556 



producing the damage, by means of which the damage for a diaphragm of a 

 given thickness can be reduced to that for a diaphragm of standard 

 thickness. A plot of these correction factors as a fimction of diaphragm 

 thickness is given in Figure 19- The given damage is simply multiplied by 

 the proper correction factor (x) to obtain the damage which wovild have been 

 obtained with a diaphragm of standard (.O85 in.) thickness. More accurate 

 correction factors may be computed by including the contribution of the 

 variation in g, but this is not necessary as long as the total correction 

 is small. 



This treatment can be applied similarly to copper diaphragms. 



(ii) Empirical determination , (aa) Lots 3 sjid $ steel diaphragms . 

 By suitably choosing diaphragms of different thicknesses for the four 

 damage gauges used in single shots, the effect of thickness variations on 

 damage was determined experimentally. For example, two diaphragms .O85 

 in. thick and two diaphragms .075 in. thick would be damaged under the 

 same conditions, and the correction factor for .075 in. diaphragms ob- 

 tained directly. Having obtained such information for diaphragms of all 

 thicknesses in the Lot 3 range it was then possible to compare the results 

 for different shots in which diaphragms of different thicknesses were used. 



In Figure 19, the empiricail correction factor is plotted as a function 

 of diaphragm thickness and compared with the curve obtained from Kirk^- ^d's 

 theory (V, k, a). The data for the empirical curve was all obtained u_j.ng 

 the steel ring mounting, although similar results (with respect to thick- 

 ness correction) were obtained when gauges were mounted on wooden frames. 

 The deviation from the theoretical curve is greatest for the thickest 

 diaphragms. Although the majority of the data used in determining the 

 empirical curve have been for damages of about 0.70 in., there is some 

 evidence that this correction curve is not r, function of the amount of 

 damage. In a series of shots in which diaphragms of .075 in. thickness 

 were compared with .093 in. diaphragms, the mean percentage increase of 

 damage for thin over thick diaphragms was constant at 20 * 1'^ for damages 

 of 0.23 in., 0.53 in., 0.7I in., and 0.93 in. (Cf results with copper 

 diaphragms below) . 



The theoretical thickness correction was partially corroborated also 

 by results obtained with larger charges. A 50 lb. chemical series of 

 12 charges of various compositions was shoti;^/ with two UERL diaphragm 

 gauges 25 ft. from the charge and two gauges 35 ft. from the charge. A 

 Lot 5a (ca. O.O38 in. thick) diaphragm was placed in one of the two 

 gauges at each of these distances, while a Lot 5 (ca. O.O8O in. thick) 

 diaphragm was placed in the second gauge at these distances. The actual 

 diaphragm thicknesses were measured before the shot and the maximum 

 depressions were corrected to standard thickness (.O38 in. for thin 

 diaphragms and .O85 in. for medium diaphragms). The corrected damages 

 obtained for the thin diaphragms were then compared with the corrected 

 damages for the medium diaphragms subjected to the same explosive shock 

 wave; the mean damage ratio was found to be 2.13 ( O^m ~ O.OI8) for the 

 25 ft. distance and 2.l6 { (r^ = O.OO6) for the 35 ft. distance. 



15/ Reported in OSRD No. 62^40 and 62^+1. 



