EXPERIMENTAL TRANSVERSAL EQUALIZER FOR TD-2 1447 



from reflections on the delay line occurs just above 80 mc and distorts 

 each characteristic near this frequency. The delay characteristics obtained 

 closely resemble the corresponding gain characteristics. The delay char- 

 acteristic with all rheostats on midrange was flat to within about ±1.5 

 millimicroseconds. 



Another measure of the performance is the amount of interaction 

 between gain and delay characteristics. As shown on Fig. 13 the gain 

 changes less than 0.2 db when the third or fifth harmonic delay terms 

 are set at their maximum values of 11 and 12 millimicroseconds, respec- 

 tively. Similarly, when a pair of rheostats are set for the maximum gain 

 characteristic, the effect on delay is of the order of two millimicroseconds. 

 These results, which are typical, indicate the interaction effect is of the 

 order of 20 per cent using one neper of gain distortion ripple as equiva- 

 lent to a ripple of one radian of phase shift amplitude. The effect of this 

 interaction on the field use of the equalizer is to require a second round 

 of equalization to correct for interactions after the gross distortions in 

 a circuit have been equalized. 



XII. FIELD EXPERIMENTS 



Models of this equalizer were installed in two channels of the TD-2 

 system between Denver, Colorado, and Omaha, Nebraska, early in 

 1956. This route is about 500 miles long and includes 18 microwave links. 

 One equalizer was installed at the center of the route and a second one 

 at the receiving end. The results of a tj^pical set of characteristics ob- 

 tained are shown in Figs. 14 and 15. The first shows the delay charac- 

 teristic of the whole channel measured at the receiving intermediate 



60 MC 80 MC 60 MC 80 MC 



mam- mmm 



(a) UNEQUALIZED (b) EQUALIZED 



Fig. 14 — ^Measurements on Denver-Omaha route. Envelope delay distortion 

 measured at intermediate frequency point. Left, unequalized; right with equalizer 

 adjusted. 



