TANK DESIGN CONSIDERATIONS 



151 



off the axis of maximum response is averaged with a 

 reflection incident on the projector axis. 



Pulse measurement, therefore, is the only feasible 

 method for obtaining directivity patterns, and even 

 this method has its limitations. As noted above in the 

 discussion of response, and as described in detail in 

 Chapter 5, the pulse method discriminates against re- 

 flections by measuring the direct signal before the 

 reflection arrives. The time elapsing between the ar- 

 rival of the direct and reflected signals is a function ol 

 the size and shape of the testing tank. The practical 

 limits to the maximum allowable size of such a tank 

 restrict, in turn, the maximum length of time during 

 which the pulse measurement may be made. The 

 maximum elapsed time needed for a significant meas- 

 urement of the response depends on the transient re- 

 sponse of the transducer to a suddenly applied sinu- 

 soidal signal, that is, the pulse. For transducers whose 

 frequency response characteristic is fairly uniform, 

 the response reaches its steady-state value in a short 

 time, i.e., a few cycles, and very short pulses and con- 

 sequently small testing distances may be used. For 

 resonant transducers, the time needed for the response 

 to reach its steady-state value is long. In this case, the 

 directivity patterns can be measured only by using 

 long pulses, and correspondingly long testing dis- 

 tances are required. 



The quantitative criterion for the minimum allow- 

 able path difference AL between the direct and any 

 reflected wave, a quantity which must be just larger 

 than the pulse length, has been given above as 

 AL > cQ/f. A theoretical analysis indicates that a 

 much shorter ptdse length, and so a much shorter 

 minimum path difference, may be used in taking the 

 directivity pattern at the resonance frequency of a 

 highly resonant transducer. The criterion for Ai in 

 this case is AL ss transducer diameter. However, this 

 theoretical analysis assumes that the transducer dia- 

 phragm moves rigidly, a condition not generally ob- 

 tained in practice. e 



The directivity pattern as measured should be a 

 close approximation to the directivity pattern that 

 would be obtained at large distances. In order to meet 

 this requirement, the test distance should be such 

 that proximity effects are small, that is, the spherical 

 wave correction should be less than a few db. It is 



questionable whether patterns taken at a shorter test 

 distance can be compared with corresponding pat- 

 terns taken on a secondary standard. There is un- 

 doubtedly some correlation between the directivity 

 patterns of the standard and of the test unit at large 

 and at small distances, but the relation is in general 

 a very complex one. 



83 TANK DESIGN CONSIDERATIONS 



Since the pulse method seems best for directivity 

 measurements, the specifications for a tank suitable 

 for that method will be discussed. Assume that the 

 test distance and pulse length are chosen on the basis 

 of the foregoing discussion. The absorption of the 

 tank walls is of little moment for pulsing. Frequently, 

 however, steady-state measurements of response and 

 impedance are to be made in the same tank. For these, 

 the absorption of the walls should be made as great as 

 practicable. Tank walls made of wood or concrete 

 offer several db of absorption. If the concrete tank is 

 set in the ground, the damp earth provides additional 

 acoustic loss on reflection. A steel tank is to be 

 avoided, if at all possible, because of the high reflec- 

 tion coefficient. A bubble layer may be applied to the 

 walls to increase the absorption if it proves necessary. 



The maximum repetition rate of the pulses used 

 depends on the reverberation time of the tank. If a 

 high repetition rate is desired for ease of measure- 

 ment, some wall absorption should be supplied. f 



While many tank shapes are possible, the simplest 

 to build is a rectangular one as deep as it is wide. It 

 can be shown that, for a given testing distance and re- 

 flection path length, the minimum size is obtained 

 when the line joining the projector and hydrophone 

 is parallel to the long dimension of the tank. This as- 

 sumes that the reflection path lengths from the sides, 

 top, and bottom of the tank are equal to those from 

 the ends. For this arrangement (Figure 1) the length 

 of the tank is given by / = d + AL, where / is the 

 length of tank, d is the test distance, and A/, is the 

 path difference required. 



The width and depth necessary are equal and are 

 given by the relation w = \/2dAL + AL". 



It may be possible to reduce the width of the tank 

 somewhat by the use of completely reflecting baffles 



<• See reference 55 for report on investigation of the effect of 

 pulse length on pattern for a representative resonant transducer 

 (Q = 50) operating at the resonance frequency. 



f For example, in a tank with an average dimension of 15 ft 

 and a wall absorption of 3 db per reflection, about 23 milli- 

 seconds are needed for the intensity of a reflection to tall 45 db. 



