150 



PRODUCTION TESTING OF SONAR TRANSDUCERS 



measurement of response in a production test per- 

 formed in a comparatively small tank may involve a 

 smaller testing distance than is required to render 

 negligible the various proximity effects. Such a test 

 can nevertheless be satisfactory, provided there is a 

 definite correlation between comparison measure- 

 ments on the standard and on the units under test at 

 large and at small distances."' 5 



Unless proper precautions are taken, response 

 measurements made in a tank are not precise because 

 of the interference between the direct and reflected 

 waves. If the testing distances required because of 

 proximity effects are not too large, it may be prac- 

 ticable to use a fairly small tank with walls having 

 from 15- to 20-db absorption. 1 ' Such absorption will 

 reduce the intensity of the reflected waves sufficiently 

 to reduce the error in the axis response to ± 1 db 

 (under steady-state conditions). If the units to be 

 measured are directional, and if directional standards 

 are used, the sources of the more bothersome reflec- 

 tions may be treated with the recently developed 

 "bubble" layer, which provides 5- to 10-db absorption 

 and probably renders negligible the overall remain- 

 ing reflection interference. (See Chapter 6.) 



In addition to, or in place of, the use of absorb- 

 ing walls, electrical methods for eliminating the ef- 

 fects of reflections may be employed. These methods 

 involve the use of noise or frequency warble, or of 

 pulses. The relative advantages of these methods are 

 discussed in Chapter 5. Their usefulness depends on 

 the nature of the response of the transducer being 

 tested, since each method entails a loss of resolving 

 power in the curves of continuous- wave response ver- 

 sus frequency. For a unit with fairly uniform response 

 (Q small), all of these methods are quite satisfactory, 

 since little resolving power (RP) is required (RP = Q). 

 For a highly resonant transducer, the use of all these 

 methods with a tank of given size is limited by re- 

 solving power consideration. The resolving power 

 will have to exceed Q, and this implies a minimum 

 allowable path difference AL between direct and re- 

 flected waves, greater than <"Q//, C for example, greater 

 than 3 meters for / = 25 kc and Q = 50. Regarding 

 the relative merits of continuous-wave noise or warble 



b An absorbing tank approximately 3 ft long by I1/2 ft wide 

 by 2i/2 ft deep has been built by the Bell Telephone Labora- 

 tories.'^"' 



<■ Here c is the velocity of sound, / is the frequency, Q is de- 

 fined precisely in Chapters 4 and 5. A derivation and discussion 

 of this equation is given in Chapter 5. 



versus pulses, the latter may be considered superior, 

 since with their use reflections can be eliminated 

 completely from the measurements, while with the 

 former the reflections are averaged in and yield a 

 time-independent, but not always known, correction. 



8.2.2 



Impedance Measurements 



The measured electrical impedance of a transducer 

 depends on the terminating acoustic impedance. (See 

 Chapters 3, 4, and 5.) This fact deserves consideration 

 because the presence of reflected waves incident on 

 the face of the transducer constitutes a change in the 

 terminating acoustic impedance and so will affect any 

 electrical impedance measurement. If the impedance 

 measurements are made in a tank with absorbing 

 walls, the reflections may be sufficiently small to allow 

 the measurements to be made in a conventional way 

 with an impedance bridge. 



On the other hand, when the reflections are appre- 

 ciable, their effect can be eliminated by pulsing. The 

 pulse length is presumably determined by the same 

 criterion as in the directivity pattern and response 

 measurements (see Section 8.2.3). For pulse measure- 

 ments, a wattmeter with a time constant small com- 

 pared to the length of the pulse may be used. d In this 

 case the impedance is obtained from voltage, current, 

 and power. 



8.2.3 Directivity Pattern Measurement 



The measurement of directivity patterns imposes 

 the most severe test requirements. This is chiefly due 

 to the fact that measurements of intensity 40 or 50 db 

 below the response on the axis must be made in com- 

 petition with reflections of the main beam. Even with 

 the best of the available absorbing materials, it is im- 

 possible to prevent some reflections from interfering 

 with the measurement of the directivity pattern of a 

 moderately directional transducer. Noise and fre- 

 quency warble are not suitable for directivity meas- 

 urements because they average in an interfering re- 

 flection of unknown magnitude with the direct signal. 

 The error introduced thereby may be considerable 

 for certain directions, as when a direct signal received 



d The recording wattmeter described in Chapter 6 can be 

 used in conjunction with a pulse recording system. Indeed, this 

 system has the additional advantage of being able to measure 

 impedance at high power levels without unduly heating the 

 transducer. 



