252 SHOCK WAVE MEASUREMENTS 



0, and the pressure is rapidly built up to a maximum, actually by for- 

 mation of a shock front. The pressure then falls off, until after a time 

 interval ^i = L{l/c + l/D) the last, much weaker, signal arrives from 

 point L. The duration of the wave is then very nearly ^i. At point B, 

 the first signal arrives from point L, and at a time ^i = L{l/c — l/D) 

 later the last weaker signal from arrives. The duration at point B, 

 in the direction of detonation, is thus shorter than at the other end, and 

 the pressure-time curve is, so to speak, compressed at point B. Cal- 

 culations based on Eq. (7.7) show that, approximately, the pressures at 

 B would be increased over those at A in the ratio that the time scale 

 is reduced. This ratio is (D -\- c)/{D — c), which for D = 20 ft. /msec, 

 c = 5 ft. /msec, is 5/3. The pressure-time curves at opposite ends 

 should therefore look roughly as sketched in Fig. 7.13. 



The qualitative predictions represented by Fig. 7.13 are confirmed 

 surprisingly well by experiment. Calculations based on Eq. (7.7) have 

 been applied to experimental data for charges of cordtex (similar to 

 primacord, 3 grams of explosive per foot) and found to represent the 

 general characteristics of the observed pressures quite well with 

 Pm = 5,000 Ib./in.^ and ^ = 17 ^tsec. The experimental records un- 

 fortunately did not show the full duration of the waves, the later por- 

 tions of which were cut off by surface reflections. Similar experiments 

 with larger charges (50 pounds of TNT in a 25 foot length) in much 

 deeper water^ gave durations of 6.5 and 3.8 msec, for positions corre- 

 sponding to A and B in Fig. 7.12, as compared to values of 6.25 and 

 3.75 msec, calculated using D = 20 ft. /msec, and c = 5 ft. /msec The 

 peak pressures were roughly 40 per cent higher at points A, as com- 

 pared with the estimated increase of 67 per cent. The measured im- 

 pulses, however, differ much less, being of the order of 15-20 per cent 

 higher at B, as expected from the approximate estimates. 



For points on the perpendicular bisector to the charge, such as point 

 C in Fig. 7.12, the situation is quite different. The duration of pressure 

 for continuing detonation depends on the distance from the axis, but 

 must always be shorter than at points off either end, and decreases as 

 the distance increases. At points sufficiently far from the charge, the 

 peak pressures are much higher than off the ends, as shown in reproduc- 

 tions of the observed pressure-time curves in Plate VII. Near the 

 charge, however, the peak pressures increase less rapidly than they 

 would for a spherical charge, and the peak pressures off the end increase 

 more rapidly. This difference in behavior is shown in the plots, on 

 logarithmic scales in Fig. 7.14, of peak pressure P,„ as a function of 

 l/vo for 50 pound, 25 foot TNT charges. At distances greater than 

 about 30 feet (from the nearest point of the charge), all the curves have 



8 These results are given in more detail in a report by J. S. Coles and U. II. Cole, 

 to be issued as a NavOrd Report by the U.S. Navy Bureau of Ordnance. 



