II. EXPERIMENTAL RESULTS 



2 . 1 Averaged Data Analysis . The acoustic signals received on the Rexburg 

 were recorded on multi-channel magnetic tape. Later, in the laboratory, the 

 tapes were played back through narrow band audio filters and the signals re- 

 corded on galvanometer strip -charts . The circuitry is described in Appendix 

 C, as is the detailed analysis involved in converting an observed galvanometer 

 pen deflection to a measurement of transmission loss . Since each run provided 

 data for nine transmission loss curves (1 projector depth, 3 receiving hydro- 

 phones at different depths, 3 acoustic frequencies) the entire program provided 

 data for a total of 54 such curves . 



Experimental data for all 54 curves were first reduced by an averaging process . 

 The amplitudes of the ten signals received at one frequency during a five -minute 

 period were averaged and converted to a received pressure amplitude (db re 1 

 Mbar). Since the source level was known, the transmission loss was derived im- 

 mediately. The chosen five -minute periods were separated by ten -minute inter- 

 vals . The range was determined from the time of transmission of the acoustical 

 ping. The experimental results so obtained are presented in Curves 1-54 in 

 Section IV. Table 2 provides an index to the curves, listing source and receiv- 

 ing hydrophone depths, frequency and the experimental region for each curve. 



The data reduction was complicated by problems that arose during the experi- 

 mental program . These included (a) interference between the clock and radio 

 synchronization signal channels on the one hand and the receiving hydrophone 

 channels on the other, (b) cross -talk between the audio -radio and receiving hy- 

 drophone channels, (c) extreme distortion, amounting to uninte legibility, of the 

 radio voice link, and (d) frequency instability of the timing clock. In consequence, 

 messages concerning calibration and attenuator settings were missed on several 

 occasions . Analysis was made more difficult by the absence of calibrations at 

 all frequencies on all receiving channels . The analytical method followed and 

 the assumptions made to circumvent these omissions are described in detail in 

 Appendix C . In the absence of calibration signals, a uniform cakbration estab- 

 lished elsewhere in the program was employed (see Appendix C). Occasionally 

 when attenuator settings were obviously incorrect, the curves were normalized 

 by comparing the observed noise level (ONL in the curves) with the expected 

 noise level (ENL) and adjusting the levels of the curve in steps of 10 or 20 dbs. 

 (Attenuator changes less than 10 dbs were not available while the original re- 

 cordings were made .) Appendix C contains a detailed record of all attenuator 

 adjustments that were considered necessary. Despite all efforts, it was not 

 possible to produce internally consistent and reliable data from the shallow hy- 

 drophone (Channel 1) on the Rexburg. It was concluded that this channel suffered 

 from an intermittent electronic fault, probably in the play-back amplifier section 

 of the magnetic tape recorder. The argument in support of this conclusion 



3rthur ZH.mttle.3nc. 



