METHODS OF RECORDING 



recording, which uses a non-consumable medium and which can be edited 

 so that only sections of importance are finally photographed. Magnetic 

 tape is a very suitable and convenient possibility. As conventionally used, 

 tape recorders operate in the audio range, about 50-10,000 c/s, and are 

 therefore satisfactory for storing patterns of action potential discharges. 

 The theory of tape recording has been discussed by Daniel and Axon*, and 

 practical requirements for recording machines by Carter^. An ordinary 

 commercial tape recorder was used for spike-potentials by Leithead and 

 Thompson^, but distortion of the spike on 'playing back' into an oscilloscope 

 is usual. This may be explained in a simplified analysis as follows: 



Conventional magnetic recording uses the 'constant current' method, 

 where the recording head causes a magnetization of the tape passing it 

 proportional to the current flowing through its coils and, since it is fed from 

 an amplifier having a high output impedance, proportional to the instan- 

 taneous value of the input signal, e. In other words, back e.m.f 's. occurring 

 across the recording head are swamped and 



^ = ke 



When the tape passes under the replay head an e.m.f. is induced proportional 

 to the rate of change of magnetization 



M 



dt 



e' = k' 



Thus if e == £■ sin ojt, e = kk'ojE cos cot, that is, its amplitude rises pro- 

 portional to frequency for constant E. Thus the signal from a magnetic 

 playback coil rises at 6 dB/octave. In practice it does this up to a limiting 

 frequency at which the whole system breaks down because the playback 

 head gap width becomes comparable with the wavelength of tape mag- 

 netization {Figure 40.5). This kind of frequency response can be compensated 



Log output 



on playback 



for constant 



recorded 



amplitude 



6 dB/octave 

 rise 



Rapid fall 



Log frequency 

 Figure 40.5 



for — by suitable filters in the replay amplifier — satisfactorily for audio 

 work, but in commercial machines the compensation is not normally 

 sufficiently exact for the overall frequency (and therefore phase) distortion 

 not to mar the reproduction of waveforms. 



To avoid this difficulty Coaton and Whitfield' employed constant-voltage 

 recording, driving the recording head from an amplifier of very low output 

 impedance. Then the back e.m.f. across the head is proportional to the 



641 



