402 BELL SYSTEM TECIIMCAL JOLRXAL 



The voltage across this condenser is compared with a reference voltage. The 

 magnitude of this reference voltage corresponds to the d^c i)ulse amplitude 

 of Fig. 1. The voltage has a magnitude of 16 units. If the magnitude of 

 the condenser voltage exceeds the magnitude of the 16-unit voltage, a 

 positive pedestal voltage is obtained in the output of the comparing circuit. 

 This pedestal voltage is amplified, limited and applied to the pedestal 

 modulator. The pedestal modulator serves as a gate for timing pulses from 

 the timing pip generator. If the pedestal voltage and timing pulse are 

 applied simultaneously to the pedestal modulator, a pulse is obtained in the 

 output. In the jjresent case this pulse corresponds to the presence of the 

 16-unit digit in the code group which represents this PAM pulse. This digit 

 pulse after amplification and limiting is (1) sent out over the line (PCM out) 

 and (2) fed back through a suitable delay circuit to a subtraction circuit. 

 The function of the subtraction circuit is to subtract a charge from the con- 

 denser corresponding to the 16-unit digit. The charge remaining on the 

 condenser is now compared with a new reference voltage which is h the 

 magnitude of the first reference voltage or 8 units. If the magnitude of the 

 voltage across the condenser exceeds this new reference voltage the above 

 process is rei)eated and the second digit pulse is transmitted and another 

 charge, this time corresponding to the 8-unit digit, is subtracted from the 

 remaining charge upon the condenser. 



If the magnitude of the voltage across the condenser is less than the 

 reference voltage, in either case above, then no pedestal will be produced and 

 no digit pulse be transmitted. Since no pulse is transmitted, no charge 

 will be subtracted from the condenser. Thus the charge remaining 

 upon the condenser after each operation represents the part of the orig- 

 inal PAM pulse remaining to be coded. The reference voltage wave 

 consists of a series of voltages each of which is ^ of the preceeding one. 

 There is one step on the reference voltage function for each digit to be 

 coded. 



A better understanding of the coding process can be had by reference to 

 the various wave forms involved. For completeness, wave forms from 

 audio input to the coded pulse signal are shown for the transmitter in Figs. 

 ^ and 5 and from the coded pulse signal to audio output for the receiver in 

 Figs. 7 and 3. In the diagram the abscissas are time and the ordinates are 

 amplitudes. Some of these wave forms have already been discussed in 

 connection with Fig. .^. Since the coder functions in the same manner for 

 each PAM pulse the detailed wave forms of the coding and decoding proc- 

 esses are shown for only two amplitudes. The block schematic for the 

 transmitter is given on Fig. 4, while that for the receiver is given in Fig. 6. 

 The letters on Figs. 4 and 6 refer to the wave forms on Fig. 3, while the 

 numbers refer to the wave forms in Figs. 5 and 7. 



