1 68 



NA TURE 



[June 13, 1901 



ELECTRO-MA GNE TS. 



TN this article it will be shown what a great advantage results 

 from constructing electro-magnets on scientific principles, 

 instead of making them according to everyday notions, and to 

 give an idea which is the best form to adopt for producing very 

 strong magnetic fields. 



To understand the matter we must first consider the magnetic 

 circuit, which is very analogous with the simple electric circuit. 

 Just as when we have an electric current flowing in a copper 

 rod, say, we know that the current is flowing round a complete 

 circuit of which the rod forms only a part, so in the case of an 

 iron rod magnetised by a current flowing round it, we consider 

 that magnetism flows round a complete circuit of which the 

 iron rod forms only a part. Fig. i is an ordinary bar electro- 

 magnet ; and in Fig. 2 B represents a cell and c A a copper rod 

 I he ends of which are joined by a great many thin wires of high 

 resistance. 



Now the flow of current in Fig. 2 is exactly analogous with 

 the flow of magnetism in Fig. I. The cell replaces the coil of 



wire on the magnet, for whereas the cell sends the current in 

 Fig. 2, the current flowing through the coil sends the magnetism 

 in Fig. I through the iron rod, corresponding with the copper 

 rod, and to complete the magnetic circuit the magnetism passes 

 through the air in paths shown by the dotted lines, Fig. i, back 

 to the south pole of the magnet ; so that the magnetic circuit in 

 this case is formed partly of iron and partly of air. 



The current flowing in the coils of wire on the magnet pro- 

 duces what is called a " magnetomotive force," which is pro- 

 portional to the current and to the number of turns of wire ; 

 and a certain fraction of this quantity is used to send the mag- 

 netism through the iron rod and the remainder to send it 

 through the air, or, in other words, every little piece of the 

 magnetic circuit requires a certain magnetomotive force to drive 

 the magnetism through it, and the sum of all these, taken all 

 round the circuit, is the whole magnetomotive force due to the 

 current in the coils ; just as a certain part of the electromotive 

 force of the cell is used to send the current through the copper 

 rod and the remainder to send it through the thin wires forming 

 the rest of the circuit. In fact, even the law governing the pro- 

 duction of magnetism in a magnetic circuit is very similar to 

 Ohm's law for the flow of current in an electric circuit, namely, 

 that the amount of magnetism produced is equal to the magneto- 

 motive force producing the magnetism, divided by the magnetic 

 resistance, or " reluctance," as it is called, of the entire magnetic 

 circuit. Hence, if the amount of magnetism is to be as large as 

 possible it is just as important that the reluctance of the entire 

 circuit should be small as it is that the current and number of 

 turns of wire be large. 



Now the reluctance of any little bit of a magnetic circuit, say 

 from S to N, Fig. i, for example, is proportional to the length 

 of the piece, inversely proportional to its cross section, and also 

 inversely proportional to the magnetic conductivity, called per- 

 meability, of the material. Therefore, to make the reluctance of 

 our circuit small, we have to make : (i) its length small, (2) its 

 cross section large, (3) and make it of a material whose perme- 

 ability is as large as possible. 



But the important thing is that the reluctance of the whole of 

 the magnetic circuit must be small, not only of any particular 

 part of it. F"or example, in Fig. I, making the diameter of the 

 iron bar large simply makes the reluctance of the circuit from 

 S to N small, while the reluctance of the rest of the circuit, 

 from N through the air to S is still very large, because the per- 

 meability of air is very small compared with ihat of iron. But 



NO. 1650, VOL. 64] 



if the bar is bent round into a ring, Fig. 3, then the reluctance 

 of the whole circuit is reduced, and consequently a larger 

 amount of magnetism will be produced in the bar for the same 

 current flowing round it, and the density of the flow — that is, 

 the strength of the magnetic field in the air space — will be very 

 much increased. 



There seems to be a popular idea that if a magnet is to pro- 

 duce as strong a field as possible it must be wound with an 

 enormous number of turns of wire and a very large current sent 

 round the coils, and that nothing else is of the least consequence. 

 The following is a description of a large electro-magnet, made 

 only about three years ago, which well illustrates this. It 

 formed part of an electrical instrument intended to be used in 

 connection with submarine telegraphy, the sole function of the 

 magnet being to produce a very strong magnetic field. This 

 result was certainly not obtained because it was not properly 

 designed. 



The magnet consists of two iron cores, 6 centimetres in 

 diameter, each wound with about 1500 turns of wire, making 

 the outside diameter more than 16 centimetres. To illustrate the 



Fig. 3. 



uselessness of this great amount of wire compared with the cross 

 section of the iron, it has been found by experiment that if only 

 one-third of the ordinary current is sent round the coils the 

 strength of the magnetic field is thereby reduced only 15 per 

 cent. If therefore the magnet had been wound with one-third 

 the number of turns the cost of materials would have been 

 about halved, and the power used to excite it only one-third, 

 and even a less reduction than 15 per cent, in the strength of 

 the field would have resulted. 



The cross section of the piece of iron joining the two cores, 

 i.e. " the yoke," is less than half that of the cores, and conse- 

 quently the density of the flow of magnetism in the yoke is very 

 large, and this means that the yoke will offer a great resistance to 

 the magnetism for two reasons: (l) because the area is small, 

 (2) Because the density being so large the permeability will 

 be very small, for the magnetic conductivity gets rapidly less the 

 greater the density ; in fact, in this case the reluctance is so 

 large that when the magnet is excited with its ordinary working 

 current it produces a field of only 7900 C.G.S. units, and it can 

 be calculated that the magnetomotive force used to send the 

 magnetism through the yoke is then more than four times that 

 required for the air gap, whereas in a properly designed 

 magnet nearly all the magnetomotive force is used to send the 

 magnetism through air gap and pole pieces. Doing what has 

 been done here is exactly analogous with trying to send the 

 strongest current that you can through an electrical apparatus 

 by connecting to it the nfiost powerful battery obtainable with 

 two very long thin high- resistance wires. Analogy, therefore, 

 shows us that the cross section of the yoke should have been 

 made at least equal to that of the cores. 



In order to see what sort of saving might have been effected, 

 I have designed a magnet (Fig. 4) to produce the same effect as 

 this one. 



It consists of a cast steel ring, rectangular in section, the wire 

 being wound on ten bobbins made of thin wrought iron, and 

 not straight on the ring, for convenience in winding. 



The design is made by starting with the assumption that a 

 magnetic field exists of the strength desired in an air gap of the 

 dimensions of the last magnet. Then the flow of magnetism at 

 the section a a, Fig. 4, is calculated, ditto for section bb, where 

 it is greater than at a«, by the amount which leaks out of the 

 iron between these two sections. Similarly, the flow is ob- 

 tained at all the sections, cc, dd, &.C., round the circuit, the 

 area of the iron at all these sections being made such that the 

 density of flow has a value fur which the magnetic conductivity 



