NEGATIVE IMPEDANCE TELEPHONE REPEATERS 



1071 



with a +180° phase shift instead of an admittance as used previously. 

 I'he letter A^ designates a numeric or i)i'op()rti()iiality constant. It will 

 be observed that the product of the shunt and series impedances is a 

 real or positive impedance and hence the image impedance is a positive 

 impedance, Zo . The gain is d(>termined entirely by the value of N. 

 Thus if the characteristic impedance of a transmission line is known, 

 together with the gain that the line can support without risk of oscil- 

 lation, then N is known and the repeater network can be adjusted to 

 give the required gain. 



The advantage of the bridged T as compared to a single series nega- 

 tive impedance such as the E2 can be demonstrated by comparing the 

 relative transmission gains obtainable from the two arrangements. Fig. 

 13(b) shows the insertion loss of a single impedance Za connected in 

 series with a transmission line having a characteristic impedance Zo . 

 If Za is a negative impedance such as that produced by the E2 repeater 

 then the repeater gain becomes a function of N as shown in Fig. 13(c). 

 If N equals 2 the gain is infinite and the system will oscillate. Thus N 

 must always be less than 2 where Za is a negative impedance of the 

 series or open circuit stable type. Practically, the impedance of the 

 transmission line is not a constant Zo but varies with termination, hue 

 construction and temperature. Thus A^ should be decreased until the 

 negative impedance is always less than the sum of the two line im- 

 pedances in series with it taking into account all possible variations in 

 these impedances. 



The same limitations on A^ apply to the bridged T repeater of Fig. 12 



Zi = IMAGE IMPEDANCE 



Zi= fzTzE 



ATTENUATION 



Zi IN DECIBELS 



-20 LOG,, 



Fig. 11 — Schematic of the bridged T network. 



Zo = IMAGE IMPEDANCE 



GAIN IN DB = 20 LOG, 



-^ 



Fig. 12 — Schematic of the bridged T repeater. 



