The Aircraft Engineer 23 Feb 1928

Embed Size (px)

Text of The Aircraft Engineer 23 Feb 1928

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928


    February 23, 1928 Supplement to FLIGHT


    February 23, 1928CONTENTS

    PAGEMetal Construction Development. By H. J. Pollard, Wh.Ex. ,A.F.R.Ae.S '. 13

    Seaplane Stabili ty Calculations 17Technical Literature ... .. . ... ... ... ... .. . 20

    EDITORIAL VIEWSIn this issue, Mr.Pollard carries out some interesting calcu-lations relating to metal fuselage structures of " strip " andsolid-drawn tubes. He arrives at the conclusion that, inthe example taken, the weight of the solid drawn tube-struc-ture, is, exclusive of weight of longeron fittings, 18 per cent.greater than the weight of the " strip" structure. Asregards the fittings, he expresses the view that those for the" strip " fuselage will be lighter than some forms of jointsused in tubular construction, but concedes t ha t, in view of therecent production of some very light, if costly, joints forsolid-drawn tubes, the weight of fittings may be taken asequal for the two types. Thewelded frame, hestates, " wouldshow up very badly indeed beside these two cases if mildsteel tube was the material used."As Mr. Pollard's article is by way of being a challenge tothe advocates of certain other forms of construction, perhapswe may hear from readers with experience of these. Forinstance, it would be interesting to know how the new Sop-with-Sigrist form of flat-sided tube construction, developedby the Hawker Company, compares with Mr. Pollard's" strip " construction. Certainly, the Hawker fittings areabout as simple as anything could be, and they could notby any means be called costly. While on the subject ofMr. Pollard's article, we should like to congratulate Mr.Miles,also of the Bristol Company, on his drawings of the " strip "structure. Mr. Miles has introduced a new style of sketchingwhich, we think, lends itself admirably to showing technicaldetails.Mr. Munro is a newcomer to the pages of THE AIRCRAFT

    ENGINEER. He is, we understand, employed by the GlosterAircraft Company, under Mr. Folland, and his article on" Seaplane Stability Calculations " should be of very greatassistance to those who, having had no previous experienceof such work, are suddenly called upon, as they may well bein these days of seaplane progress, to tackle it. Unfortunately,lack of space has compelled us to divide Mr. Munro's articleinto two instalments, but this was unavoidable. The con-cluding instalment will be published next month.


    (Continued from Page 3.)Before amplifying some of the statements made in theprevious article, we will study a simple feature of strip metalconstruction and demonstrate its advantages. In doing this ,one or two of the principles governing economic structuraldesign will appear, and later some observations on the methodof manufacture will be made.In Fig. 1 is shown a side view of a frame which might be aportion of a fuselage tail. Fig. 2 is a view in perspective ofthe structure, and Figs. 3 and 4 alternative nodal points.The bulkhead bracing has been omitted from Fig. 2 for thesake of clearness.From these illustrations the details of the construction arequite clear, and no elaborate description is necessary.For such a structure to be light, safe and rigid, two veryimportant conditions must be fulfilled, and in certain specialcases there is an equally important third condition. Thefirst is that the built-up longitudinals must be continuousthroughout their lengths. The best results cannot beobtainedif the smaller of the two strips is cut away at intervals so thatangular fittings may be secured to the flats of the largersection, because this would introduce a series of sections ofdiscontinuity along the longerons with consequent s ubstan tialreduction of strength at these points. The second construc-tional feature to be observed is the method of securing thebracing members to the gusset plates.These members consist of two similar sections rivetedtogether along their edges, forming a circular or approximatelycircular sectioned member, having two diametrically oppositeoutwardly extending flanges. It might appear safe to cut offone of the component sections level with the outer edge ofeach of the gusset plates, forming a junction, as shown inFig. 5. The only object in doing sowould be to save a littleweight, but here, again, the necessity for continuity makes itimperative that the strut ends be divided, a section passingeither side the gusset. Two other advantages are derivedfrom this, one being ex act centroidal loading of the member,and the other tha t the securing components are put in doubleshear, thus making it possible to effect an appreciable savingin assembly time due to the use of fewer rivets. The thirdcondition is only of importance when the struts are " short,"that is, when they are subjected to considerable intensities ofstress. The load is transferred to the main section of thestruts through the narrow riveting edges, and these edges inconsequence are subjected to a stress much in excess of the


  • 7/27/2019 The Aircraft Engineer 23 Feb 1928




    Fig. 1.average P/A for the section ; this stress round the end rivetsmay exceed the compressive yield stress of the material,causing crinkling of the flats and premature end buckling ofthe whole section.

    It might be possible to calculate the load at which the strutends would fail if the direct forces only had to be considered,but owing to flexing of the compression boom and the changein shape of the frame bays due to the displacement of thepanel points under load, a very complex stress system is setup round the rivets common to the bracing struts and gussetplates. The computation of this stress is not possiblemathematically, but tests of a rather simple nature can easilybe devised from which data can be obtained as to the endreinforcement necessary, so that the end of the struts maycarry their loads up to the point of central failure by buckling.The " fixing" couples at the strut ends are probably ofconsiderable magnitude ; the end load effect on the compres-sion boom is to produce a condition as shown in Fig. 6, which,as stated, is resisted by the nature of the end connections ofthe bracing. It is seen, therefore, that a much greaterradius of gyration is required in a strut about an axis at

    right angles to the line joining the riveting edges, than aboutthe other axis of symmetry. Instead of the edges being"waste metal," as is sometimes alleged, they play a reallylarge part in giving strength and rigidity to the frame, andapart from difficulties of riveting, if the edges are narroweddown excessively, it will be found on test that the strutswill fail in the plane of the frame due to the above-mentionedcauses.In Fig. 7 is shown a simple method of counteracting thetendency to local end buckling. (Also in this figure is shownthe socket attachment used for connecting one length oflongeron to another length.)Two short lengths of section wrapped round the strut endsand continued above the gusset a short distance are sufficientto distribute the load evenly across the section of the st ru t;these reinforcements need securing only at the riveting edges,and not separately by rivets to the main body of the section.In cases of very high stress intensities, additional reinforcingmay be made by means of a narrow strip the width of eachriveting edge running the length of the strut, the thicknessof which is equal to the thickness of the gusset plate. This


    Member. Length.L. Area.A.Radius ofGyration.K. L /K P/A=p. P .

    ActualLoad inMember Description of Member.


    I0050-050050 05


    550"]1,280 I1,875 f2,26OJSection as shown in Fig.(0-009 ins. thick, S. 40).







    55511,2941,900 y2,280I2,560J

    Section as shown in Fig.(0 009 ins. thick, S.40).

    VerticalStruts20-625-030-83 6 0

    0 0220-0220-02570-0257





    Section as shown in Fig. 9(0-006 ins. thick, S.40).Section as shown in Fig. 9(0-007 ins. thick, S.40).


    23 033-541044-047-5

    0 03100310 0310-0310-031

    0-40- 40-40-40-4

    57 '83 '




    900"]930760 V530435JSection as shown in Fig. 10(0-006 ins. thick, S. 40).


  • 7/27/2019 The Aircraft Engineer 23 Feb 1928


    l\ . ,1.


    u * ,\ Walso obviates the hitherto wires have been used, but from the experience gainednot only lends stiffness to the free edge but also obviate^he n be ^ ^ w i r e g ^ ^Q o tnecessity for "joggling "the edges where the stru^leaves to date t ^ P ^ from ^ ^ ^ ^ a n d s t r u t 8 o n l y u s e dthe gusset. None of these ^ ^ ^ 3 ^ Si their place. There are several things that could be arguedordinary fuselage construction, * f . ^ j S ^ A infavou? ofsuch astructure, probably the most importantSoft S3S iS5 SSJ is lwnrin%-. 3; point being th e freedom of the rigid members from initia l

    1 1 6 c

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928



    To pLongerons.








    2 3!

    Length. ! Area.L .

    2633312914-226-333-431-329-320-62 5 030-83 6 023 033-542-04 4 047 5

    A .

    0-0570-0570-0570-0570-0570 0570-0570-0570-0570-0290 0350 0350 035










    P .




    ActualLoad inMember.

    55011,280 11,875 f2,260j55511,2941,900 }2,2852 ,560 j70 0565"]455 >385J9 0 0 ]930760 y530435J

    FEBRUARY ?3, 1928


    Description of Member.

    lj in . o/dia. x 28 S.W.G. (T.5).

    l i in. o/dia. X 28 S.W.G. (T.5).

    f i n. o/dia, X 28 S.W.G. (T.5).| in. o/dia. X 28 S.W.G. (T.5).

    4 B.A. tie rods.

    stresses ; apa rt from military aircraft, where members areliable to damage in action, there is no need for bulkheadbracing at all, since it is found experim entally, and by calcula-tion, that such bracing does not affect the strength or rigidityof the structure. A panel point having no bulkhead bracingis shown in Fig. 4.A simple comparative weight and strength estimate will bemade of a structure as described, and a similar frame builtfrom T.5 tube and wires.The dimensions of the uni-planar structure are given inFig. 1. It is assumed tha t a load of 800 lbs. is suspendedfrom 0 and the flat frame is held at X X. Figs. 8, 9 and 10are sections of longerons, ties and struts made from steel stripto Specification S. 40. These have been designed to supportthe loads given in column 8.The sizes of the struts have been derived from the appro-priate curve, as shown in Fig. 11.In Tables I and I I, the full pa rticulars of the " strip " and" tubular " fuselages are given.For the section of Table I marked " Diagonals," the loadhas been reversed ; these members have to act both as

    ties and compression members, and obviously the case toconsider is when these diagonal bracings act as stru ts. Amoment's thought will show that this procedure does notalter the numerical value of the load in the members, butmerely the signs.In Table II it is assumed that these struts are replaced bytwo swaged wires, complete with fork ends and pins. Ineach case, column 1 denotes the me mb er; column 2 its length,L ; column 3 its area, A ; column 4 the radius of gyration, K ;column 5 the ratio, L/K ; column 6 the corresponding valueof stress, P/A, obtained from the graph ; column 7 the loadsfrom columns 6 and 3 ; in column 8 the forces induced by theapplied load ; and in column 9 a description of the member isgiven.A comparison of the figures in column 7 in the tables givesthe relative strengths of the two frames, which, in the worstcases, are approxima tely equal. A simple computation ofthe relative rigidities of these frames is not possible, buttests which have been made show this to be decidedly infavour of the strip construction. From the lengths and areasof members given, the weight of each is q uickly derived ;

    Figs. 8, 9 and 10.U6d

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928


    \ v;







    a. 70.000

    60.000Sizi~ 50000










    > = ;



    -^ .c


    TH E








    On heshs Prom buiir up sh -u fhas been Found rtiar* rhe reswhe n plotted . l ie slightl y abcc u r v A


    > M a

    s i^ulfsve

    " ^ ^

    = ^


    * ^s








    0 IC 20 30 40 SO 60 70 80 90 100 110 120 130 140 150Fig . 11 .

    comparison, therefore, may need revising as experience withthese higher tensile tubes is obtained. Advances are, however,to be expected in the design and methods of manufactureof components made from steel strip.It is not suggested tha t the whole weight of 800 lbs. could betaken locally on the strip longeron section, but the sameremark applies to the solid drawn tube. Provision for restingon trestles, lifting, etc., is easily made, and a fitting andmethod of attachment suitable for this is shown in Fig. 12.The above comparison is presented in as simple a wayas possible. At the same time, the overall dimensions andexternally-applied loads are such as might apply to a portionof the structure of an aeroplane of 4,500 lbs. gross weight orthereabout. If the investigation is pursued further, it willstill be found to favour the strip construction, particularlyin the matter of fittings for the attachment of equipment,control surfaces, cable guides, et c. ; the numerous " freeedges " obviously lend themselves to this purpose. One suchtype of fixing is shown in Pig. 13. This is a tail plane sparattachment.Space does not permit of further illustration or descriptionof fittings, but in general, a simple bent or flat plate is allthat is necessary ; there is a sharp contrast between this andthe machined fittings or clips with bolts that are common totubular construction.While the writer believes that aircraft frames as describedhave only been built by the Bristol A eroplane Co., yet descrip-tions and drawings of the various component sections haveappeared from time to tim e; for instance, particulars ofbracings made from two similar semi-circular channelsjoined together along their edges were advocated for aircraftmore than 30 years ago ; similarly, drawings of longeronsmade from two parts shaped approximately as illustratedabove have been published fairly recently, but such longeronshave been shown discontinuous along their lengths, and itmay be that this lack of continuity has been the reason forthe abandonment of the m ethod. Only one aspect of thisconstruction has been dealt with : it may be possible in thefuture to describe further developments along these lines.

    allowance must be made for tube sockets, pins, fork ends,rivets, etc., exclusive of longeron fittings, the percentageincrease in weight of the wired over the strip frame is foundto be 18 per cent. There is also the weight of fittings toconsider. The gussets would be 24 G., with suitably-shapedlightening holes. These would certainly be lighter thansome forms of joint used in tubular construction, but asrecently several very light, if costly joints for solid drawntube work have been designed, it may be assumed that theweight of fittings is equal in each type of structure.The above is a fairly complete weight comparison of twometho ds of steel construc tion. A welded frame would show upvery bad ly indeed beside these two cases if M.S. tube, the strutcurve of which is shown on the cha rt, was the m aterial used.Molybdenum or manganese steels would show up better, butit has been admitted tha t where tubes have been used, notablyin America, finished structures are on the heavy side. This isprobably due to the fact that it is not considered safe tojoin tubes by welding where the wall thickness is less than22 G. It should be noted th at if the material of the gussetsis distributed over all the corrugated members, the thicknessof the m aterial would only be raised one-and-a-half thousand thof an inch. This fact should give the welding enthusiastfood for thought.To further this comparison, it should be stated that thesections shown in Figs. 8 to 10 are practical propositions,although it would be wrong to give the impression that,without some experience on the part of the producer, suchsections could be readily made. The question of the assemblyof these members will be dealt with in a later article. A morefavourable case could have been made out for the tubularStructure if a larger diameter and thinner gauge of T. 5 hadbeen taken, but comparison with a tube outside the practicalcommercial range is useless. Tubes are now being offeredto the aircraft industry of quality superior to T. 5, and theseare said to be quite suitable for structural work : the above


    In the design of seaplanes it is quite as necessary to deter-mine by calculation the statical stability of the machine onthe water, as in th e case of ocean-going vessels, and the calcu-lations involved are solved along very similar lines.The statical stability is defined as the tendency the sea-plane has to return to the upright when inclined from thatposition, say by wind or waves.This stability is measured by a comparison of the " meta-oentric he ig ht " calculated for any given machine with themetacentric height of similar craft known to be successful,and is very largely a matter of experience and tabulated dataIt is proposed to outline the method adopted.Fig. 1. Shows the machine inclined at a small angle, andindicates the two equal forces acting.(i) Wt. acting down vertically through the C.G.(ii) Buoyancy acting vertically u p throug h the newcentre of buoyancy ; th at is, the C.B. with machinetilted.When the machine is tilted, the total displacement remainsthe same, but the shapeof the underwater surfaces changes,so th at the centre of buoyancewhich is the centre of gravityof the underwater volumealso changes from B to H vThe point M where the vertical through B x cuts the centreline of machine, is termed the transverse metacentre.

    If, now, a line is drawn GZ perpendicular to the verticalthrough Bj, then the equal forces (i) and (ii) act at a distanceGZ from each other, and the moment tending to right theseaplane is WxGZ.As the point M is generally assumed to remain constantfor small angles of heelto about 8we can substitute forGZ and say tha t the righting moment, or moment of staticalstability is W X GM sin 6.

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928






    G_ ^ _




    Fig. 1.If the point Mshould be below G the resulting couplewould tend to overturn the boat. This is the case inmostflyingboats, and the volume of thewing-tip floats iscalculatedto overcome this tendency to roll over.A glance at Fig. 1will show that a large amount of meta-centric height i.e., the distance G.M.means a quick" snatch-back" to the upright position. It is because ofthis that naval architects design for a minimum G.M., com-patible with safety, to eliminate unnecessary discomfort to

    passengers and a minimum of "wracking " of the structure.An important point on seaplanes is thenecessity of havingample propeller clearance from the bow-wave system, whichmay easily wreck even a metal airscrew.The point M shown inFig. 1 istermed thetransverse meta-centre, and is governed by the shape of the floats and theirdistance apart; its distance from G is used as a measureof theseaplane's stability when rolling from side to side.W hen considering the longitudinal movement of themachine in the water, we have another point to considerwhich is termed the longitudinal metacentre. See Fig. 2.The transverse metacentre is the more important, but tofind either position wemust first determine theposition of thecentre of buoyancy and then show how the distance B.M. isfixed.The following calculations arenecessary :1. Total displacement of thefloat.2. Area of load water plane.3 . Centre of flotation.4. Moment of inertia of the load water plane.5. Displacement to load water line.6. Position of centre of buoyancy.Fo r thesake of clearness wewill take each separately.

    The designed shape of the floats for the purpose of thisarticle is assumed to be already determined, and the methodof checking out thestatical stability only is being considered.The offsets are given in table below for the assumed floatdesigned and the calculations throughout are based on thisshape.Station.

    121 H1 1109876Top ofStep54321I0

    BelowKeel.5-3517-121-427-1829-530-2530 0629-828-726-023-320-617-915-314-012-6

    DatumDeck.5-352 -81-220-050 - 00 - 00 - 00 - 00 - 00 - 00 - 00 - 00 -00 - 00 - 00 -0


    ChineHalf-Brdths.9 112-315-7817-117-3817-216-8616-8616-415-7714-412-38-65-56

    RadiusofDeck.9-3811-9815-4717-0817-3817-1916-8616-8616-2915-0213-4510-686-754 0 7

    I t is assumed also that the flotation system is the normaltwin-float type with rounded deck and vee-bottom, notfitted with hydrovanes or stern stabiliser.Taking calculation 1, wewill refer toFig. 3.The volume shown shaded below the load waterline repre-sents the displacement of the float, and is equal to one-half2 427the weight of machine ; in this case 2,427 lbs. or 'cubic feet.The volume of the float above the waterline representsthe " reserve of buoyancy." This may be taken as 90 percent, minimum of the " displacement."Therefore, wehave :Weight of machine = 2,427 X 2 = 4,854 bs.

    Fig. 2.116/

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928


    FEBRUARY 23, 1928 19


    Figs. 3 and 4.Reserve buo yancy = 90 per cent, of 4,854 lbs. = 4,368 lbs.Total volume required = 4,854 + 4,368= 9,222 lbs.As this is taken on two floats we get:9,222Total volume required for one float = x = 4,611 lbs.

    The length L of the float is divided into 12 equal stations,as shown in Fig. 4.By using the planimeter over the body-plan (Fig. 5),the area of each half-section is quickly found, and the resultsare ta bu lated as below :

    in the load water line on each station. These positions aretransferred from Fig. 3, e.g., AB on Fig. 3 is equal to ABon Fig. 5, and CD on Fig. 3 is equal to CD on Fig. 5.The half-breadths of these waterlines are set out as shownin Fig. 6, and the area of the load "water plane found bytabulating as below :


    1 1 *111098765432140

    2Area of Half-section in sq. ft.0 - 00-571-162 1 02-682-732-712-702-161-761-410-990-630-370 - 0


    4Functionsof Areas.0 - 01 1 41-748-405-3610-825-4210-804-327-042-823-960-940-740 - 0

    1ition.12" I1110987654321


    2Semi-Ord.(in ins.).0 - 06 - 012-01 6 116-817-017-016-01 5 013-511-58 -55 - 54 -00-0



    4Functionsof Areas.0 01 2 018-064-433-668-034-064-03 0 064-023-03 4 08-258 -00 - 0


    Total volume of float = 63-5 X63-50

    X j - x 6 4 = 4,630 lb s.For each station the area in column 2 is Tmultiplied bythe multiplier given in column 3, the result being placedin column 4 under Fun ctions of Areas. The Function s of Areasare totalled u p, multiplied by one-third of the interval betweenstations in feet, and this result multiplied by two, to includeboth sides of the float. Multiplying further by 64 bringsthe result to lbs.This checks the size of the float and, as already shown,

    should, in the case taken, be at least 4,611 lbs.As Simpson's Multipliers are universally used, it is notconsidered necessary here to do more than show the methodof their application in each instance.2. Area of Load Water Plane.On the body plan (Fig. 5), which represents sections ateach of the stations numbered upon Fig. 3, we now draw


    STN.II .,STN. 10 ^ JfTNS.619 - - ^ f -

    STN7 U DSTN.8 J /It;




    ^. OF FLOAT

    STATION \5 TN. I

    V \ \ \ \,

    Fig. 5.

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928



    Fig. 6.461-25 19-32 2 1 , aArea = ; X s X 7 X i~7J = 41-25 sq. ft.The area of the load water plane is found by total offunctions of areas multiplied by one-third the intervalbetween stations, and this result multiplied by two for bothsides of the float.As the ordinates used are measured in inches we thendivide by 144 to bring the result to square feet.Finding the position of the load waterline is a matter ofjudgment and " trial and error." The first considerationis water-clearance for the airscrew, which fixes the positionof the floats below the fuselage. The trim fore and aftdepends upon the nature of the machine.

    3 . Centre of Flotation.For transverse inclinations the centre of flotation, whichis defined as the centre of gravity of the water plane, is takenas on the centre line of machine, but the position of centreof flotation for inclinations fore and aft must be calculated.As we are dealing first with the transverse metacentricheight we will leave this calculation for the moment and

    pass on to No. 4.4 . Moment on Inertia of Waterplane.This must be found:

    (a) About the C.L. of float, (b) About the C.L. ofmachine.4 (a) is tabulated as follows : .

    Then Io = I + (area of W.P. + \ track2)= 21-45 + f41-25 x




    2Semi-ords.l" in ft.0-00-51-01-341-401-421-421-331-251 1 20-960-710-460-330 0

    3Cube ofSemi-ords.0-00 1251-02-42-742-852-852-351-951-400-880-350-090-0360-0

    4Simpson'sMultiplier.I2l i4242424242 i

    5Functionsof Cubes.0-00-251-509-605-4811-405-709-403-905-601-761-400 1350-0720-056-197

    Moment of inertia about centre line of float56-2 1-725 3=21-54-z

    The moment of inertia about the centre line of float is foundby adding up the functions of cubes, multiplying this by one-third of the interval between stations, and multiplying theresult by two-thirds.Let I = moment of inertia of W.P., about C.L. of floatand let Io = moment of inertia of W.P., about C.L. ofmachine.

    = 856-8.Then Io for both floats = 2 x 856-8 = 1713-6.[T o be concluded.)




    R. & M. No. 1099 (E. 25). (7 pages.) March, 1927. Price4d. net.As a result of a long series of experiments at the RoyalAircraft Establishment, it appears that the power of an engineis a function of the pressure rather than of the density.

    Recently, Mr. Capon* has suggested that the law should bedefined more precisely as a function of pressure to the twothirds and density to the one third power. Other investiga-tions at the R.A.E. have previously been published as R. & M.Nos. 462,t 960J, and 961.The whole question of the variation of engine power withheight has been reviewed, and the experimental results havebeen examined to find confirmation or otherwise of Mr. Capon'ssuggestion.The relative importance of pressure and density in deter-mining the power of an engine appears to vary with height,and different methods of experiment lead to slightly dis-cordant results. The simple pressure law is undoubtedlybetter than the simple density law, and for greater refinementMr. Capon's suggestion should give a very close approximationto the truth.* R. & M. 1 0 8 0 . The R educ t i o n of Performance Tests to the StandardAtmo spher e . B y R. S. Capon .t R. & M.4 6 2 . The V ar i at i o n of Eng ine P o w er w i th He i g h t .B y P in sen tand Ben wick.J R. i M. 9 6 0 . V ar i a t i o n of Engine Power wi th Heigh t .By H. L.Stevens, B . A . E . R. & M. 9 6 1 The V ar i a t i o n of Engine Power wi th H eight.By H. M.Garner and W. G. Jenn ing s , R . A . E .

    These Reports are published by His Majesty's StationeryOffice, London, and may be purchased directly from H.M.Stationery Office at the following addresses : Adastral House,Kingsway, W.C. 2; 28, Abingdon Street, London, S.W.I;York Street, Manchester; 1, St. Andrew's Crescent, Cardiff ;or 120, George Street, Edinburgh; or through any book-seller.AERODYNAMIC INTERFERENCE.

    We very much regret that Mr. Stanley H. Evans has beenso busy recently in other directions that he has not been ableto find the time to complete his second article on " TheProblem of Aerodynamic Interference " in time for it to beincluded in the present issue. We are, however, promisedan interesting instalment for next month.ED.116*

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928


    FEBRUARY 23, 1928

    THE SHORT " CALCUTTA " : On the left, a " close-up " of the beaching chassis. The front hatch, which alsoforms steps, can just be seen. On the right, the port wing engine. Above the nacelle can be seen the crane, usedfor lifting the engines into and out of the machine. In the photograph, the engine has open exhaust. A collectorring, shaped liked the cowl in the photograph, will be fitted later.(Concluded from p. 116)has also been arranged to drive a mechanically-operated bilgep u mp as well as the general purpose dynamo for lighting andradio when the main engines are not running.

    SpecificationAs the general arrangement drawings of the Short" Calcutta " were published in a recent issue of F LI GHT wedo not propose to reprint them here.

    The main dimensions and areas are : Span of upperplane, 93 ft. (28-35 m.) ; span of lower plane, 76 ft.6 in. (23-15 m.) ; total wing area, including ailerons, 1,825sq . ft. (170 ma) ; length, o.a., including servo rudder, 64 ft.9 in. (19-75 m.) ; wing chord, 11 ft. 6 in. (3-51 m.). Area ofailerons (total), 150 sq. ft. (13-95 m2) ; area of tail plane,112sq. ft. (10-4 m.2) ; area of elevators, 105 sq. ft.(9-76 m.2);

    THE SHORT CALCUTTA On the left, a wing in skeleton, and on the right, the hull,the faired rear step.117

    Note particularly

  • 7/27/2019 The Aircraft Engineer 23 Feb 1928


    FEBRU ARY 23, 1928

    area of fin, o6 sq. ft. (5-2 m.2) ; aiea of main iudder, 49 sq.ft.{4-55 m 2 ) , area of servo rudder , 7-6 sq. ft. (0-706 m.B).W eigh t of m achin e em pty , 12,600 lb (5,730 kg.) ; wei ghtfully loaded, 20,200 lb. (9,185 kg .) ; wei ght avai lable for load,7,600 lb. (3,455 kg ). The available load may be composedas follows: A crew of three , wit h baggage, food and wa ter,768 lb. (319 kg.), and 320 gallons of pet rol and 30 gallons ofoil, 2,730 lb. (1,241 kg.). Wirele ss, electrical equip me nt, in stru -ments, f ire extinguishers, cooking and marine equipmentaccount for a weight of 562 lb. (255 kg.), leaving a pay loadof 3,540 1b. (1,610 kg.), which is equivalent to 15 passengerswit h baggage, food and wate r (at 236 lb. per head ). The fueland oil capacity given does not represent the maximum, asthe tank s have been designed to hold 480 gallons of petrol and45 gallons of oil, so that by sacrificing a certain amount of payload the range can be correspondingly increased. W ith thequantities mentioned, the range is 5i hours, or 500 miles(805 km.), and with full tanks and a smaller pay load theduratio n is 8-2 hou rs, and the range 740 miles (1,190 k m ).The wing loading is 11-05 lb./sq. ft. (54 kg./m. 2) , and thepower loading (at full power) 12-8 lb./h.p. (5-83 kg./h.p.).

    P e r f o r m a n c eAlthough the official performance tests of the " Calcutta "have not yet been carried out, it may be of interest to giveth e estimated performances. The top speed at sea level is120 m.p.h. (193 km./h.), and the cruising speed 100 m.p.h.(161 km ./h.) . The landing speed is 57-5 m.p.h. (93 km ./h.) .Ra te of climb at ground level is 800 ft. /min. (244 m./m in.).The service ceiling is 10,000 ft. (3,050 m ) . The " W ingPowe r " is ? 0-863 h.p. per sq. f t. = 9-26 h.p./ mA Asthe to p speed is 193 km. /h. , the Everting " High-speed Figure "(metric) is 14, which is an extremely good value for a three-engined flying-boat. The Everting " Distance Figure " a ttop speed is 4-2, which is also a high value for a machine ofthis typ e, As this refers to the top speed, it is not, of course,an optimum value, but as we have no information relatingto the power at which the machine cruises most economically,it is not possible to give the maximum value of the " DistanceFig ure." Th at it is well above the average seems more tha nprobable.

    R O Y A L A E R O C L U BA MEETI NG of the Committee of the Royal Aero Club and theManagement Committee of the Society of British AircraftConstructors was held on February 15, 1928, to consider thequestion of handicapping formula.King' s Cup. It was decided that the handicap pingshould not be on formula but on known performances.Ae r ia l De r by .It was decided to hold the Aerial Derbythis year on a course round London. Also an Aerial DerbyHan dicap on formula. The S.B.A.C. put forward a formulawhich it was agreed to submit to an independent authorityfor his report.Off ices: THE ROYAL AERO CLUB,3, C LIFF OR D STR E ET, LONDO N, W. 1 .H . E . PER R IN, Secre ta ry .

    a s