July 7. 1910] 



NATURE 



25 



all sides, is rather a complicated phenomenon. By " phase 

 relation " we mean the position of the current with regard 

 to the voltage producing it. To understand what " phase 

 relation " means, let us take the analogy of a pendulum 

 in motion. 



The force keeping the pendulum swinging is a maxi- 

 mum at the end of each swing, while the greatest velocity 

 resulting from this force is at the middle of the swing ; 

 obviously the times of greatest speed and greatest force 

 are not coincident ; the one is out of phase with the other 

 by what mathematicians would determine, in the case of 

 the pendulum, as 90°, or a quarter period. 



Now the current leads the voltage at the sending end of 

 the cable by 45°. If a series condenser is introduced to 

 diminish distortion, it still further increases the lead, and 

 reduces the effective power into the cable. The effective 

 power can only be a maximum when the current and 

 voltage are exactly in step, or in ether words, when there 

 is no " phase relation." 



A receiving condenser is also harmful for the same 

 reason as a sending condenser. By abolishing the sending 

 condenser and replacing the receiving one by a magnetic 

 shunt placed across the suspended coil of the siphon re- 

 corder or relay in 1898, the speed and accuracy of signal- 

 ling were materially increased. 



A magnetic shunt, as employed on the cables, consists 

 of an insulated copper wire wound round a closed circuited 

 iron core. The resistance of the shunt is about 30 ohms ; 

 its inductance varies up to a maximum of from 20 to 40 

 henrys, and its weight from 1 to 3 cwt. In the case of 

 a siphon recorder used as the receiver, the shunt short- 

 circuits the suspended coil and the series condenser is 

 abolished. In the case of a cable relay, the series con- 

 denser is usually retained, to ensure that earth currents 

 are effectually stopped, but the condenser is made large. 



.A shunt inductance has a similar time action on the in- 

 coming current to that of a series condenser, but with this 

 improvement — that it helps to reduce the phase distortion 

 of current with voltage rather than accentuate it, as is 

 the case with the condenser. 



Having obtained the best value of the shunt alone, the 

 following curious effect was discovered : that adding a 

 condenser as an additional shunt, the size of the signals 

 on the recorder got larger and more distinct. The mathe- 

 matical reason for this is as follows : that for any par- 

 ticular frequency, say the highest frequency of the cable 

 signalling, the shunts of inductance and capacity when 

 properly proportioned act as a shunt of infinite resistance. 

 For frequencies much below this it is as if we had no 

 condenser at all. For frequencies much above this, it is 

 as if we had no inductance, but only a condenser. 



To reduce still further the harmful effect of phase dis- 

 placement, series inductances have latelv been introduced 

 at the ends of cables, particularly at the sending end. By 

 placing an inductive coil of low resistance in series with 

 the battery at the apex of the duplex bridge, not only has 

 the speed of signalling been increased, but the effect of 

 what is known as " jar " on the duplex balance has also 

 been greatly reduced. 



Before proceeding to describe the instruments that work 

 the cables, I will say a few words about "duplexing." 

 All cables are now duplexed, that is to say, are arranged 

 so that messages can be sent and received, at the same 

 time, at each end simultaneously. The first cables were 

 duplexed by Stearns, and later ones by Muirhead and 

 Taylor. Duplex reduces the speed of simplex, or of work- 

 ing one way only, by 20 per cent., but the total carrying 

 power of the cable, irrespective of direction, is raised by 

 some 70 per cent., and is for this reason valuable, and 

 repays the trouble in maintaining the balance. 

 _ Cables are duplexed by arranging an artificial or imita- 

 tion cable, which is an exact electrical copy of the real, 

 in parallel with the real cable. The current from the 

 sending battery flows through two equal arms of capacity 

 or inductance of a Wheatstone bridge arrangement and 

 into the real and artificial cables. 



The inductive or magnetic bridge which I have applied 

 lately is, I think, the best to emplov, because it gives in 

 practice higher speeds than any other form of 'bridge. 

 The receiving instrument is joined to the commencement 

 of the cables, and is thus not interfered with bv the send- 



NO. 



2123, VOL. 84] 



ing currents, because there is no tendency for the current 

 to flow one way or the other, the real and artificial cables 

 having exactly the same electrical properties and acting 

 on the sending current in the same way ; but the current 

 that is received flows only from the real cable, and is not 

 balanced by any from the artificial, so that the receiving 

 instrument is worked by it. 



When duplex is properly adjusted it is said to be in 

 balance, from its similarity to the adjustment of an 

 ordinary balance used for weighing goods. Take the 

 ordinary balance as an illustration of the electrical one. 

 Let one scale-pan represent the cable, the other the 

 artificial ; if equal weights are placed in each pan the 

 beam will not turn, but the beam will turn if, while equal 

 weights are or are not in the pan, a small weight is added 

 or placed on one pan. 



In the cable " duplex," the receiving instrument will 

 not be affected by the sending current, because the voltage 

 is always the same on each side of the instrument, but 

 will turn to indicate a signal when a voltage is received 

 or is added to or subtracted from the voltage already on 

 the cable side, due to a voltage being applied to the cable 

 at the far end. 



In Fig. 6 is shown the simplest diagram of a cable 

 "duplex," and Fig. 7 illustrates its mechanical equivalent; 

 the lettering is similarly related. 



ns of the balance 

 difference of voltage 



the pan ; C and AL a 

 pans of the balance. 



If the battery B sends equal currents into cable and 

 artificial line, as it should do if there is a perfect balance, 

 no current will flow through S, and thus the receiver S 

 is unaffected by the sending voltage ; or, if the pans of 

 the balance have equal weights B placed on them, the 

 indicator S will not move. On the contrary, if a voltage 

 is received from the cable C, this voltage is added to or 

 subtracted from whatever voltage may be in C at the 

 time, due to the sending battery, and thus there will be 

 a difference of potential across S, and the receiving instru- 

 ment will be worked from currents sent from the far end 

 of the cable, and from these currents only. 



In the mechanical analogy a small weight W is added 

 to or taken from one of two equal weights in the pans C 

 and .AL, and the beam will be tilted and will be moved by 

 this weight only however the weights B B are varied. 



The voltage of the battery as applied to the sending end 

 of a cable is very much greater than that received from 

 the cable to work the instrument, say in the relation of 

 40 volts to 1/20 volt in the case of a moderately long 

 cable, or as 800 is to i, and the sending and received 

 currents resulting from the same follow a similar propor- 

 tion. 



In the mechanical illustration I have therefore indicated 

 the weights B and W as squares having this proportion to 

 give a visual indication of what this means in the balance. 

 The proportion I have given is only the relation of the 

 sending voltage to that received. If the balance were out 



