REPRODUCTION OF MAGNETICALLY RECORDED SIGNALS 1149 



spacing between the head and the medium. With this in mind it is easy to 

 see that at high frequencies only a thin layer of the medium nearest the 

 reproducing head will contribute to the reproduced signal. In this case 

 (short X) increasing the thickness of the medium beyond a certain amount 

 can have no effect on the reproduced level simply because the added part 

 of the tape is too far from the head to make its effect felt. Consider the 

 effect of increasing the thickness of the medium from 0.3 mil to 0.6 mil 

 when the wavelength is one mil. Since the spacing loss for 0.3 mil spacing 

 at X = 1 mil is 16.5 db, the signal contributed by the lower half of a medium 

 0.6 mils thick cannot be less than 16.5 db lower than that contributed by 

 the upper half and hence the increase in thickness can do no more than to 

 raise the reproduced level by 1.2 db. 



At a lower frequency for which X = 100 mil, however, the corresponding 

 spacing loss is only 0.165 db and in this case the two halves of the tape can 

 contribute almost equally with the result that doubling the thickness of 

 the medium can almost double the reproduced signal voltage. 



Qualitatively, then, one might expect that increasing the thickness of 

 the recording medium, other things being equal, would increase the response 

 to low frequencies and leave the high frequency response relatively unal- 

 tered. This is in agreement with data published by Kornei.^ 



The estimates of magnitudes just given rest on assumptions which cannot 

 be proved except by further experiments. It has been implicitly assumed, 

 for example, that the medium is uniformly magnetized throughout its 

 thickness and this may not be the case. It does seem perfectly safe, how- 

 ever, to conclude that at a wavelength of one mil that part of the medium 

 which lies deeper than about 0.3 mil from the surface cannot contribute 

 appreciably to the reproduced signal. Furthermore, as the wavelength is 

 decreased beyond this point the thickness of the effective part of the tape 

 decreases in inverse proportion to X with the result that the available flux 

 also decreases. For this reason the "ideal" response curve cannot continue 

 indefinitely to rise at 6 db per octave as it does at low frequencies. In fact, 

 when the effective part of the tape becomes thin enough, the available flux 

 will decrease at 6 db per octave and just cancel the usual 6 db per octave 

 rise, giving an "ideal" response curve which rises 6 db per octave at low fre- 

 quencies but which eventually becomes flat, neither rising nor falling with 

 further increase in frequency. 



Spacing loss may contribute in still another way to the frequency response 

 characteristic of a magnetic recording system in which the reproducing 

 head makes contact with the medium. It is well known to those who work 



'Otto Kornei, "Frequency Response of Magnetic Recording," Electronics, p. 124, 

 August, 1947. 



