The Sail-Plane Citation Kiesling, Louis (1924) The Sail-Plane. Bachelor's thesis, California Institute of Technology. doi:10.7907/YB5X-9C87. https://resolver.caltech.edu/CaltechETD:etd-05052004-132633 Abstract No Abstract. Item Type: Thesis (Bachelor's thesis) Subject Keywords: Engineering and Economics ; Aeronautics Degree Grantor: California Institute of Technology Division: Engineering and Applied Science Major Option: Aeronautics Engineering Minor Option: Economics Thesis Availability: Public (worldwide access) Research Advisor(s): Merrill, Albert Adams Group: GALCIT Thesis Committee: None, None Defense Date: 1 January 1924 Record Number: CaltechETD:etd-05052004-132633 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-05052004-132633 DOI: 10.7907/YB5X-9C87 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 1617 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 07 May 2004 Last Modified: 03 Oct 2019 23:09 Thesis Files Preview PDF (Kiesling_l_1924.pdf) - Final Version See Usage Policy. 4MB Repository Staff Only: item control page

California Institute of Technology Pasadena, Ualifornia Class of 1924 Department of Aeronautics SAIL-PLANE - 'iThesis of Louis Kiesling 1924 anger OF | 7 ‘ fed CAR Lee he Ue. og as The Cohfornia Lnshfose of Technology. So// plore. Built and designed by Gerard F. Velfee. Supervised by Prot AA. [lerrii. Assisted b y JF. AW. Cloverie, Pilot: J. AW. Cloverre. Thess Wark of Lows Fiieshing . Soh. 2k 1@Paa. In the years 1920 to September 1922 a great stir of interest was caused by the unheard of successes of certain German soaring machines. For two years, because of the prohibiting by the Allies of large machines, their efforts had been directed to the design and perfection of light weight motorless ships. They held gliding meets in the Rhon Hills and established soaring and gliding records that by far surpassed anything that had ever been done before. These events incited interest in France and England and provoked certain students of one school in the United States, the Massachusetts Institute of Tachnology, of Boston, Mass., to build a glider which they subsequently entered in the French meet. At the same time this new spirit could not have failed to come west to California, and so we find two students of the California Institute of Technology en- deavoring to put into practice some of the engineering principles and theory that they had acquired in the previous three years. Our practical engineering experience was very meagre, and we had to rely greatly on common sense and the helpful suggestions of our competent advisers. Realising the great success of certain glidere in Germany, we naturally took these for models and in particular the ideas of the foremost German glider, the Hanover Vampyr, of that time. In doing this we were favored by having access to the published theory of the designer of this last machine. Our ship is not an imitation or replica of the Hanover, far from it, but it does embody the general aerodynamic features of this plane. In construction the two are entiraly different and we can consider our constructional work as completely distinct, the design following the regular prescribed engineering processes and formilae. We were aided also by the fact that each part of the design must be a compromise or have the approval of all four of us, all different in tempertmerit. Furthermore we all desired, and for financial reasons were forced to embrace, simplicity af design and constructional cheapness. We used standard aeroplane parts such as pulleys, horns, etc. wherever possible. Our small fund of practical building knowledge made it imperative, for safety reasons, to use simple straightforward designs. We were also Limited in the way of wood and metal tools. We have friends in the Pasadena High School, BarnHardt Alroraft Company of Pasadena, Douglas Aircraft Company of Santa Monica, and our school authorities to thank for their unlimited advice and material aid. Above all mare we grateful to Mr. Merrill and Mr. Claverie in their unceasing enthusiasm and aid in the consummation of this work. Mr. Merrill helped us with our engineering difficulties and made the financing of our project possibbe. Mr. Claverie, with his long experience and deep knowledge, did all of the metal work, a not incon- siderable amount, and has piloted the final result of our efforts. - When judging or condemning this plane, bear in mind the relative lack of experience, age, shop facilities and above all the size of the project. We assert ourselves in saying that we would never have attempted the job if we had realized only in part the time, trouble, and effort required, bu’ we are now sincerely glad that we stuck to it and put it over, good or bad. Most of air work and all assembling and model research was done in the Aero Laboratory of the California Institute of Technology. Photo, #1. It has been mentioned that we were greatly governed by the precedence set by the German gliders and by imperative low cost, and by the lack of facilities. It must not be thought that safety of construction was en- dangered. Weight considerations were relegated to those of secondary importance. On the whole we were extremely conservative in everything, a very wise thing for our first attempt. Our materials were of class A, in fact, they conformed to government requirements. We used basswood plywood instead some of the more expensive kinds, but not at the expense of safety. Our considerations lead us to first adopt the cantilever monoplane type of wing because of our ow private penchants, and because of the general adoption of this type in foreign practice in gliders. We chose cantilever construction to eliminate all the parasitic drag we possibly could. The wing in relation to the fuselage is of parasol type with the cockpit in front of the leading edge of the wing. This affords excellent visibility to the pilot. The monoplane or biplane polemic has a®guments for and against each, and no absolute basis of decision has yet béen evolved, hence our choice. The landing gear of the Hanover was unique but had several disadvantages. We first decided to have skids, made out of hickory. This would have lifted the fuselage away from the ground and thus avoided contact witk stones and brush. It would have been of easier construction than that of the Hanover. We ended by designing a wheel landing gear without across axle. This eliminated to a degree the chance of nosing over when landing in brush country. Again, the ship would be easier to launch and transport with wheels than with skids. We arranged to have our rubber shock absorber inside the fuselage, thus eliminating some parasitic drag. The skids were designed and made but we found it so hard to bend them into the double curve necessary that this difficulty pre- oipitated us happily to a choice of wheels which we now think are very much better. The salient feature that all soaring machines must have ie quick maneuverability. This is obtained by having large control surfaces and small moments of inertia of the parts about the center of gravity. We followed the Hanover in having a short tail. We decided to have a small horizontal stabllizer and to have our elevator of twice its area. This combination was preferred to that of just the elevator alone as it would give a small inherently etable force and also the desirable effect of a large elevator. We made our fin and rudder after the design of the Hanover. The large fin would tend to nose the machine into the gusts of wind. Our rudder was slightly larger than that of the Hanover. The new interest in gliders and small horsepower “machines is the result of a desire to free the aero- plane industry from the inefficient methods of war machines. Small motoroycle engines could now be made use of, The lack of excess power would tend to lead to improvement along lines which were being neglected. Increased aerodynamiocal efficiency was the goal. Ex- perimerits by the Germans sought to the development of more efficient wings and lead in part to the cognizance of the thick high lift wing and it's improvement. The designer of the Hanover was first to emphasize the _ekhking velocity or rate of descent and to develope formulae concerning this. To soar means to sink more slowly than the wind rises, hence the great importance of a small minimum sinking velocity. Good gliding angle was of secondary importance in soaring. The new donditions of soaring make it imperative to teach pilots the necessary facts in relation to making use of gusts, rising currents of air, etc. In our problem, which was very similar to that of the Hanover, we must find an aerofoil section which had a very deep camber, deep enough to make cantilever con- struction possible, and to allow an aspect ratio of at least 10. The sectio st have a high Litt coefficient, a good lift over areg (bl r3t10, and primarily a large value for lift, cubed, over‘@rag, squared.(m} The latter . oriterion is embraced in the factors‘for minimum sinking speed. The wing must have an area and wing loading low enough to permit a low landing speed, around 20-25 miles per hour. This must be attained to render landing and taking off easy. With a speed of 23 miles per hour and a 1i tt coefficient of about .73 we obtained a loading of Our weights were the next thing to assume. We cal- culated from these weights and the loading that an area of about 150 or 160 sq. ft. was necessary. In view of the above facts we then looked about for thick aerofoil sections and narrowed these down to a choice of one of three, namely, the U. S. A. 27, Martin #2, and Fage and Collins #5. We desired to get an aspect ratio of approximately 10, so our first task was to reduce the results of the reports from a basis of 6 to that of 10. The drag of a wing being divided into two parts, namely section and induced drag, with the section drag remaining equal for all aspect ratios, we computed the section drag, found the induced drag for aspect ratios of 10, added ese two for total drag at 10, and then found values of a) and which are plotted on accompanying gmaphs. e chah¢e from 6 to 10 changes the drag enormougly, ‘Va4dN3% {OS ‘Hass3 © WLEVEL.. LSS Loe i at LALEK LAL. ae. i i ’ 3.12: get i i ? i t i i i i ee HOD PEE CON CMEOA MAGN OO (YS Lor 1 mb i 7 one i 1 ; et i* i i : aoe! i dee i : i N “| : - i t : e : } uN ! , Lad i he : ro To woe + : Ss aa feed i ' | : tani ee eT ~~ —e : a i so ng} ee Pd \ : : a Nl Z ‘ i ae ow : ioe DL. H be Z. nes NJDEL cenebivee mere etree 1 . on JO HaSSa FP V3s4Na» HO PES CON SMYOA MAAN Ty: te po ane ies poctte te & e bothes 2 a oad oe eee i gives. 2 | [ 4 2 t | i 3 SW tos | | Yak ijeaatae tee “Yass 8 T24anay web fe ees | approximately one-third, by virtue of the lowered induced drag. Thus it 1s seen why we endeavored to get a value of 10 if possible. From the graph it can seen that Fage and Coll #5 has the best value of /i,) and is second in regard to (#) . The outstanding quality is it's camber of almost 19%) permitting lighter and stronger construction. We chose that section. See figures 1, 2, 3, and 4. We then met a problem which seemed difficult and almost militated against our using this section. We found that the tralling edge was so thick that it might be Gifficult to construct ailerons and hinges. Again, we thought that even if we could make an air proof hinge, the ailerona would not be very efficient. Next we hit upon the scheme that the Hanover used, a tapering outer section with flaps. This was adopted with flaps giving rolling moment comparable to or larger than that of the Hanover. We did not adopt the sweep back that they used, but we did make our top surface a straight line thus giving us a slight dihedral at the outer tips of about 2.5 degrees. A tapered wing has the advantage, according to recent tests, of higher efficiency and less end loss. From this it can be seen that the flaps will be of triangular plan shape, to give the entire com- bination a rectangular form. Our efforts to decide the shape of the ends of the wing tips lead us to just chop them off square. This was easy. Exhaustive tests have not lead to the adoption of a standard shape by manufacturers. The tapered wing has another advantage in that it produces @ lighter wing loading on the beams at the tips, this effect moving the center of pressure of each half toward the axis of symmetry. Our fuselage is @ modified Hanover having about the same size and shape. This shape, although reotangular in cross section, lends itself to ease of constunction. Recent papers of the 1922 and 1923 German gliders show the almost general adoptiqn by other Builders of this type. This seems to bear out the contention of the Hanover designers that this type is very efficient in conjunction with mono- plane wings. They claimed that the combination of wing and fuselage gave a lift value equal to that of the wing alone and gave in addition a low parasitic drag. Our changes in it's shape consisted in making it adaptable to skids. To our chagrin we later found that we could not make the con- templated skids and so we had to design at the last moment a wheel landing gear. Happily this turned out to be a success,and approval obtained. The tail portion had to be thickened to allow the easy flow of the air around the new position of the fuselage to the wing, which position we obtained from wind tunnel tests. Our pilot must sit in front of the leading edge of the wing. We did this to give him good visibility and to make a favorable location of the center of gravity easy of attainment. We made extensive model tests in the wind tunnel. This tunnel has an old N.P.L. balance, a 4 ft. sq. channel and gives a maximum air speed& of 44 ft. per second. Photo. #2. : ey =, Wid Tunnel Ga TF: Our models were made to 1/20 scale and our wing had an aspect ratio of 6, to make our tests comparable to tests in other tunnels. We tested the wing alone first and found that the flow of the air broke at 9° thaldence instead of the 17 according to the N.P.L. reports. We attributed this to our low speed, which gives a low VL (velocity x chord) of only 11. Mr. Norton of the Washington D. C. laboratory was written to relevant to the safety of our using this wing. He replied that the high value of VL in actual flying conditions would wash out this break in the air flow. Our results below this break were the same ag those of the N.P.L. tests, as far as they went. The model was made of wood except for the skids: and tail surfaces. We tested this first by varying the position of the fuselage to the wing until we obtained a position which gave us the beat fh ratio. This necessitated our changing the shape of the fuselage slightly. Having done this we put on a horizontal stabilizer of span of 9 ft. and chord of about 2 ft. No elevator was put on because we wanted to determine the smallest stabilizer which would be necessary to just make the ship inherently stable. We assumed a center of gravity, ‘corresponding to current practice, and constructed the model spindle so that the model would rotate as a whole about this assumed center of gravity. By varying the angle of the stabilizer with respect to the wing and by gradually cutting down it's area, we obtained a small surfact which»would give a stable couple with no catastrophic instability. That is to say, with this small stabilizer we obtained a normal flying position to which the model would always return of iteelf, if changed by forces over which the pilot has no control, to another position of larger or smaller incidence. This normal flying position was made to correspond to the position at ‘which the sinking velocity, was smallest,or in other words, to the position where the was maximum. With this stabilizer, then, the ship would come ground, by itself, if perchange, the pilot took his hands off the control stick. Our stabilizer needed to be only 1 ft. x9 ft. and at an angle of -1.3 to the wing. He changed our results from aspect ratio of 6 to that of 10 by use of Prandtl's formula for induced drag, and ob- tajned for the whole ship a best of 16 at 2.5 and a best é at 4". This means that the bést angle of inoldence for sdaring or staying aloft longest is larger by 1.5 than the best angle for gliding or going farthest. Our computed sinking velocity was 2.7 ft. per second, which is equal to or better than those of several German gliders. This we attribute to the Fage and Collins #5 aerofoil section. See figure (5). 'g3SS3 Y T344N3y ome att decd besnnlenemals HEF ; c riot ems pees e WY coe EE ie Seeed esas? ii rit Photo. 3 Spon - 40. Wheel tread 5. Width Fuselage = aig Height above ground 64. Front View. Photo. +. Center Section Wing 22 Ovter - "9 ee. Elevator Span 4 Back View. Leng th overoll | 8. Depth Fuselage 33, Chord at tip wing 2 3. Side View. Photo, ©. Span Wing sll i Chord +t Length overall 18’ Spen Tail q Chord 1 3’ Plon View. Photo. Ve no 7ech, Aero. Lobora/ory, Sor/ plane. at : Foss) Fie /d. : = ; Arcocha, Se Cob GAIT. 8% Our work was done for the most part in the wind tunnel, except fof the metal parts which were made at the Barnhardt Aircraft Company. Pasadena High School kindly allowed us to use their wood working machines. We were able for a while to use the wood shop in our own physics department. The continuity of the work was approximately as follows, with overlappings here and there: (1) The tail portion of the fuselage, (2) the tail surfaces, (3) the center portion of the wing, (4) the front portion of the fuselage, (5) the . outer wing sectiorm and ailerons, (6) the controls, (7) the landing gear. Oe The materials used were grade A spruce, which we tested for checks and spiral grain etc., 5/32" basswood plywood, cold rolled sheet steel, Shelby seamless steel tubing, Roebling aircraft wire, casein glue, etc. The general specifications for the whole ship are as follows: : Wing Area, including flaps 180.0 sqe ft. Aileron-net area-each of (2) 10.0 sq. ft. Horizontal stabilizer area &.& sq. ft. Elevator area 17-6 sq. fte Vertical stabilizer area 54 6 Rudder area jy 7" Angle of wing to direction of flight at normal flying attitude 4 degrees Angle of stabilizer to wing-zero - getting ~1.3 degrees Angle of wing to ground at rest - 13.5 degrees Movement of Ailerons plus-minus 30 degrees Movement of Elevator plus 24=minus 30 degrees Movement of Rudder plus-minus 20 «degrees Weight including pilot 410 =o lbs. Weight without pilot 270 lds. Wing Loading including pilot 2.2 #/sq.ft. From the advice proffered, we decided to use the two beam type of wing construction, the spacing of the beams also following conventional practice and being in the ratios of about 1/5 and 5/8 from the leading edge. To avoid difficulties, we made the beams of I beam section because this form allowed the placing of spacer blocks at any time after the gluing of the web and flanges. The web was made of 5/32" basswood 3-ply veneer,and flanges, of which there were two above and two below, were of spruce with the Plywood glued between. Since our span was 40', we were forced to split up the wing into three sections. Spruce of 22 ft. length was the longest obtainable, and so our center section was 2eft. long and each outer section 9 ft. These sizes allawed ease of handling. . Since we used no external bracing, our bending moments assumed those of a cantilever, the center bay being only 26 in. wide. The loading for normal flying was 2#/sq.ft. We computed on a basis of load factor of five, raising this, therefore, to 10#/sy.ft. From a knowledge of the center of pressure movement and the beam spacing, we assumed, conservatively, that each. beam would have to be able to withstand the total load alone. Computations for beam sizes were made simply according to straightforward beam formulae and cantilever” moments. In order to cut down on the weight, we tapered the flanges from the center to the outers along a straight line untih it became impracticable far them to be made smaller, and so the outer section beams were overstrength. To determine the strength of the web in shear, we subjected a 4ft. length of beam to test and foumd it had a safety factor of 2 above our load factor and also that it required no stiffeners. Our standard rib was made as the drawing indicates (pa #/) mainly because of it's simplicity and strength. It was heavier than necessary. The strength was computed on empirical formulae prevalent, assuming a total load of about 55% for each rib. The web was unique in that it was made of 7-1/8" square strips glued together to the shape of the section. This was done by gluing 7-1/8" x 3" wide strips in a jig to the aerofoil shape and then ripping these into webs 1/8" thick. The rid had it's flange running over and unde# the beams and was fastened by gluing and stapling the flange to the beams and placing blocks at the side of the web. The ribs were spaced at 15-16 in. apart. 10 Standard fib yinerdence bracing -Sprvee je * Fa ° Sorvee Web - + thick. [deep - made of 1-§ 94 glued together. yk right my , tom jay & vA Flange - ateed + stapled % Beam, 254. flange ,one side only. Sprucq sat A) glued to web Q) and Flange (B) . Our box ribs took the compression of the drag bracing and were similar to the standard ribs except for the center section having a solid web of plywood. The flanges, of couree, were larger according to size of loads. The drag bracing we adopted aroused a bit of skepticism on the part of many. One engineer advised us to use strips of spruce 1/16" thick in place of wires. This plan was very much simpler and easier to make and we followed his advice. I have since changed my opinion as to it's wisdom. Our drag bracing was a double system, one above, and one below, allowable because of the thickness of our wing. Either system was more than strong enough to support the total weight of the ship in a dive at limiting speed and | the double system also gave us the advantage of resistance to torque of the whole wing. The strips of spruce were merely glued to the under sides of the beam flanges. . We placed two nose ribs in every standard or box rib division, which gave us a spacing of about § in. We had to do this to retain the shape of the nose, a very important condition to be sought for, in order to obtain maximum wing efficiency. The nose ribs were of spruce ply similar to the webs of the standard rib, but only 3/8" deep and with no rtange. They were glued and stapled to the beam and leading edge. The leading edge was made of 1/16 spruce 3" wide, curved around the nose and with grain running longitudinally. This gave optimum stiffness with light weight and something to which the nose ribs could be easily fastened. The trailing edge wag made of wire that was soldered to copper ferrules at the end of eadh rib. This was easy and also conventional. The entire wing was attached to the fuselage at four points. The center fittings which did this are as sketched: li ay ao me me ) ” .Strop Continues atin aye braved to sep OF a ich Stoel Bolt wage © ocr ta Fie Fy i 4H hold of seamless Tg ~~ Forelege Tuna braved to w Center Wing Fittma. Caveat) Web . Cover Plate , tube Ae St. tube and Strap are Brazed Togethe ™~ Beam Skids were put on the ends of the wings for the sake of protection and to serve as handles in ground work and launching. The pulleys for the controls were standard Jn. 4 parts. The covering consisted of 90# sheeting. Two thicknesses were put on the nose in order to preserve the section shape. The whole received four or five saute of lacquer and two coats of varnish. 12 Fnote.J - Center wing Section: —~/9 23, Photo. li Center Wing Section Bi, CYT ~ 1023 12 @) | Phoro. /5. ArTtmg Wing on Fuselage. The ailerons can best be show by the photo accompanying: Photo 1G - Aileron. laa, * Harn Rib, C.J. 7 ~1923 The horns and horn-ribs were cantilever in design so that there would be no wires to the trailing edge. The horn rib or main rib was as shown below: i: ° a“ \ _ 3- Ply Web — ™ Ply Veneer ————" All wood parts glued and bradoled . Sire Blakes Sprvee The leading edge was covered by aluminum sheet to prevent atr leakeage. 13 Our fuselage was rectangular in cross section with the bottom longerons running in one piece from tail to nose. The upper longerons ran from the tail piece to the rear fittings, which held the wing. We both assumed or computed our loads and then stressed our Fuselage members for both symmetrical and torsional leadings. We found that our interstitial bracing carried no load, theoretically, if the load was considered concentrated on the tail piece. The longerons were then in straight tension or compression, and we computed their sizes én the assumption of pin joints be- tween bracing diagonals. We assumed reasonable sizes for the diagonal and incidence bracing, and we put these in such @ direction that they would be in tension on landing and compression in flight. We did this because we were sure that landing in rough country would produce greater loads than those of ordinary flight. Here should come the explanation of our bracing system. It is seen that we used no wires, put a straight wooden truss with permanent jointe. This method we found would be much easier to make and would, after onee done, obviate the necessity for continual tuning up of wires and inspection of turnbuockles, joints, eta. Against this, of course, was the greater difficulty of re- air in case of smash-up, but we chose the fixed joints for he above reasons. These joints were accomplished by means of gusset plates made of plywood. We computed the sizes of the gussets upon a glue strength of only 1000#/sq.in. which may actually attain 2500¢/sq.in. In addition to this, we put a small screw at the end of the plywood along each member of the truss to prevent the glue joint fracture, in case there were one, frem spreading, which would be aggravated by infinitesimal flexures in the whole system. This method of construction demanded great accuracy of setting up. We succeeded fairly well, the fuselage warping only an almost inconsiderable amount. The tail portion ended in a box like construction of plywood. This produced rigidity to torsion and gave us a desirable foundation for the assembling of the tail group. It will be noticed that on the fuselage the sides are out in under the wing. We did this to allow the air un- obstructed passage after it passed the pilot's head. The struts, then, were exposed a short length and the incidence diagonal on that bay could not run to the wing fitting. This demanded that we put in some bracing between this point and the wing to make the system complete. We worked out a yracing of small tubing which satisfied both the external fuselage shape and strength requirements. The exposed atruts were streamlined with aluminum. The wing fittings are shown on the photos. 14 CLT. - 923 a Photo. |8. Fuselage “ Photo, 20 - Showrng Fuselage under wing . Photo .19. Skeleton Fuselage ¢ Center Wing Section . — i Photo. 2). Skeleton Fuselage 4 Center Wing Seation. ‘. ; s Fire 42 $933 C27" -1923 l*+ (a) The tail group wae made so that it could all be taken off with a minimum of effort. The fin could be slipped off, and of course, with it the rudder. The stabilizer and _ elevator were more difficult and will be taken up in the discussion of that group. The landing gear was ultimately designed for wheels. We were obliged to use Smith Motor wheels and tires (20x2) vecause of the doubt as to the strength of bicycle wheels. The smallest airplane wheels that we could get were much too heavy and large. We wanted to encase our shock absorber inside the fuselage in order to eliminate parasitia drag. We also wished to avoid a cross-axle, which might cause spills in brushy or rocky country. Last but not lease was the fact to be contended with that we had to adopt our design to the already fixed fuselage trussing instead of doing the reverse as would te normal. All in all, the design, as arrived at, was one with which we were all pleased. It was all constructed of seamless tubing, without any special heat treatment and the joints and ends of the members were all of brazed construction. The two large side struts were computed for size on a basis of both pehding and compression and the entire structure was designed to take an impact load of 3, which corresponds to a pancake drop of 18 inches. It is interesting to note that in flight the plane was stalled and pancaked purposely to test the landing gear and that one axle point, on one side only, was bent but slightly. This was bent back and the whole is still as it was originally. The shock absorber was made with 3 wrappings of 1/2" dia. rubber oord on each side. These will deflect about 4 in. displacement under full load. We also put a oheck in the shape of a web strap in oase of overtravel or breakage of the rubber. The rubber is wrapped around a cross tube at the bottom of the fuselage and the corss tube of the landing gear, the latter being higher. This cross tube of the landing gear ia allowéd to play up and down in the groove formed by the fuselage struts. Excessive side play is prevented by the rubber itself. We do not need to fear breakage of struts in case of a side swipe, because the wheels would collapse first, being only 3" across the hub. The wheels were streamlined with cloth tO make less drag, the cloth being olinched by the tire and doped and varnished to tightness. The whole gear was made adjustable, fore and aft, by having the braces running to the front, of two parts, telescoping into each other. Two pine were put through them to fix their positions. 15 Photo. 22. Landing Gear. The top of the fuselage in front of an/in back of the wing was made open to inspection by means of two detachable turtle backs. Photo 23 Turtle Back 16 The tail skid was a conventional hickroy one with a shook absorber unit. The skid support was designed so that the skid would turn automatically in taxying. Glimpses of this feature may be obtained from other photo- graphs. The seat was a standard one, equipped with life straps and arranged so as to not give the pilot undue discomfort. It was made adjustable with the controls able to be moved through a distance, fore and aft, of 8". The control system was of the joy-stick-rudder bar type and can be easily sketched as follows: &—~ Te Ailerans Nate: yo wre ‘When stick 1s moved perpendicular to paper Torque Arm brazed te main tobe. Torque Arm moves with it te predvce movemert ol Ailerons , Fricheon Cpa vf wires 5 af aT ——— vTe Elevater hovas. \ BES: 7 Universal Saint / fh ley 3. The launching device is one of which we are °%: cq. The plane is launched, that is, for it's preliminary .....18, by means of a 300 ft. cable attached to an automobile. The pilot is able to free thé ship from the cable at will, by merely pulling & string inside the cookpit. The device on the nose is such that when the pilot pulls this string the tension of the cable opens the jaws of the device, which hold the end of the vable, affording an absolutely sure re- lease, impossible of catching. The fuselage and turtle backs were covered, doped and varnished similar to the wing. 17 The fin was built to be easily removable and the rudder was attached to it by means of three hinges. The group was of cantilever construction. The horn-rib of the rudder was similar to that of the ailerons. The section was @ symmetri- cally streamline one we devised ourselves and we conformed the rudder section to certain new principles. It was found, from governmental testg, that if the rudder was a continuation of the fin section or/even smaller thickness, there was a certain dead area between whose limits the rudder, when moved, was ineffective. This could be eliminated by making the maximum thickness of the rudder larger by 1/4 than the trailing edge of the fin. They were built to stand 24#/sq.ft. of load. To eliminate dead area make fiomes thickness ot rudder ter than thickness of Fin at hinge. ” nge Photo 25 showing Fin, rudder Q stabilizers. Photo 26 showing Elevator. Photo 27, Locating c.f g> (see p20.) by weighing in two positions, (Showing Asse mbly et Elevator X stabilia Ge p- 19) Strap hinges- Ad justoble feature. 13 @) The stabilizers and elevator were of unusual construction. We wished to make the stabilizer adjustable, so that the general attitude of the ship could be changed when desired. he seotion chosen, mainly so that we could get the required thickness to suit the 14" beam, was that of a wing (Sloane #1) and we inverted it so that it would be most efficient when a down pressure was to be developed. A ship rarely flies with an up load on the tail. The total load necessary in a dive at limiting velooity was figured from center of pressure movements to he 3007 downward. We doubled this to 6004 to give a factor of safety of 2, thus giving a loading of 23#/aq.fts We would not make this cantilever in construction and so we had two tubes as braces on each side, one running from the front beam of the stabilizer to the middle post of the fin and the other running from the main beam to the rear post of the fin. The main beam of the combination was to act (1) as the trailing edge of the stabilizer, (2) as the leading edge of the a@levator, and (3) as the hinge rod or center. It was a large tube, 14" dia.°x 22 ga., with fittings brazed to it, at the center, to make it attachable to the tail piece of the fuselage. The stabilizers and elevator were attached by means of strap hinges. A method of trial computations was the only way possible to compute the size of the beams because they were subjected to both bending and compression. This: was done by two ways, the Berry and the Perry methods, each of which checked the other. The front beam which was a tube 4" x 20 ga. was a hard problem to solve. The only method we could see, to accomodate it to the limitations of detachability and adjustdbility, was to make it so that it could be pulled right out from througk.the fuselage. The side bracing tubes were each made with/tufnbuckle in them to vary their effective length. The fittings which held this beam at the sides of the fuselage, were attached to arms,one on each side, which could be ad- justed up or down to any position desired. These arms, in conjunction with the tubes having turnbuckles, made the whole stabllizer adjustable. The horn-ribs of the elevator were cantilever just as the rudder and aileron horn-ribs. 19 When we had completed the whole plane it was then necessary to balance it and to fix it's flying attitude. It will be remembered that we had three adjustable features, namely, the seat, the stabilizer, and the landing gear. The senter of gravity, according to praatice, must be at about .28 or .30 of the chord from the leading edge when the plane is in normal flying position. We had, at the peginning of our computations, assumed weights and dimensions, the resultant of which would approximate this locus of the center of gravity. To determine it's exact position, we fixed the seat and wheels temporarily, and weighed the whole ship, including the pilot, in two positions, one, tail low, the other, tail high, by means of three scales, one under each wheel and the third under a point of the tail skid support. The day was comparatively frea from wind and the weighings checked in totals, missing only by 1/3 of one percent. By computing the resultant of the weight for each weighing, we obtained the existing center of gravity and found that it was off from the desired by 1 3/4". To obtain this desired locus required that we move our seat as far forward as it wae possible. This shows that we reckoned wisely in making the seat adjustable. After we had found the center of gravity, as it now exists, we were then able to definitely locate our wheels. We fixed them in such a position, that the line through the point of ground contact and the center of gravity would make an angle of 28 degrees with the vertical when resting on the ground level. This angle has been found to be the desirable one for the making of an easy three point landing. The final adjustment to make was that of the stabilizer. From our wind tunnel teste we found that the setting of the tail stabilizer should be at -1.3 degrees to the wing to give a normal flying attitude for best soaring at 4 degrees incidence. This we did. Having made our adjustments and finished wiring the controls, we were all ready to fly. The ship was taken to Ross Field, Arcadia, where there is a field with a length of one-half mile. The method adopted to test the ship was to hitch her to an aatomobile with a 300 ft. cable. The first trial consisted in merely taxying over the ground, Mr. Merrill wisely limiting the speed of the automobile to 25 mi./nr. This gave the pilot a chance to test the tail surfaces and to prove the strength of the landing sear. Without much ado, the next trial found the plane in the air and thereafter many flights were made, with a view to testing the maneuverability of the ship as a whole. Newspaper and motion picture photographers took advantage of these flights to ply their trade. 20 A few views of the firs/ fights RQXa) The general consensus of opinion is that the plane ie successful, inasmuch as it is the first work of two ngineering neophytes. The pilot gave as his opinion, (he having never done any gliding before without motive power) that the controle were slower than for a power machine. The pitching and rolling controle were safe, although slow. The real difficulty occurred in yawing. The rudder was totally inadequate. It is a coincident fact, that the rudders of all gliders subsequent to the Hanover Vampyr, in Germany, have deen made greatly larger. One cause of our difficulty may have been due to the fact that the inertia of the wing tended to keep swinging the plane around notwithstanding the force of the small rudder to bring it back to norm. The two students are well satisfied with their work as a whole, but oan see that the need for large efficient control surfaces is paramount in any glider design. The reason for this is that the absence of propellor slip stream reduces greatly the (according to the ratio of the squares of the velocities) forces on the rudder and elevator. he ailerons are not affected in this regard but in another, equally potent. The wing of a glider is one of the large, perhaps the largest, weights of a ship and the span is conoomitantly large. The centroid of each side of the wing is moved outward according to the above facts and hence the wing requires greater rolling moment to produce accelerations in rolling equal to those in power flight. Therefore, it is seen that all control surfaces should be considerably larger for sailplanes in general than for power machines. A new scheme for control in yaw has since been devised by Mr. Merrill, and eonsists of rudders placed at the tips of the wing, the old fin and rudder being re- moved altogether. The same effect is accomplished with these by turning the rudder on the side of the ship in the direction it is wished to turn. In other words, a drag is introduced on one side only, thus slowing down that side, to produce the yaw. This also helps the aileron action in that i$ drops the side whose rudder is turned. This method of control in yaw is very effective and has, I think, advantages over the conventional type. This has been tried on our glider and the pilot reports 300% better yaw control. The monoplane has, in our opinion, proved itself to be the most efficient type in soaring practice. The elements to be strived for to produce efficiency are (1) good wing sections, (2) light weight construction and extreme refinements in streamlining and aerodynamic improvements to reduce, to the lowest possible, the parasitic drag. el The argument is specious, (I think it can be proved so) that is sometimes heard, that because of the slow speeds encountered in soaring, the parasitic drag is non-important. It mist be remembered that all the forces, including lift, vary according to the same law e.of velocity squared. Hence, since relatively, the drag depreciates with the velocity and the lift depreciates in the same proportion, it is seen that any diminution of the arasitia drag coefficient will raise the L/D and L fo Tatios. We have now built one machine successfully, and I fully believe that if we were to build another one now that we would build the new one better and more efficient than the first, although we have gained a lot of satisfaction from the sucoess of our first attempt. It is to de hoped that we may see this glider attempt soaring flight over a suitable terrain some day. But at the present time the pilot and Mr. Merrill have many research problems they wish to solve first. A gliding meet in Oakland, California ie contem- plated for next summer, and 1f it actually occurs, it is our wish that we may enter this plane. 22