430 BELL SYSTEM TECHNICAL JOURNAL 



It will readily be seen that with neutralizers applied at each end 

 of the four circuits, transmission of signals on one of them will gen- 

 erate the proper impulses for neutralizing both sending and receiving- 

 end crossfire from that circuit into the other three. Furthermore, 

 neutralization will take place with all wires operating simultaneously 

 in either or both directions. 



Application to Different Circuits 



It is in general not practicable to compute the constants of the 

 neutralizing devices, but this is unnecessary since it is an easy matter 

 to determine them experimentally. In making trials to determine 

 the proper amount of capacity and inductance required to neu- 

 tralize crossfire effectively, it is fortunately possible to design the 

 various parts independently of each other to a considerable extent. 

 For example, the diagonals of the six-condenser network may be 

 determined after the condensers in the sides have been approximated 

 very roughly, or vice-versa; likewise, the amount of inductance 

 required between each pair of circuits may be approximated inde- 

 pendently, but if a single coil is to be used for coupling more than 

 two circuits, all the circuits should be connected up in making the 

 test. Sending and receiving-end crossfire may, of course, be treated 

 separately. 



It is convenient to vary the capacity of the condensers, but not 

 usually the inductance of the transformer. In the latter case a coil 

 with excess mutual inductance may be used together with a variable 

 resistance shunt. 



In order to design neutralizing arrangements or to determine 

 whether or not they are effective, tests may readily be made by 

 observing the deflection of a milliammeter connected in series with 

 the polar relay of a bridge polar-duplex set while signals are sent on 

 the parallel circuit. In a similar way a differential meter may be 

 used in a differential duplex set. A somewhat more accurate test 

 may be made by observing the response of the receiving relay, pref- 

 erably with variable electrical bias. The disturbing signals are of 

 course sent from the same station in checking sending-end crossfire 

 and from the distant station in checking receiving-end crossfire. 



Representative anti-crossfire capacity values are given in the fol- 

 lowing table for No. 8 B.W.G. (0.165 in., 2.5 mm.) composited open- 

 wire copper circuits, 300 to 500 miles (500 to 800 km.) in length and 

 No. 12 A.W.G. (0.104 in., 1.5 mm.) circuits 150 to 300 miles (250-550 

 km.) in length. No timing resistance is required usually. In practice 

 there are material variations from one circuit to another. 



