84 



EXPERIMENTAL PROCEDURES 



it is not likely to be a dominant cause of error in a 

 determination of the transmission loss. 



Transmission work at single sonic frequencies 

 began only recently, and the analysis procedure has 

 not yet been very well standardized. Records ob- 

 tained up to the present appear to indicate that 

 fluctuation of signal intensity is much less severe at 

 sonic frequencies than at supersonic frequencies. Onv 

 the other hand, because of image interference, sys- 

 tematic changes in signal level are observed at short 

 ranges which vary so rapidly with range that any 

 averaging procedure would obscure them. For this 

 reason, in transmission work at frequencies from 200 

 to 1,800 c individual signal levels rather than sample 

 averages are reported. 



In the records obtained at sonic frequencies the en- 

 velope of the signal trace, as a rule, is badly serrated. 

 The fuzziness of the envelope is probably caused by 

 the unfavorable signal-to-noise ratio, which is about 

 1 db for 1.8 kc and lower, somewhat higher for 

 22.5 kc, and by the relative narrowness of the filters 

 used in the recording channels. If random noise is 

 received through a wide filter, the oscillograph trace 

 has a typical "spiked" appearance, that is, the noise 

 is characterized by sudden high peaks of short dura- 

 tion. If the filter is narrow, as it must be in sonic 

 transmission work, the individual peaks are lowered 

 and broadened, and their separation from the 

 single-frequency signal is more difficult. For this 

 reason, the person reading the film record does not 

 attempt to measure the "peak" level, which would 

 be fictitious, but estimates and reports the average 

 amplitude. It has been found that the uncertainty 

 introduced by this estimate is less than 1 db, on 

 the average. 



The final step in the processing of a transmission 

 run consists of the recording of the computed signal 

 intensity. Since in most transmission runs the range 

 is altered by a factor of 10 to 100, the signal levels 

 change in the course of a run by a large number of 

 decibels. It has, therefore, been found useful not to 

 plot signal level in decibels below transducer output 

 directly, but to take out the bulk of the variability 

 by plotting the transmission anomaly. Usually, the 

 transmission anomaly is plotted as the ordinate down- 

 ward, with range as the abscissa. Theoretically, this 

 curve should pass through zero for zero range. In 

 view of the great experimental difficulties involved in 

 the determination of the signal level at very short 

 ranges, the curves usually stop at a range of 100 yd 

 or more. 



4.4 



ECHO RUNS 



As mentioned before, transmission runs are by far 

 the most important useful method of obtaining in- 

 formation on the propagation of sound in the ocean. 

 The other two methods, which are of secondary im- 

 portance, will be discussed in the next two sections 

 for the sake of completeness. 



Echo runs have been carried out both on specially 

 designed standard bodies and on chance targets, such 

 as wrecks, in order to study the dependence of echo 

 level on range and in order to study fluctuation and 

 coherence. The principal purpose of echo runs has 

 usually been to study not the propagation of soimd 

 between echo-ranging transducer and target, but 

 rather the effect of certain targets on the received 

 echo. (See Chapters 18 to 26 of this volume.) 



The equipment used in echo rims differs from that 

 used in transmission runs in that sending and re- 

 ceiving stacks are aboard the same ship and the same 

 sound head is used both for sending and receiving. A 

 change-over relay connects the sound head first with 

 the sending stack and then, immediately following 

 the emission of the signal, with the receiving stack. 



Artificial targets have been developed for research 

 and training purposes. Natural targets usually re- 

 flect very differently at different aspects; most arti- 

 ficial targets are designed to minimize this kind of 

 directionality without sacrificing too much overall 

 reflecting power. There is one geometrical shape 

 which remains the same regardless of any twisting of 

 the cable from which the target is suspended. That is 

 the sphere. From the point of view of constant re- 

 flecting power, spheres constitute ideal artificial 

 targets. 



Unfortunately, the reflecting power of a sphere, 

 while constant, is fairly small. To obtain useful 

 echoes from spheres at distances similar to the ranges 

 commonly encountered in practical echo ranging, one 

 would have to use spheres with a diameter greater 

 than 30 ft. It was found, however, that a 10-ft sphere 

 was almost unmanageable at sea. The only spheres 

 which could be handled with ease were spheres with a 

 diameter of 2 or 3 ft. 



In the search for an artificial target with a large 

 target strength, the best solution found so far has 

 been the triplane^' shown in Figure 15, which com- 

 bines ease of handling with a reflecting power com- 

 parable to that of a submarine. It is a well-known 

 fact, sometimes used in optical signaling, that a ray 

 whichhas been reflected from three planes which are 



