November 21, 1884. 



SCIENCE 



477 



EDISON'S THREE-WIRE SYSTEM OF 

 DISTRIBUTION. 



The three-wire or multiplex system of distribut- 

 ing currents for electric lighting over large areas, 

 as devised and used by Edison, is highly ingenious, 

 and effective in reducing the necessary size of the 

 large copper conductors. The size of the conductor 

 must be proportioned to the maximum number of 

 lamps which it will ordinarily supply. This number 

 being given, the size should be such that the resist- 

 ance of the metallic part shall bear a fixed ratio to 

 that of the lamp part of the circuit ; and the value of 

 this ratio will be determined by the condition that the 

 additional running expense due to the resistance of 

 the conductor shall equal the interest on its first 

 cost, so far as this depends upon its cross-section. 



In the two-wire system, A is the dynamo, which 

 we will suppose to keep up a difference of potential 

 of a hundred volts between the conductors I and 

 II. Across these are bridged twelve lamps of equal 

 resistance, representing what would be, in practice, 

 several hundred dwellings, factories, churches, thea- 

 tres, etc. 



(S 



5- 



4 ,5 6 7 



9 io II 12. 



*© <) <> <> <> <> <> O <p <) o O 



TWO-WIRE SYSTEM. 



THREE-WIRE SYSTEM. 



T 



i 







4- 







7 







io 





>l~ 



<> 



2. 





o 



5" 





O 



8 





\ 



ii 



t®ut <-- 



> 



7 



3 





o 



6 





o 



9 





° I2_ 



r m l. 



--> 





o 







o 







o 





♦ 



FOUR-WIRE SYSTEM. 



The next figure shows Edison's three- wire modi- 

 fication of this. A and B are two dynamos coupled 

 in series, with conductors I, II, and III leading 

 out as shown. As A and B each keep up a hundred 

 volts, as before, the difference of potential between 



I and III will be two hundred volts. The twelve 

 lamps are now, however, equally divided between 

 the circuits I-II and II-III, connected as shown. 

 If the resistance of the six odd-numbered lamps, 1-11, 

 exactly equals that of the six even-numbered, 2-12, 

 and if A and B keep up the same difference of 

 potential, no current will flow in II, between the 

 dynamos and where the first lamp joins it. Suppose 



II to be cut, there will then be a single circuit 

 through A and B, I and III, and the twelve lamps, 

 as shown, with a difference of two hundred volts in I 

 and III. The resistance of the twelve lamps, as 

 now arranged, will be four times what it was before; 



and hence only one-half as much current will flow 

 through I and III. But each of the lamps will get 

 just as much as before, and will shine the same. The 

 conductors I and III, since the resistance is now 

 four times as great, need only be one-fourth as heavy, 

 according to our adopted principle. This is also 

 proper as regards heating-effect in them, which, pro- 

 portional to the square of the current, is now only 

 one-fourth what it was before. 



If this were all that was needed, we should now 

 have the same amount of lighting done, and the con- 

 ducting-mains, which are the most expensive part of 

 the plant, of only one-fourth their size and cost in 

 the two-wire system, and with only the additional 

 expense of another dynamo. Moreover, since the 

 current is only one-half as much, these two dynamos, 

 though giving the same potential as before, can be 

 smaller. But on account of the difficulty in keeping 

 an exact balance in the two sets of lamps, especially 

 about the time of lighting up at twilight, it is neces- 

 sary to introduce the third conductor from between 

 the two dynamos, and then neither circuit can be 

 exposed to a difference of potential greater than 

 either dynamo is generating. Also, if the balance is 

 not kept, a current through II, and a galvanometer, 

 shows on which side the lamp-resistance or the 

 dynamo-potential is in excess; and Edison restores 

 the balance by variable resistances in the circuits of 

 the field-magnets, or, in some cases, by bringing an 

 extra conductor from one or two large buildings, like 

 factories, theatres, etc., when near by, so that they 

 can, at will, be thrown into either circuit from the 

 central station. 



This middle wire need not, for most purposes, be 

 so large as the other two; but, in the case of a break- 

 down of I or III, it will have to do equal work with 

 the other, so that it is safer, simpler, and better to 

 make them all of the same size. The cost, then, of 

 conductors, is that of three wires, each of one-fourth 

 the section of the two in the first case, or f * £ = 

 .375, or a saving of sixty-two and a half per cent. 



The four-wire system shows a still further reduc- 

 tion of expense. The law on which this percentage 

 of economy proceeds, as far as cost of conductors is 

 concerned, may be shown as follows, in units of the 

 cost of the two-wire system : — 



For 2 wires, we have, f (t) 2 — 1.000 

 " 3 " " " f (^) 2 = .375 

 " 4 " " " -f W) 2 = .222 

 " 5 " " " I (|-) 2 = .156 

 " 6 " " " ±(i) 2 = .120 



A limit of economy or practicability, however, will 

 soon be reached in the increased number of dynamos, 

 the complexity of the system, and especially in keep- 

 ing up an approximate balance between so many 

 circuits. In practice, probably, the three-wire sys- 

 tem, with its saving of .625 of the cost of the two- 

 wire, will be found all-sufficient; except, perhaps, in 

 the case of a long main through a large scattering 

 district, when the four- or five-wire plan might be 

 preferable. 



One other advantage, available in all these systems 



