REPRODUCTION OF MAGNETICALLY RECORDED SIGNALS 1157 



By a completely different sort of measurement,^ J. R. Anderson has 

 arrived at a similar value for eddy current loss in this type of head and 

 has shown that approximately the same loss occurs in both the recording 

 and the reproducing process. For this reason it seems proper to assume 

 that eddy currents account for just twice the loss measured by the method 

 of Fig. 8. 



The loss due to the finite gap in the reproducing head is computed from 

 the well known relation.^ 



Gap loss = 20 logi- '^^^^ 



sin irg/X) 



where g is the effective gap width in inches and.X is the recorded wave- 

 length in inches. 



Thickness loss is computed from equation (5). It must be remembered 

 that this loss was derived on the assumption of uniform magnetization 

 throughout the thickness of the recording medium. This may be a fairly 

 good approximation to the actual state of affairs for a thin medium such 

 as the one being considered, but obviously if the thickness of the medium 

 is large compared with the width of the recording gap then the recording 

 field will not penetrate uniformly through the medium and the derived 

 thickness loss function will not apply. 



The derived equation (3) indicates that at low frequencies the reproduced 

 voltage should be proportional to the thickness of the medium. If the thick- 

 ness of the medium is increased beyond the limit to which the recording 

 field can penetrate, this will no longer be the case and further increase in 

 thickness will have no effect on the response. 



Data presented by Kornei^ on the cobalt-nickel plating being considered 

 here shows that the low-frequency response is approximately proportional 

 to the thickness of the medium for values of thickness between 0.075 mil 

 and 0.5 mil. This may be taken as an indication of approximately uniform 

 penetration through these thicknesses and hence tends to indicate that 

 the derived thickness loss function should be applicable in the case of the 

 0.3 mil plating being considered here. 



The effects of these losses are shown in Fig. 9 along with measured fre- 

 quency response data. Consider first the experimentally measured response 



5 In unpublished work, J. R. Anderson of the Bell Telephone Laboratories has made 

 use of the fact that eddy losses depend on frequency while all other magnetic recording 

 losses depend on wavelength. By recording a single frequency and playing back at vari- 

 ous speeds he determined the loss on playback. By recording various frequencies with 

 recording speed adjusted to give constant recorded wavelength and using a single play- 

 back speed he evaluates the eddy loss in the recording process. 



6 S. J. Begun, "Magnetic Recording," p. 84, Murray Hill Books, Inc., New York. 



