'8 



THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1956 



iWWWMMMWWIWWMilJflM ^ I I . rlil l lT- i \m....: i n i T i inr- IH. 



Fig. 10 — The regeneration of band-limited pulses. A — Input to regenerator, 

 first two groups. B — Output of regenerator, first two groups. C — Output of 

 regenerator, 24th trip. 



amplitude with each trip around the loop indicating that loop gain was 

 slightly greater than unity. Without the sheer it is difficult to set the 

 gain to exactly unity and the amplitude tends to either increase or de- : 

 crease depending upon whether the gain is greater or less than unity. 

 Results indicated by the pictures of Fig 9 are possibly not typical of a 

 properly functioning system but do show what happened in this par- . 

 ticular sj^stem when regeneration was dispensed with. 



Another important function of regeneration is that of overcoming . 

 band-limiting effects. Figs. 10 and 11 show what can be accomplished. . 

 For this experiment the pulse groups inserted into the loop were as shown i| 

 at the left in Fig. lOA. These pulses were 15 milli-microseconds wide at 

 the base and spaced by 25 milli-microseconds which corresponds to a j 

 repetition frequency of 40 mc. After passing through a band-pass filter 

 these pulses were distorted to the extent shown at the right in Fig. lOA. 

 From the characteristic of the filter, as shown on Fig. 12, it is seen that 

 the bandwidth employed is not very different from the theoretical min- 

 imum required for double sideband transmission. This minimum char- 

 acteristic is shown by the dashed lines on Fig. 12. Fig. lOB shows that 

 at the output of the regenerator the effects of band limiting have been 

 removed. This is borne out by Fig. IOC which shows that after 24 trips 

 the code group was still practically perfect. It should l)e pointed out 

 that the pulses traversed the filter once for each trip around the loop, 



