196 



MEASUREMENT OF PRESSURES 



completely quantitative measurement, even though tests seem to indi- 

 cate its unimportance. As an extreme example, suppose that the 

 maximum signal developed by an exponential pressure wave is fifteen 

 per cent of the peak signal from the gauge. This signal will decrease 

 very gradually and the relative gauge and cable signals as a function of 

 time will look somewhat as sketched in Fig. 5.22. It is evident that 

 the indicated signal will be double the true value when the latter is 

 fifteen per cent of the original peak signal. The error in indicated area 

 under the pressure-time curve also becomes increasingly great; for the 

 final time in Fig. 5.22, it is already forty-five per cent of the true value. 



Fig. 5.22 Effect of cable signal on piezoelectric gauge records. 



The indicated value of rapidly changing pressures, such as peak pres- 

 sure in the example, will not be in error because of the integrated, or 

 "low-frequency," nature of cable signal. 



With a given cable, there are several expedients which can be used 

 to minimize the signal. The most obvious is that of leading the cable 

 directly away from the source of pressure in order to minimize the 

 length of cable exposed and reaUze attenuation of pressure as rapidly as 

 possible. A second, which has been used in measurement of bubble 

 pulse pressures, is to lead the cables to the water surface as directly as 

 possible and run them just below the surface, a region in which the 

 pressure wave is almost immediately cancelled by the reflected pressure 

 wave from the surface. A third expedient makes use of the "season- 

 ing" effect already mentioned, whereby cable signal is reduced after 

 application of pressure. Although this technique has been used with 

 some success, it is clearly makeshift and, it is to be hoped, is of historical 

 interest only. 



