646 BELL SYSTEM TECHNICAL JOURNAL 



13 TU being observed at 20,000 cycles, probably due to the tuned 

 circuits in the receiver. Since modification of these circuits to obtain 

 a flatter characteristic would have been difficult and would have 

 occasioned a noteworthy sacrifice in selectivity, a compensation net- 

 work was designed for use in the 600-ohm output circuit of the receiver 

 which introduced a negligible loss at 20,000 cycles, a substantially 

 constant loss of 13 TU at frequencies below 2000 cycles, and for inter- 

 mediate frequencies losses represented by the height of the "normal 

 overall" curve above the horizontal line representing — 13 TU. 

 With this network connected between the receiver and the television 

 equipment, the average level throughout the band was, therefore, 

 reduced some 13 TU but the resulting characteristic as measured 

 beyond the network was that which has been designated "overall 

 equalized." Above 20,000 cycles the characteristics all fell very 

 rapidly which is an indication of the degree of selectivity attained. 

 This was contributed to by the radio-frequency tuned circuits, the 

 band-pass filters in the intermediate-frequency amplifier and the 

 20,000-cycle low-pass filter between the final detector and audio 

 amplifier. The individual characteristics of the various filters were 

 designed to be 60 TU down 20 kilocycles from the specified cut-off 

 frequency. 



Similar measurements were made upon the speech channel but a 

 less thorough study was deemed sufficient in that case due to the 

 existing background of experience. 



Effects of Fading 



With the system as outlined above, very satisfactory performance 

 was obtained during the afternoon and early evening hours when 

 reasonably stable transmission conditions were prevalent. Later at 

 night, however, when marked fading became evident, some rather un- 

 expected but easily explainable phenomena were observed which may 

 be of sufficient interest to warrant brief mention. 



When marked fading occurred, the normally clear reproduction 

 was accompanied by "ghosts" or additional images which faded in 

 and out in an erratic manner, sometimes appearing as positives and 

 sometimes as negatives. The effect was most clearly observed when 

 using one of the various types of test screens employed, a white card 

 bearing a black diamond-shaped outline, approximately a square with 

 its diagonals vertical and horizontal. With this simple type of 

 pattern, it became evident that the secondary images were additional 

 reproductions which were "out of frame" by a greater or less amount. 

 In other words, each of these additional images consisted of a portion 



I 



