1156 THE BELL SYSTEM TECHNICAL JOURNAL, OCTOBER 1951 



following two kinds of loss should be added to this list: 



4. Spacing loss due to imperfect magnetic contact between the repro- 

 ducing head and the recording medium, and 



5. Thickness loss. 



Of these five losses three can be evaluated quantitatively either by direct 

 measurement or by calculation from theory. The remaining two are self- 

 demagnetization and spacing loss. 



In this section the known losses will be evaluated for a particular re- 

 cording system. This leads to a response curve which can be compared with 

 the measured curve. The difference between the two curves should be due 

 to self-demagnetization and to spacing loss provided the above list of losses 

 is complete. 



0.6 0.8 1.0 2 4 6 8 10 20 40 



FREQUENCY IN KILOCYCLES PER SECOND 



60 80 100 



200 



Fig. 8 — Measured eddy current loss as a function of frequency. 



The recording system used is the one shown in Fig. 1 with the speed set 

 at 15.5 in./sec. for both recording and reproducing. A constant signal cur- 

 rent of 0.1 ma was used for recording with the 55 kc bias adjusted to give 

 maximum open circuit reproduced voltage. 



Eddy current losses were measured as indicated in Fig. 8 by sending a 

 measured constant current i through a small auxiliary winding around the 

 pole tip and measuring the open circuit voltage developed across the normal 

 winding of the head. Any departure of this measured voltage from a 6 db 

 per octave increase with increasing frequency is due to losses in the head 

 which will be loosely called eddy current losses. Other kinds of loss may 

 enter into this measurement (as, for example, loss due to the self-capaci- 

 tance of the winding) but in the frequency range of interest, eddy losses 

 predominate. 



