278 BELL SYSTEM TECHNICAL JOURNAL 



reduction in cross-talk caused by shielding may be written 



a = 20 logjo^^ = 10 logio ( 1 + ll) • (3) 



In the 1/32 inch copper shields ordinarily used for filters, the para- 

 meter "P" has been found to have values from 1250 at 500 cycles 

 to 2500 at 15000 cycles. 



From Equation 1, the change in impedance, AZ of the coil, caused 

 by the presence of the shield, may be written 



. „ _ , WK'^L.UR,] _ . WK'L.U^ ... 



This change is made up of a decrease in reactance and an increase 

 in the effective resistance of the coil. The percentage change in both 

 resistance and reactance may be written 



ALi 

 and 



Lj " P2 + co^ ' ^^^ 



The percentage change in the inductance of the solenoidal air-core 

 coils mounted inside ranged from 2 per cent to 4 per cent, depending 

 upon the proximity of the coil to the shield. Since these changes 

 depended upon the distance of the coils from the shield, slight vari- 

 ations in the winding diameters of the coils caused variation in the 

 magnitude of this cover effect, so that although this change in in- 

 ductance was taken into account by specifying inductance values for 

 the filter coils 2 to 4 per cent higher than indicated by the design 

 computations, the corrections were only approximate. No such 

 correction was, of course, necessary for the condensers; the condensers, 

 therefore, as measured outside the filter shield, were still correct, but 

 the coils under the same conditions, measured higher than specified 

 values. It was obvious that the resonant adjustment should not be 

 made with the coils outside their shields, as the approximate nature 

 of the cover correction would vitiate the precision of the adjustment, 

 but it was equally obvious that it would be very difficult to make a 

 coil adjustment inside the soldered cans, as the elements would then 

 be inaccessible. 



