effect of modifications, Thus, the actual test data for wings with fairly low angles of sweepback is mich more voluminous, end covers more variables, than does the data for highly swept back wings, At this time the best materieluw for determining the correct wing shape can be found in references 2, 5, 6, 19, end 25, Other material applicable to special wing shapes, special problens, or as general reference material, can be found in references 3, 4, 21, 22, 26, 27, and 29. In the above discussion no mention has been made of the pro= cedure to de followed in the design of a slow speed tailless air- plane. It is felt that if any procedure is to be established, it ms of necessity be extremely indefinite due to the large influence of the type and application of any airplane under consideration, but an attempt will be made to set up the basic steps in design as timated follows: For a given gross weight the wing loading can be from conventional values or from figure 16 of reference 2. For an arbitrary selection of an aspect ratio and taper ratio, the wing span can be calculated and the sweep angle cen be approximated from the stability range curve shown in figure 1, The span and sweepback angle will give the effective tail length and thus the trimming momen‘ of ailerons can be calculated, using the data of references 2, 6, and 20, This moment must be sufficient to overcome the adverse diving moment dpe to flaps, references 2, 6, 20, 26, and 29, The design of the ailerons and fleps, above, will be a function of the necessary Lift and stability coefficients, which in turn will depend on the aspect ratio and sweepback angle. If the moment due to flap deflec~ tion could not be taken out by the ailerons, either the flaps could de reduced, the ailerons enlarged, or the swoop angle increased. It can now be seen that the various wing elements mst be varied ina systematic order, if possible, in order to get the most efficient18. wing to perform a given function, It is obvious that the above method is not precise, but until a great deal more systematic design data is made available it is the only possible procedure. ‘This paper has so far covered all parts of the basic wing de- sign except for twist and dihedral. These are also a function of the purpose of the airplane, as, for instance, the twist of a wing de- signed for very high speed flight should be sero in order to mininize the drag, while the sane wing should have washout in order to help prevent tip stalling et high lift coefficients. In general the good effect of twist is not enough to offset the adverse drag effect at high speeds, but it can be useful for low speed airplanes, as shown in references 19, 21, 23, and 28, The dihedral of a swept back wing affects the lateral stability and control to a great extent, ‘as shown in references 2, 28 and 30, Since a swept-beck wing has an “effective dihedral" due to sweepback, the actual dihedral angle can vary from positive to negative, depending on sweep angle, to get the correct rolling moment coefficient for the airplane. ‘The final element to determine for the basic wing is the aero- dynamic section to be used. It must be remembered that in swept- pack wings the section considered is that perpendicular to the line connecting the 25% points. ‘This definition loses its effectiveness, as wes previously explained, when applied to highly swept-back and tapered wings, so for these wings the section perpendicular to the Line of maximin thickness would be more apt for design purposes The wing section for.a high speed plane vill be considered first. The requirements are that the drag be low for high Mach numbers,16. the critical Mach number high, the maximm lift coefficient high, and the pitching moment curve stable, All this in addition to having as mich space as possible in the wing to hold fuel, control lines, etc. Since all these conditions cannot be met at the sane time, it is necessary to anend then in view of the desired perform ance of the airplene. They can be evaluated as follows: a very thin wing, of the order of 56 t/c, will have low drag and a high crit Mach number, but the maximm lift coefficient will be of the order of 0.7 to 1.2 depending on tho shape of the section. Furthermore, ‘these thin sections will have poor stelling characteristics on wings and will certainly not afford mich stowage space. A selection of a wing section will depend on the aerodynamic problens listed above, and the structural problems involved in manufacturing very thin wings, The question of stress determining the thickness will eppear in the discussion of spanvise lift distribution, The airfoil for a large airplane of slow speod need not be designed so mch for drag as for lift end stability characteristics. Furthermore, in large airplanes it is advisable to consider fairly thick sections so that the airplane can be designed as a flying wing where 1 equipment is carried in the wing and the fuselege is eliminated entirely. Also in thick wing airfoils it 1s mich easier to accomodate the various high lift devices listed in a subsequent section, Thus,the selection of any airfoil will be a function of the desired speed and load carrying capacity of the aizplene, and also, to a certain extent, upon the actual size. In any event the selection can be based on the standard compilations of airfoil data.a. In either high speed or slow speed airplenes a critical fector ie the spanwise 1ift distribution, Fron the material given in the references it is obvious that many of the problems typical of swept- ack wings occur because of the peculiar lift distribution, The general effect of sweepback is to increase the lift at the wing tips and reduce it at the root, while the opposite effect is observed with sweep forward, Thus it is seen that the high lift at the tips is a primary fector in premature tip stalling of swept-back wings of reasonable aspect ratio, which is one of the main reasons why wings with sweep forvard are now being considered for tailless airplane: ailable for ‘There are at present several different methods estimating the spanvise lift distribution of swept-back wings. The easiest methods, and those which give a very close approximation to the actual lift distribution, are applicable only for relatively slig angles of sweep, of the order of 15°.to 20° at the most, while larger angles entail more computations of which the results are still not to: good, The problem arises from the fact that the Prandtl lifting line theory cannot be truly applied to wings which are not straight, therefore either approximations mst be made to the lifting line ‘theory or else other methods developed, such as the lifting surface theory described in reference 31, In this paper Weissinger note that two methods are available. These are called the Fonethod, or lifting surface method, and the L-method, or lifting line method. In either case account is taken of the distribution of the circula- tion over an area. The F-method, when applied to swept-back wings, requires about eight hours of computations while the L-method, which gives almost as good results, can be calculated in about one third18. the time. Experimentel data from reference 20 indicates that the Inmethod gives fairly close agreement to the actual lift distribution. Other methods for finding the lift distribution of swept wings can be found in references 32 and 33, while reference 27 can be ex- tended to calculate the effect of small angles of sweep on the lift distribution. An estimate of the spanwise lift distribution for the wing of the airplane shown later in this paper, /1 = 63° is given on figure 2, where the distribution is calculated from reference 27 and ap- proximated from the curves given in references 32 and 33, From examination of the data of references 33 and 31 it appeared that the ‘two methods gave mich the same type of lift; therefore, figure 2 represents all the different methods, In each case the section lift coefficient corresponds to a totel wing lift coefficient equal to unity, so the curves are directly comparable, Elliptic lift distrib- ution is also plotted since Jones, in reference 8 shows that an elliptic 1ift distribution is approeched for delta wings lying well inside the Mach cone, The lift distribution with flaps or ailerons deflected can be approximated from reference 29. ‘The practical epplication of the lift distribution curve comes in finding the stresses on the wing and the position of the center of pressure of the entire wing behind the feference point on the wing root chord, Thus, if it ie known that the center of pressure of each section lies at the 25% chord point, the center of pressure of the wing must be referred to a point 25% of the root chord behind the wing apex. Jt is seen that this determination of the c.p. of the airplane1. 4s dependent on the wing characteristics, For wings with no sweep the e.p. of any section will be at approximately the 25% point for M<1, and at the 50% point for M> 1, references 15, 14, 16, 18, and 24, For delta wings of small vertex angle the c.p. theoretically remains at the center of the wing area, 2/3 of the wing root chord from the vertex, for all Mach number: references 17, 18, and 24, For wing planforns in between the two extremes the c.p. will have values depending on both planforn and Mach number.DESIGN OF CONTROL ELEMENTS In the first part of this report considerable nention has been made of the very high degree of importance the problem of stability and control assunes in the design of tailless airplanes, These prob- lens arise from the single fact thet the horizontal tail is not used, bat instead the wing performs all control functions through the use of sweepback end special control devices. That sweepback is not an inherent part of the static longitudinal stability of wings is a well known fact, but if straight wings with reflex airfoils are used, the other aerodynamic qualities of the airplane are adversely affected, The different aspects of this problem are shown by Jones, ference 25, Since the design of control elements for flying wings is a com- plete problem in itself no attempt will be made to discuss the design of the controle or the calculations and tests made to determine the stability coefficient derivatives. Instead the reader is referred to reference 19, which is the most complete and thorough source avail- able at the present time. In this report the items discussed are: remedies for tip stalling through the use of wing twist, change in airfoil section, flat plate separators, changes in planform at the wing tip, leading-edge slats, and taper; effects of power on longi- tudinal stability; longitudinal control through the use of bevels, special venting, slots ahead of the elevators, automatically controlle tabs, or spoilers; lateral stebility; directional stability of prop- ellers, fuseleges, fins, turned down wing tips, and automatic control; directional control; aileron control, using spoilers or elevons;ale end dynamic stability, Many of the itens listed above are quite thoroughly Giscussed in reference 2, while various stability probe lems are exanined in references 5, 6, 21, 25, 34, 35, and 36, Afleron effectiveness for two dimensional airfoils at supersonic speeds is given in reference 13, High 1ift devices are included with control elements because they have many probleme in common, and in sone ceses are actually the same elenent performing two different functions, as for ine stance the outboard flep which can also be used for aileron or rudder control, Furthermore high lift devices ere an inherent factor in stability considerations since they cause many of the stability problens and are also used to help solve sone of then, es does boundary layer control on the wing tips increase the lift and also helps control the tip stell, which affects longitudinal and lateral stability. The basic type of Lift inereesing device is the trailing edge flep, the enslysie of vaich is given in reference 55. Various other 1ift devices are given in references 37 and 38, These var dations are usually modified wing leading edges, such as a drooped nose, slot, or leading edge slat. An additional meens of increas ing the lift is through boundary layer control, either by removing the boundary layer by suction or else preventing boundary layer separation by blowing. This control, when applied to the wing tips, cen be extremely beneficial in controlling tip stelle. 4 combina tion of a nose flap, split trailing-edge flep, and blowing air through @ alot in the upper surface was the subject of a recent test, reference 39, on a 10% double wedge airfoil. The maximm value of the Lift coefficient was2.35, while for the basic airfoil the coefficient was 0.81. Any attempt to summarize the effect and use of high lift devices on taillese aircraft mst concern itself with the wing planform and airfoil shape, This is seen particularly in the use of trailing~ edge split flaps where the effect of sweepback is of prime importance in determining the effectiveness of the flaps, since for sweep angles somevhat greater than 60° the adventages of the flap are completely lost, and the flaps may even Decone detrimental to airplane perform ance. The wing airfoil is important in that it is very difficult to incorporate mechanical devices in thin wings with very sharp leading and trailing edges. Since the specificetions of high sweepbeck and thin wing apply directly to high speed aircraft it is seen that the design of an effective lift coefficient for these types can be ex- tremely difficult, but nevertheless a necessity since the need for low drag at transonic and supersonic speeds leafs to snall wing area end high wing loading, necessiteting high lift coefficients for landing. Even the design of flaps for large airplanes with fairly low swee angles is not simple since a large flap deflected along the rear edge of wing would lead to @ very strong pitching moment which the aileront could not effectively or safely trim out at high lift coefficients and high angles of atteck, This problem is also discussed in referenc 18 while one type of solution may be seen on the experimental airplant A.W,52, described in reference 40,FUSELAGE AND FILLETS The engineers who did the first experimental work on tail~ less aircraft-had in their minds the idea of the complete flying wing, that is, the wing alone performing all the functions usual~ ly assigned to wing, fuselage, empennage, and nacelles, The realization of their drean is seen to vary quite a bit with the actuality of the present experimental tailless aircraft, While the very large wings have almost no fuselage or external drag producing elements, the small transonic and supersonic airplanes are rapidly approaching the shape of uissiles, with a large body and relatively tiny supporting wing. Thus,the problems range from finding the shape of the most efficient wing to support a given internal lead, to finding an efficient wing to control the flight of a missile. At the present time the question as to whether a fuselage shall be incorporated in the design of a tailless airplene depends primarily on space considerations. Yor example, if en airplane is contemplated to fly at a given gross weight and speed then the wing can be designed on this information, but the wing which will fit the specifications might not have enough free internal space to accommodate either the engine or other necessary items, such as @ pilot, so provision must be made for a parasitic struc- ture to supply the extra volume, The extra structure ranges from canopies, nacelles, or turrets, to entire fuselages. In gen- eral the large, slow speed airplenes can have nearly all their structure and equipment submerged into the wing, while the snell ultra high speed craft must have a fuselage to carry engine and24, pilot, Because of the critical effect of the fuselage on the max- imum speed of, the latter airplane, it will be considered in more detail. If the wing plan form can be determined even approximately for a supersonic airplane then the fuselage volume requirenents can be determined, This is true because of the small percentage of the total volume contributed by e very thin and relatively emall high speed wing, The fuselage shape must be such as to give least drag for the minimm wing-fuselege interference effect, and least detrimental effect on stability. In other words, the fuse- age must be designed to give the maximum overall airplane perforn- ance, but not necessarily the least fuselege drag. From the stend- point of the fuselage alone, the design would be based on ballistic considerations, as given in reference 41, but the dest fuselage shape to give minimm dreg would include an extrenely long nose cone which has en adverse effect on stability. Some speculation, and work, has been done on the shape of the wing-fuselege intersection best suited for supersonic flight, and the conclusion drawn by the investigators was that the inter~ section should conform to the flow pattern of the air over the wing. The streamline shape for subsonic flight, calculated end tested in reference 42, seems to agree at least approximately to the shape useful for supersonic flight, qualitatively described in references 12 and 14, but in all cases the important fact is empha~ sized that the desired shape is a function of the lift coefficient and Mach number of the airplane in flight. Since the flight con- ition 4s not fixed the wing-fuselage Jinction could not work most25. efficiently at all times, but whether it could be designed to offset the flow disturbance at the junction of wing and fuselage for even a ronge of flight attitudes is not know, Furthermore the prob- lem of combining the wing-fuselege intersection to a favorable fuselage design presents further problems, since the best condition for supersonic flow indicates a cylindrical body of revolution with as few surface contortions as possible. The cylinder with end cones, 4s the obvious choice since it gives the maximum cross sectional area for a given perimeter, and also is subject to drag computations, From reference 41 it seems that the best fuselage shape would have a round cylindrical body with a conical or ogival nose in front and a partial cone in back, if a Jet exhaust is used. Formules devel~ oped by von Karman indicate an ogival shape is better for a given length/diameter ratio for the nose, see references 41, 43, and 44, ‘The wind tunnel tests of missiles are given in reference 45, ‘The design problem is further complicated by the necessity of having air ducts for turto-jet or ram-Jet engines, In large or slow speed airplanes the design of the duct is difficult but not critical, since the duct entrance is a small part of the total air- plane geometry, but for high speed aircraft the opening and duct must be designed for efficient engine operation at low speeds, such as for lending, and elso to give minim drag and entrance loses at ‘the design top speed of the airplane. The current manner is to place these ducts either at the nose of the airplane or ahead of the wing-fuselage juncture, The former method can reduce the over- all fuselage length while still keeping a low effective cone angle, while the latter method has possibilities of preventing the formation of shock wave at the fillet.26. RECOMMENDATIONS FOR FUTURE WORK Tt is recommended that extensive tests be nade to determine the characteristics of wings with sweepforverd in order to form a comparison between evept-forvard and evept-back wings, Also, tests should be made to find optimum airfoil shapes for super sonic flight, such airfoils in addition having good low speed characteristics. Extensive work mst be done to eystenatically evaluate high Lift devices and control surfaces for highly swept wings of low aspect ratio. In addition the effect of fuselage duct opening and wing~ fuselage junction on dreg coefficient mst be known,27. DESIGN OF SUPERSONIC TAILLESS AIRPLANE Yo partially explain the use of the references given in the preceding sections the design of a tailless high-speed experimental airplane will now be considered, This type of airplane is used as an example because of the variety of problems to be overcome, These problems sten from the fact that the airplane must take off and land at low speeds, but also be able to fly as fast as possible, ‘The latter requirement implies large seepback, low aspect ratio, 2 smell ving ares, and a very thin airfoil section. The attempt mast be made to reconcile these characteristics with the require~ nents for low speed lift and stability. Yor simplicity the airplane will be considered to be an ex~ perimental prototype. Tt will carry no unnecessary equipment for flight except for 300 pounds of testing apparatus, Hnough fuel will be carried to give an endurance of approximately one hour at full speed. The pilot must have sone means for parachuting safely from the airplane et any speed, also on air conditioning system must be carried to protect the pilot from extremes of temperature, If possible the airplane will be fitted with a tricycle landing gear. The pover will be supplied by a hypothetical turbo~Jet engine, of 4000 pounds stetic thrust at sea level, ‘The preliminary estimate of the airplane specifications is nade in the following manner, Since it is desired to have a very high top speed the airplane should be small and the weight low, with a high ratio of engine thrust to gross weight, Ifa ratio of 2/3 is teken, the weight would be about 6000 lbs. for en engine sea level static thrust of 4000 lbs, This ratio is considerablyhigher than thet of conventional alversft, but it 1s felt that {4 is justified in an experimentel alrolane subject to unknow bagards, A conventional velue of wing lonéing should be used in view of the extrenely rocr asxinom Lift cvefficlents expected fron . In other vords, the thin alrfotle associated with supersonic f1i to ettain even e landing speed of 130 mp.b, it is necessary to use wing loadings of the order of 60 Ibs/su.ft. which is rbout the high- est conventional velue now used, Thus the combination of weight end wing leeding lead to a wing with an aren of 100 sq. ft As a basis for first estinates it was assumed that Cy = 0.03, a nnmber besed on pest experience ir the desien of supersonic aim craft, If it is aleo assumed thet et 50,000 feet altitude the engine thrust 1s 1500 Ibs., then . = 4222 = 500 Is /H* TGS 7 3300 This corresvonds to a velocity of 1140 mh, or @ Mach number of K = 1.72. To obtain 2 velocity component normel to the wing of M = 0.75 means thet the wing must be swept back 65°, Since a teper retio of the order of 1/2 will de used, for low speed charecteristics, the lending edge of the wing can be considered to be swept beck 65° with~ out introducing very much error, After several attempts the wing shown in figure 4 was evolved. The wing has 98 squere feet of area, 8 root chord of 10 feet, teper retio of 0.513, aspect ratio of 1.70, and a spon of 15 feet. The leading edge is swept beck 65° while the 50% chord Line is svent 61°. It wee felt thet a delte wing vould be less desirable than atapered wing due to the very large percentage of wing area covered by the fuselage, The apace covered on a delte wing would be even combination lazger, end since the lift distribution across th nown it wes felt of win ter to use g and fuselere is not yet the wing plen shown. The dianeter of the fuselage, forty inches, is the least thet can fit eround the engine, which is assumed to have a dianeter of thirty-seven inches end to be fourteen feet long. The fuselege shape fore and aft ves determined by both drag end stability con- siderations. The after cone has the best angle for minimum drag, 15°, referencé 41, while the fore cone engle, also 15°, is a com promise between drag, stability, and visibility. The entrance duct to the engine is located at the nose of the fuselage, the actual entrence dianeter corresponding to the area required for en assumed flow of 73 1bs, of air per second at sea level at en airplene velccity of 300 i.F,H, This velocity wae arbitrerily teken as e design criterion to give the best possible valance between low speed end high speed engine performance. The leading and trailing cone lengths were determined by their coinci- dence with the entrence ond exit ducts, respective The entrence duct must de divided to have the air flov around the pilot and then converge at the ergine, The duct should be shaped so as to have as few friction and compression losses as possible, The nost efficient duct for supersonic speed vould heve close & converging section to slow the alr down,to so velocities,30, ake velocity. then a diverging section te reduce it to the engine 4 The shape of the fillets, figure 4, wes dictated nore by = desire to incorporate a tricycle landing gear in the airplane then from the sercdynsmic standpoint, though they showlé preserve the ation Lift distribution over the fuselage and help prevent the fo: of shock waves at the wing fuselege juncture. The fillets were made symmetricel, top and betton, deceuse of the anull incidence of the wing. ‘The vertical teil was designed to have en aron about one tenth (L to the wing's, This the wing area, and leading edge angle equ lea to 2 tell of 10 aq. ft. in area, aspect ration of 0.75, and leading edge sweenback angle of 65°, The rudder chord was assumed to be 25% of the teil chord, though this is subject to experimenta- tion, Other factors that must be checked experimentelly are the din mensions of the elevons and fleps, For a first approximstion it wes oseuned the elevons extended along S0$ of the sesi-spen, and the flops vere in the rest of the freo space along the trailing edge. For both the chord was teken ae ‘7 of the wing chord. For ease of construction the eirfotl was taken to be double wedge, of 106 thicknoss at the root ond 44 at the tip, in each cese measured to 2 chord perpendicular to the line joining the 50% choré. points, This thickness wes assumed, not for reasons of strength, ut deceuse It anneared to be about the minimum thickness thet would flow the incorporetion of the flans end elevons, and the Leading edge high 14ft device. the chord of the nose flep vas teken as 10% of the wing chord.a. To estimate the wing weight it is necessary to know the forces acting slong the wing. If it is essuned that the lift distributions of figure 2 are represontetive, to a first avproxima- tion, of both subsonic and supersonic flight, then the spanwise forces can be calculated. For this case the curve of Cohen, ené Weissinger, was assumed to represent the most likely condition. The results of the celculetions appear in figure 3, for a unit weight, or lift of 1000 1bs. The critical condition will occur at the juncture of the wing end fillet when @ totel wing lift of 6000 Ibs. is applied. A design factor of 18 is used to allow for unknown conditions. This fector is used to determine the design loads at the fillet, Knowing the loads, and thickness, the rew quired skin area can be estimated, end thus the wing weight cen de found to a first epproximation. Calculations showed the thickness of the skin should be about 0.02" to 0,03" for steel, 0.07" to 0,08" for duraluminun, and 0,12" for magnesium, Since skin stiffness, or absence from wrinkles, is extremely important in suversonic flight, the use of the magnesium would be preferable, since the weight of the wing would in eny case remain almost the game, nemely, ebout 400 lbs. The last point to determine is the position of the wing in relation to the fuselage, The a.c. of the wing can be approx- imated from the lift distribution curve of figure 2, Since the location of the a.c. is extremely doubtful for highly swept~ back wings, it was assumed that the esc. of eny section wes at the 25% point at subsonic speeds, and it wes in relation to this point that the wing wes located on the fuselege, It is realized that this ie a procedure that is definitely arbitrary,32, dat the justification lies in the fact that the center of gravity of the aixplene must lie ahead of the center of pressure, which in no cose will fall forvard of the 254 point, for stable flight. It 4s very obvious thet the above assumptions mast be checked by means of wind tunnel tests at subsonic and supersonic velocities to de~ ternine the actual location and travel of the a.ce ‘The airplane weight end belence chart is shown in Table I, which indicates thet for either loaded or unloaded flight the atm plane is statically stable, since the fartherest forvard position of the a,c, 43 approximated to be 151" from the nose, The airplane drag estimate is shown in figure 6, This will be the miniman drag because it was assuned there would be no entrance duct drag or wing-fuselage interference effects. The performance estimates of the airplene ere shown on figure 6, for a sea level stetic thrast of 4000 1bs., while in figure 7 the estimate of performance is made for a sea level static thrust of 6000 lbs, It should be noted that the critical velocity near sea level will probably be limited by the increase in temperature of the airplane when flown at high Hach numbers. At an altitude of 50,000 feet the endurance will be approximately fifty minutes at full speed, with the first engine in use.TABLE I WEIGHT — BALANCE Iten Weight Distance from nose - lbs. = inches Engine 2200 176 Pilot 200 60 Radio 30 22 Instruments 20 36 Test equipment 300 92 Fuselage 400 140 Rudder 60 294 Controls 30 140 Hose lending gear 70 90 Mein landing gear 140 173.5 Air conditioner 80 92 Wing 400 116.5 Hose fuel 400 92 Teil fuel 420 260 Wing fuel 21000 180.5 Loaded condition 5750 1bs. 150,8 inches 866,860 Unloaded condition 3930 Ibs. 150.4 inches 590, 360 ‘The wing location was arbitrerily selected so that the fartherest forward position of the center of pressure will be 161.0 inches from the nose of the airplane, Therefore,the airplene will be statically stable at all times,34 PRELIMINARY WIND TUNEL RESULTS A 1/t2-qcale model of the aizplene shown in figure 4 was tested in the Pasadena Junior College low speed wind tunnel, The model was the actual size of the drawing in figure 4, The intake duct and engine were simulated, approximately, by a round hole Yored the length of the fuselage, the diameter of the hole being equal to the intake and exit duct diameter. Also, because of the very short length of time evaileble for testing the model, the tests were made without fillets between wing and fuselage; there fore, the fleps, when used, extended inboard to the fuselege. Four aifferent tests were mede, “he originel test was for the wing and fuselege, then wing and fuselage with a pertial span 256 chora split flap, deflected 60°, extending inboard, from the elevon position shown in figure 4, The third test was for the full span flep deflected 60°, where the elevon chord was also 256 of the wing chord, The last test was made with the full span split flep on the treiling edge of the wing and a 10% chord nose flep deflected 150°, In each test the Lift, dreg, and pitching moment were meas~ ured, The dynanic pressure, q, of the tunnel was computed using static pressure readings end the tunnel calibration curve. The average tunnel speed was 106 ft/sec. The lift and drag coeffic~ lents were computed from = = Bed Mapes _ . GBs + Cn, Le where L is the 1ift measured, D is the drag measured, and M is the moment meesured in the tests, while d ie the drag due to the35, eupporting struts, 4 = 0,01 q, and # is the tare moment. The meen aerodynamic chord, 0, of the model wos 0.653 feet and the model wes supported at 0,651 0. The location cf the aerodynamic center was found by using the relation c Sn — &(E- E) hae. = If the Oy, 4, if constant, then Ac) re Wa) ~t- = where Oy, ie the moment coefficient referred to the suoports and & As the location of the wing support, while £ is the locetion of the ace Plotting Cy, ageinst G gave Ae) 0.308 fer sero Gy, therefore ~& = 0.343. Since in the original desizn it wes assumed that the a,c, would te at — = 0.250, it ie epper- ent that the wing is located too far behind the center of gravity. Hience, though the airplane is steble, the moments required to trim it for level flight would be unnecessarily high. The moments ebout the a.c. were calculated from Mae ~ Coty ~ 0.3086. Enowing the pitching moment coefficient, the complete wind test results are plotted in figure 8, Comparison of the resulte to preliminary eetimetes indicate several points thet affect the design of the airplane. The most Anmortant point is that the airplane is unstable at relatively high lift coefficients; therefore, modifications must be nade in the design to eliminate this instability, The other factor is that the small slope of the lift curve neceesitates either » very high36. angle of attack for reasonable landing speeds or else landing speeds somewhat higher then ordinary, Another interesting feature is that higher lift coefficients can be obtained with trailing edge split flaps, though previous references had indicated that this would not be the case, The continued increase of the lift coefficient with angle of attack is interesting in respect to the tuft studies, figure 9, Though tufts were used on the wing for ell configurations the results seemed to de about the same, so only one set is shown. The flow over the wing was uniform and parallel to the free stream flow at negative and very emall positive engles of attack, wing A, At some angle between 2° end 5° the air would suddenly flow along the outer half of the leading edge of the wing, and renain nearly parallel elsewhere, The turbulent action of the tufts indicated that that part of the wing was stalled in the spanvise direction, wing B, For all angles of attack greater than the critical angle the tufts gradually assumed the cheracteristics of wing C, that is, more end more of the outer portion of the wing in the stalled condition, At no angle of atteck up to 34° did the trailing e of the wing appear to be stelled,Le 2. 3. 4, ue a7. REFERENCES Lee, @. H.a The Case For the Tailless Aircraft. Journal of the Royal Aeronautical Sciences, Wov, 1946. Soules Hartley A.: Influence of Lerge Amounts of Wing Sweep on Stability and Control Problems of Aircraft, N.AsCeA. T.M, 108, 1946, Seacora, Charles L, Jr, and Ankenbruck, Herman 0,: Determina~ tion of the Stability and Control Characteristics of a Straight-Wing, Tailless Fighter Airplene Model in the Langley Free-Flight Tunnel, N.A.C.A. ARR No, L5X05, 1946. Seacora, Charles L, Jr, and Ankenbruck, Herman 0,: Effect of Wing Modifications on the Longitudinal Stability of a Tallless All-Wing Airplene Model. N.A.C.A. ARR No, 15023, 1945, Shortal, Joseph A. and Maggin, Bemard: Zffect of Sweepback ‘and Aspect Ratio on Longitudinal Stability Characteristics of Wings at Low Speeds. W.A.C.A TAN. 1093, 1946, Letko, William end Goodman, Alex.: Preliminary WindTunnel Investigation at Low Speed of Stability and Control Char- acteristics of Swept-Back Wings, N.A.C.A. T.¥.1046, 1946, GSthert, B.: High Speed Measurements on a Swept-Back Wing (Sweepback Angle = 35°), ‘Translation in N.A.C.A. TM. iioz, 1947. Jones, Robert T.: Wing Plan Forns for High Speed Flight. MACA, TN, 1032, 1946, Jones, Robert T,: Thin Oblique Airfoils at Supersonic Speed. HAC, TM. 1107, Lindsey, W. F., Daley, Bernard N., and Humphreys, Milton D,t The Flow and Force Characteristics of Supersonic Airfoils at High Subsonic Speeds. .A.C.A. TN. 1211, 1947. Hilton, W. 7, end Pruden, F. W.,3 Subsonic and Supersonic nh Speed Tunnel Tests of Faired Double Wedge Airfoil. British R & M, No, 2057, 1943, Ackeret, J.: Airforces on Airfoils Moving Fester than Sound. WekCoke TM. B17, 1925, Ivey, Reese 1. Motes on the Theoreticel Cheracteristics of ‘Two-Dinensional Supersonic Airfoils, W.A.Ged. T.N. 1179, 1947,14, 15, 16, 1, 18. 19, al. 22, 23. 24. 25. 26, ar. 28. 38. Bonney, E, Arthur! Characteristics of Rectangular Wings at Supersonic Speeds. Journal Aeronautical Sciences, Feb., 19 Packett, Allen E, Supersonic Wave Drag of Thin Airfoils. Jour- nal “Aeronautical Sciences, Sept. 1946, Legerstrom and Wall: Formilas in Three-Dimensional Wing Theory (1), Douglas Aircraft Company S.M, Report 11901, July, 194 Hayes, W. D., Browne, S. K., and Lew, R. J. Linearized Theory of Conical Supersonic Flow with Application to Triengular Wings, Worth American Aviation Report No, NA-46-816, Sept. 1946, Snow, B. M. and Bonney, E. A.t Aerodynamic Characteristics of Wings at Supersonic Speeds. Bumble Series Report No. 55, Johns Hopkins University Applied Physics Laboratory, March, 1947. Stebility Research Division: An Interim Report on the Stebility and Control of Tailless Airplanes. NACA ACR L4H19, 1944, Thiel, A. and Weissinger, J.i Pressure Distribution Measurements on a Straight and on a 35° Swept-Back Tapered Wing. NebsC.t TM, 1126, 1947, Pearson, Honry A, and Jones, Robert T.: Theoretical Stability end Control Characteristics of Winge with Various Anounts of Taper and Twist, N.A.C.A. TLR. 635, 1938, Campbell, John P. and Seacord, Charles L., Jr. Determination of the Stability and Control Characteristics of a Tailless All-Wing Airplane Model with Sweepback in the Langley Free- Plight Tonnel. M.A.C.A, ARR L6A13, 1945, Anderson, Ramond F,: A Comparison of Several Tapered Wings De- signed to Avoid Tip Stalling, NA... T.H. 713, 1939, Jones, Robert T,: Properties of Low-Aspect-Ratio Pointed Wings at Speeds Below and Above the Speed of Sound, W.A.C.A. TN, 1032, 1946, Jones, Robert T,: Notes on the Stability and Control of Tail- less Aircraft, N.A.C.A. T.N. 837, 1941. Pitkin, Marvin and Maggin, Bernardi Analysis of Factors Affect ing Net Lift Increment Attainable with Trailing-Edge Split Flaps on Tailless Airplanes. N.A.C.A. ARE Wo, 14118, 1944, Anderson, Raymond F,: Determination of the Characteristics of Tapered Wings, N.A.C.A, T.R. 572, 1936, Garbell, M. A.t Effective Control of Stalling Characteristics of Highly Tapered end Swept-Beck Wings. Journal Aero- nautical Sciences, Oct. 1946,29, a1. 32. a. 35. 36, 37, 38, 39, a, 42. 39, Pearson, Henry A, and Anderson, Raymond F,: Calculation of the Aerodynamic Characteristicn of Tapered Wings with Partial Spen Flaps. N.AsC.A. T.R. 665, 193% Maggin, Bernard and Shanks, Robert B,: The Effect of Geometric Dihedral on the Aerodynamic Characteristics of a 40° Swept— Back Wing of Aspect Ratio 3. N.A.C.A. Toi. 1169, 1946, Weissinger, J.: The Lift Distribution of Swept-Back Wings. Trensletion in N.A.C.A, T.M. 1120, 1947, Cohen, Doris: Theoretical Distribution of Load over a Swept- Back Wing, W.A.C.A, 1942, Matterperl, W.1 The Calculation of Span Load Distribution on Swept-Back Wings. N.A.C.A. T.¥. 834, 1941, Garbell, Meurice A,t Theoretical Principals of Wing-Tip Fins for Teilless Airplenes and Their Prectical Applications. Journ} Aeronautical Sciences. Oct. 1946, Bennett, Cherles V. end Johnson, Joseph L.: Experimental Dete: mination of the Damping in Roll and Aileron Bolling Bffec- tiveness of Three Wings Having 2°, 42°, and 62° Sweepback. W.A.C.A, TAN, 1278, 1947, Harper, Charles W. and Jones, Arthur L,: A Comparison of the Lateral Motion Calculated for Tailless end Conventional Airplanes. W.A.C.A, T.N, 1154, 1947, Lene, H. G.: Force and Pressure Distribution Measurenents on a Rectangular Wing with a Slotted Droop Nose and with either Plein and Split Flops in Combination with a Slotted Flap. Translation in N.A.C. T.M. 1108, 1947, Krueger, W.i Systematic Wind Tunnel Measurements on a Leminar Wing with Nose Flap, ‘Translation in N.A.C.A, T.M, 1119, 1947, Pollack, A. D, and Reck, F, ¥.1 A Study of Methods to Increase the Lift of Supersonic Airfoils at Low Speeds. California Institute of Technology Professional Thesis, 1947, ‘Twin-Jet A.W, 52: Flight and Aircraft Engineer, Dec. 19, 1946, Charters, A. C.: Some Ballistic Contributions to Aerodynamics. Journal Aeroneutical Sciences, March, 1947, fatkins, Charles E.: The Streamline Pattern in the Vicinity of an Oblique Airfoil. N.A.C.A, T.M. 1231, 1947,40. 43. Yon Karman, Th.t The Problem of Resistance in Compressible Fluids. Reale Accademia D'Italia, Roma, 1936. 44, Von Karman, Th. and Moore, Norton 3,1 Resistance of Slender Bodies Moving with Supersonic Velocities, with Special Reference to Projectiles. ‘Transactions of A.S.M.E., 1932. 45. Walchner, 0,: Systematic Wind-Tunnel Measurements on Missiles. Translation in NeA.C.A. T.M. 1183, 1947,ferevencer 4.5, &TVET STUOIES Fig 9.