1158 THE BELL SYSTEM TECHNICAL JOURNAL, OCTOBER 1951 



data shown as circles falling near the lowest curve. Some of the measured 

 points have been omitted to avoid crowding but enough remain to show 

 the trend. At low frequencies these points fall along a line of approximately 

 6 db per octave. 



A straight 6 db per octave line labeled 1 has been drawn through these 

 points and extended as shown in the figure. This line is the base from which 

 the various losses must be subtracted. Curve 2 shows the effect of sub- 

 tracting the computed thickness loss. When eddy losses and gap loss are 



50 



45 



35 



30 



<n 25 



O 20 



ui 



o 



10 



z 



UJ 



Q. -10 



O 



-15 



-20 



THICKNESS 

 ) LOSS 

 I (^=0.3 MIL 



\ MEASURED 

 r EDDY LOSS 



GAP LOSS 

 9=0.5 MIL 



SPACING 

 > LOSS 

 d =0.23 MIL 



-d = 0.36MIL 



-d = 0.81MIL 



0.3 0.4 0.6 0.8 1.0 2 3 4 



FREQUENCY IN KILOCYCLES PER SECOND 



6 a 10 



20 



Fig. 9 — Computed response curves and measured response points. 



also taken into account, curve 4 is obtained. The difference between this 

 curve and the lowest measured response points is presumably due either to 

 self-demagnetization, to spacing loss, or perhaps to both. 



There is one clue which may be of help in deciding how much of this loss 

 should be attributed to self-demagnetization and how much to spacing loss. 

 This clue comes from the fact that the form of the spacing loss function 

 is known. Any part of the loss which is due to spacing must follow the equa- 

 tion 



Spacing Loss = 54.6 ((//X) db 



