THE CORRELATOGRAPH 1181 



remains linear during the integrating interval. Such a long time constant 

 would carry over too much charge from previous intervals if continuous 

 integration were permitted so a shorting relay is provided to discharge 

 the condenser quickly to ground after the integrated value is sampled. 

 The sampling is done by a two-way clamp circuit with the timing pulse 

 generated from the trailing edge of the pulses which gate the switch 

 outputs. The shorting relay operates directly from the 60-cycle supply 

 and is of a type specifically designed to give a brief closure of about one 

 ms. every half period of the driving wave. The sampled outputs of the 

 integrator are applied to a balanced modulator to which is also applied 

 a 6-kc carrier. The resulting double sideband suppressed carrier wave is 

 amplified to form the marking voltage applied to the stylus. Instantan- 

 eous compression is incorporated in the marking amplifier to extend the 

 range of input magnitudes which are encompassed by the relatively 

 narrow recording range of the paper. The stylus responds equally to 

 positive and negative voltages and does not resolve the individual high 

 frequency oscillations. The result is like full wave rectification of the 

 correlation functions. 



Grateful acknowledgement is given to A. J. Rack, A. E. Johanson and 

 P. A. Reiling for suggested physical configurations and design informa- 

 tion suitable for the circuits which generate the various control pulses, 

 and which sample and hold the integrated outputs. G. W. Blake tested 

 and adjusted these circuits after they had been constructed by the wiring 

 shop. Performance runs were made by F. H. Tendick and N. K. 

 Poole. Also at various stages of the project assistance was given by W. 

 A. Klute, F. W. Kammerer and R. L. Carbrey. We have also received 

 helpful advice from many other associates too numerous for explicit 

 mention. 



The 200-tap delay line was constructed by the transmission networks 

 department. It consisted of one low^ pass filter section per tap with 

 mutual inductance between sections to maximize the linearity of the 

 phase curve. The taps were taken off high impedance shunts wdth the 

 output fraction tapered down the line to give constant loss along the 

 taps. C. E. Jakielski planned, assembled, tested, and adjusted the delay 

 line. The intricate task of connecting the 200 output taps to the cor- 

 responding 200 contacts of the switch was performed by M. Biazzo. 

 The delay line and switch appear in the photograph. Fig. 6. Additional 

 electronic apparatus not shown in this photograph was placed on inde- 

 pendent panels for experimental convenience but can be arranged in a 

 chassis in the same cabinet with the other apparatus, now that the 

 components have been determined. 



