ESTABLISHMENT OF SOUND FIELDS 



65 



(or current) at some frequency. A standard receiver 

 is then placed at an appropriate position in the sound 

 field, and from its generated voltage the magnitude 

 of the sound held can be obtained. The reference 

 standard is then replaced by the receiver under test 

 and its generated voltage obtained, giving its re- 

 sponse at that frequency. The frequency may be 

 swept during each test, and then, by comparing meas- 

 ured values of the generated voltage of the standard 

 and the transducer under test at ecpial frequencies, 

 the frequency response characteristic of the trans- 

 ducer may he obtained. 



It should be noted that, if a plane wave calibration 

 of the instrument is desired, then, at the position se- 

 lected for the receivers, the waves from the transmit- 

 ter must be essentially plane with respect to the 

 instruments. Whether or not this is the case depends, 

 among other things, on the nature of the instruments 

 themselves, particularly on their size. Because of these 

 conditions, the waves at one point may be essentially 

 plane for one receiver but not for the other. In this 

 case, it may be more convenient to test the two instru- 

 ments at different distances from the source. This 

 may be done, provided both distances lie in the in- 

 verse-square-law region for the source, so that only an 

 inverse-square-law correction need be applied. In 

 some cases one may find it desirable or expedient, 

 because of the presence of reflection interference, to 

 operate the instruments at closer distances and to 

 apply a spherical wave correction to the result, as has 

 been described previously. 



In calibrating a transmitter, one uses a calibrated 

 standard receiver which is located in the inverse- 

 square-law region of the field of the transmitter. The 

 pressure produced at this point can then be deter- 

 mined from the voltage generated by the receiver. 

 Again, it may be desirable or expedient to locate the 

 receiver closer to the transmitter and apply a spheri- 

 cal wave correction. 



Since a relative calibration is based on the stability 

 of calibration of the standard, frequent checks on the 

 calibration of the standards must form a regular part 

 of any testing program extending over a long period 

 of time. These checks are most conveniently made at 

 regular intervals by means of the reciprocity method. 

 The reciprocity method is particularly valuable for 

 this purpose at a test station since the procedures used 

 are the same as for relative calibrations. Thus the 

 equipment necessary for a reciprocity calibration is 

 available at all times and no extra equipment is neces- 



sary. A running check on all (esls may be made by 

 using several standards and checking the calibrations 

 thus obtained against each other. A calibrated trans- 

 mitter may serve as one ol the standards. 



5.5.10 The Choice of Standards 



For maximum accuracy, reliability, and general 

 versatility, a standard should possess certain char- 

 acteristics which are outlined below. 



.Stability With Time 



Stability is essential, since the accuracy of a relative 

 calibration is limited by any change in the response 

 of the standard from the time that it was calibrated 

 to the time of its use. For this reason, it should be suf- 

 ficiently rugged so that slight jars do not change its 

 calibration, and it should be construe ted of materials 

 whose properties do not change with time. If it con- 

 tains permanent magnets, their flux density should 

 be permanent. The stiffness of springs should not 

 vary as a result of aging, or of cold working resulting 

 from their extension and retraction. The instrument 

 should be constructed so that dampness or moderate 

 heat or cold do not change its calibration, and all ex- 

 posed parts should be resistant to corrosion. 



Temperature Independence 



It is desirable that the response of a standard be 

 independent of temperature over the useful fre- 

 quency range of the instrument, since the tempera- 

 ture rarely can be controlled in calibration tests. 

 Otherwise, it is necessary to know how the calibration 

 varies with temperature, which would require con- 

 siderable additional labor. Temperature dependence 

 of response is an important factor in working with 

 Rochelle salt crystal devices because of the rapid 

 variation of the dielectric constant of v-cut crystals 

 and consequent change of impedance in the neigh- 

 borhood of the upper Curie point (24 C or 75 F). This 

 temperature variation does not have a serious effect 

 upon the response of a Rochelle salt crystal receiver 

 if the impedance into which the crystal operates is 

 high compared to its own impedance. For a Rochelle 

 salt crystal transmitter, the response expressed in 

 terms of pressure per unit current input does not vary 

 appreciably with temperature, although the response 

 expressed on a per volt or per available watt basis 

 may vary greatly. It is therefore desirable to operate 

 Rochelle salt crystal projectors on a constant current 



