280 



BELL SYSTEM TECIIXICAL JOURNAL 



except for an integral number of cycles. The same will apply to a line CC , 

 to which the distance from each element is exactly two wave lengths more or 

 less than from its immediate neighbors, or to any other line where this dif- 

 ference is any integral number of wavelengths. 



Now in no radar antenna do we desire two or more beams for they will 

 result in loss of gain and probably in target confusion. The array must 

 therefore be designed so that for all positions of scan all beams except one 

 will be suppressed. This will automatically occur if the array spacing is 

 somewhat less than one wavelength. If the array spacing is greater than 

 one wavelength these extra beams will appear in the array factor; they 



Fig. 37 — Some Possible Wave Fronts of an Array of Elements Spaced 2.75 X. 



must therefore be suppressed by the pattern of a single element. The pat- 

 tern of an element must in other words, have no significant components 

 in any direction where an extra beam can occur. 



Where elements with only side fire directivity are spaced more than a 

 wavelength apart in a scanning array it is almost impossible to obtain 

 adequate extra lobe sui)pression. If these elements are spaced by the 

 minimum amount, that is by exactly the dimensions of their apertures and 

 all radiate in phase the}- may indeed just manage to produce a desirable 

 beam. A little analysis shows however that an appreciable phase variation 

 from element to element, e\'en though linear, will introduce a serious ex- 

 tra lobe. To get around this difiKulty elements with some end lircdirec- 

 livity must be used. 



