FlightGlobal.com
Home
Premium
Archive
Video
Images
Forum
Atlas
Blogs
Jobs
Shop
RSS
Email Newsletters
You are in:
Home
Aviation History
1933
1933 - 1322.PDF
ss SUPPLEMENT TO FLIGHT DECEMBER 28, 1933 THE AIRCRAFT ENGINEER rivets themselves are comparatively expensive. This is not due to any technical difficulty in the manufacture of the rivets, but arises from the fact that the De Bergue Company has chosen this way of collecting royal ties on the riveting process. The cost, they state, corresponds to a royalty of approximately 4d. per gallon of tank capacity, in view of the fact that one man can operate the pneumatic riveter, no holding up being required, this does not appear to be an excessive charge. The process, depending as it does on " squeeze rivet ing," demands pneumatic riveting machines capable of a pressure of several tons, and these naturally cost a good deal of money, so that the De Bergue process does not lend itself to cheap riveting in small quantities. As soon, however, as there is production work to be done, the process should be cheaper than normal riveting. Diagrams of De Bergue riveting By the courtesy of Handley Page, Ltd., we saw some De Bergue machines at work in the Cricklewood works, and secured a few photographs. The riveting machines used there are of the De Bergue No. Ill type, which weigh some 8 cwt. In this model the gap of the frame is vertical, with the ram placed horizontally. The riveting ram has a vernier adjustment for the gauge of sheet being riveted, and this governs also the degree of pressure which the finished rivet head exerts on the sheet. The rivet is inserted in the drilled hole, with its flat head against the, at that stage, flat surface of one of the sheets. The free end of the rivet shank passes into a recess in the opposite member (which may be regarded as the " dolly ",), which is also shaped to receive the cupped portion surrounding the rivet. Finally, the ram presses right home, and in so doing, forms, by plain squeezing, the rounded head of the rivet. The speed of operation will obviously vary with the nature of the work to be done, tanks of complicated shape being slower to rivet than straight runs, but as an average figure the company quote some 150 per hour, less on some jobs and more (up to 200 per hour) on straight runs. Handley Page, Ltd., fitted petrol tanks made by De Bergue riveting into the Hannibal and Heracles, and these have now done many hundred hours' flying without any trouble. The success achieved with the first tanks led to the general adoption of the system for all petrol tanks, and those fitted in the new " Hey- fords " are all manufactured on the De Bergue prin ciple. The system can also l>e applied to such items as large panels, wheel fairings, etc., and the fact that the flat rivet heads are sunk in flush with the surrounding sheet results in a very neat external appearance. For petrol-tank construction a material known as Petroquoil Jointing is inserted between the two sheets to be lap-jointed. The tanks themselves are made of Alclad, while the rivets are of Duralumin. The first tank of a new type is subjected to a pressure of about 6 lb./>sq. in., and subsequent tanks have to pass tests at 1.5 lb./sq. in. This refers to main tanks. For gravity tanks and oil tanks the pressure (" subse quent ") is 1.75 3b./sq. in. On vibration tests a tank was filled with paraffin and subjected to 10£ million vibrations, after wnieh it was tested at 3 lb./sq. in., which it withstood satisfactorily, so that it would appear that the De Bergue system ol riveting is proof against vibration troubles. To give an idea of tank weights when De Bergue riveting is used, the following tank weights from the Handley Page " Heyford " may be of interest. Main petrol tank, 103 gallons, 68 lb.; oil tank, 10 gallons, 12.5 lb.; gravity tank, 12 gallons, 11.5 lb. THE DESIGN OF AEROFOILS AND THE PREDICTION OF CHARACTERISTICS By W. Pv. ANDREWS, A.F.R.Ae.S. (Continued from page 815) The application of this interesting correction to the calculated moment of a few mathematical centrelines will illustrate how it is applied and provide data for design and analysis. Table X gives a few curves suitable for use as centre lines, together with their calculated and empirically- corrected moments. It will be noticed that 1 and 2 are of the same form, but 2 has been included as it is the special case where a — 1.0 and the trailing-edge slope is zero, i.e., the centreline is tangential to the chord at the trailing edge. For values of " a " less than 1.0 the section has a reflexed trailing edge which gives zero moment at no-lift when a = 0.8675. If an aerofoil of this series is required having the point of maximum camber at some particular point x, the value of " a " to satisfy is given by «! (2 — 3 xx) ~~2*i 1 (10) This relationship is plotted in Fig. 13. As " a " increases towards infinity, the shape of the centreline tends towards No. 3, which is the limiting case. Fig. i4 gives the value of h in terms of camber for different values of as,. There is too small an amount of data to generalise on aerofoils of this shape of centreline, but as will be shown later the maximum lift is rather disappointing, being no greater than that of a symmetrical section oi the same thickness. The form of centreline which seems to the writer to be of the greatest interest is No. 6, of which Nos. 4 and 5 are particular cases. With this form of centreline the point of maximum camber can be arranged to come at any point along the span (#,) by a suitable choice of "n." It will be shown in the section dealing with maximum lift that this value of oc, is important where the maxi mum possible lift at fixed moment coefficient is to be obtained. The value of the leading-edge slope is +nh and the trailing-edge slope —h for all values of n. For values of " n " greater than about 2, the rear portion of the wing is perhaps flatter than is common practice, but this is not anticipated to be a disadvan tage, either to lift or drag, until the nose slope becomes absurd at high values of n. R.A.F.28 has centreline of this family with n = 2. Fig. 16 gives the variation of Km,, in terms of values of h for different values of n. This curve is a maximum at n = 2.0, and is sensibly constant at 0.39 for values of n between 2 and 3, which represents the part it is anticipated would be most used. We can now pass on to considerations of the angle of no-lift. In the theory of R.&M. 910 the no-lift angle is shown to be dependent upon Fx = ^y^f^dx only, but an Jo , , analysis of the series of tests under review suggest that this is not quite borne out in practice. 1310d
Sign up to
Flight Digital Magazine
Flight Print Magazine
Airline Business Magazine
E-newsletters
RSS
Events
