366 BELL SYSTEM TECHNICAL JOURNAL 



on that account, not altogether objectionable. The Holmdel MUSA 

 employing such unit antennas represents, however, a considerable 

 departure from present antennas of fixed directivity designed from 

 statistical data, and approaches the ideal MUSA steerable over the 

 entire useful angle range. 



The curves as plotted assume that the dififerences in transmission line 

 loss for the various line-lengths have been equalized in the inter- 

 mediate-frequency circuits. By slightly tapering the amplitudes so 

 that the antennas in the middle of the array contribute more than 

 those near the ends a reduction of the minor lobes has been obtained 

 at the cost of slightly widening the principal lobe. As a result of this, 

 the directional discrimination of the experimental MUSA has been 

 improved. All data and photographic records reported in this paper, 

 however, were obtained before this improvement was introduced. 



IV. Tests and General Evaluation ^^ 

 Tests and Experience 



Numerous experiments and tests had been carried out on the various 

 parts of the MUSA system before it was first tuned to a transatlantic 

 signal. Despite the fact that all tests concurred in predicting that 

 the system would perform as designed, it was with considerable gratifi- 

 cation that a pattern was observed on the monitoring oscilloscope, 

 during one of the early trials, which was almost exactly as calculated 

 for a single wave. Patterns corresponding to two or more waves in 

 various degrees of resolution were observed from time to time. To 

 increase the angle resolution, for test purposes, pulses were trans- 

 mitted by the British Post Office on several occasions. Turning the 

 steering shaft during these tests clearly showed the principal lobe 

 sweeping through the angle range. When fairly discrete pulses were 

 received the minor lobes could be readily identified. In Fig. 21 is 

 shown a sample of motion picture oscillograms of pulse reception. 

 Two principal waves or, more accurately, wave bundles occurred and 

 were separated by the two MUSA branches as shown. For details of 

 the pulse technique employed in these tests the reader is referred to a 

 previous publication.* 



Before exhibiting sample motion pictures of typical patterns dis- 

 played by the angle monitoring oscilloscope and the delay indicator 

 oscilloscope, further discussion of the former is desirable. The photo- 

 graphs of Fig. 22 show the monitoring oscilloscope pattern with a 

 locally produced equiphase, equiamplitude input supplied to each 



^1 The theory and test results of the signal-to-noise advantage are considered 

 together in Part V. 



