]\gt ; INDEX SLIP .
HBIR-SItAW , H. and HAX , Alfred.\mdash ; Lines of Induction in a Magnetic Field .
Phil. Trans. , , vol. 195 , , pp. 303-327 .
, Alfred , and HELI ?
-SIIAW , H. S.\mdash ; Lines of Induction in a Magnetic Field .
Phil. rans .
, vol. 196 , 1900 , pp. 303-827 Magnetic Induotion in Cylinders and Shells : Mathematical ano Graphical Treatment of .
-SHAW , H. S. , and HAY , Alfred .
Phil. Tralls .
, , vol. 196 , 1900 , pp. Magnetic Lines marked out by Flow of Liquid .
-SHAW , H. S. , and HAI , Alfred .
Phil. Trans. , , vol. 195 , 1900 , pp. 803-327 .
0.$ 1 IX .
Lines of Imluction in ) Profcssor H. S. -SHAW , .D .
, , B. Rcceived June 13 , \mdash ; Read June 21 , 1900 .
following } ) aper , which is partly experimental and partly mathematical , has arisen from the discovery that two-dimensional cases of magnetic lines of force could apparently be represented by the flow of a viscous liquid .
* The original experiments upon which this assumption was made , showed that the stream lines which were obtained by the method in question , gave results very similar to those which had been calculated and plotted for the cases of an elliptical and circular cylinder .
In order to ascertain definitely that the stream lines under these circumstances actually gave the exact position and direction of the corresponding magnetic lines of force , a result which , if verified , could be used for many practical investigations\mdash ; it was necessary to undertake a long research dealing with the various points involved , a resea.rch which has proved extremely laborious , extending without intermission over a period of nearly two years .
In the first place it was necessary to devise some method by which a thin sheet of transparent or semi-transparent medium could be obtained of any lequired thickness , and on which , when placed between two sheets of glass , the required section of the body to be investigated could be formed .
Next it was necessary to determine the laws connecting the thickness of the thin film of liquid with the quantity flowing through it in a given time , so that the relative differences of thickness corresponding to the differences of permeability of the substances in a magnetic field could be ascertained .
Lastly , a mathematicainvestigation was undertaken of some cases suitably selected so as to afford , when plotted out , as severe a test as possible for if the experimental method really determined for any case , accurately , the position and character of the lines of force in a magnetic field .
It may at once be stated that after overcoming in succession a very large number of difficulties , the case selected , viz. , that of an ellipse with the major axis parallel " " Stream-line Motion of a Viscous Film\ldquo ; : 'British tiott Peport ' ( Section A ) , Bristol 1Ieeting , 1898 .
20.12 .
1900 SOIi H HAVE to the lines of force in the undisturbed magnetic field , showed unmistakably absolute agreement between the results obtained by calculation and experiment .
On account of the great practical importance of magnetic phenomena , various methods have from time to time been devised for the study of magnetic fields , and by means of such methods the approximate distribution of the lines of magnetic induction in various cases may be investigated .
One of the earliest of these methods was that in which the lines are approximately mapped out by means of iron filings .
This method , which has been known for a long time ( LA HIRE mentions it as far back as 1717 ) , was very largely used by FARADAY in his study of magnetic fields .
It must , however , be regarded as a very rough-and-ready method , and does not enable us to trace out the exact shape of the lines with any degree of accuracy .
Another method , which when carefully applied is capable of giving good results , is that in which a magnetic needle is moved from point to point of the field , the consecutive positions of the two ends of the needle being marked , and a line drawn through the points so determined .
This method gives the shape of the lines very ccurately if the length of the needle is small in comparison with the radius of curvature of the lines .
But although in a diagram so obtained the relative intensity of the field at various points may be roughly estimated by noting the conyergence or divergence of the lines , yet it is impossible to indicate this variation of intensity by the distance apart of consecutive lines , since there is no means of ascertaining how far apart the consecutive lines should be drawn .
A third method , which we owe to the genius of FARADAY , and which of very great practical importance , is that of a search-coil connected to a ballistic galvanometer .
In applying this method , the search-ccil is eitherjerked out of a given position in the field , or turned through a small angle , or else the field is suddenly removed , or reversed , while the search-coil remains stationary .
This method , however , notwithstanding its great importance as a method for ascertaining the magnetic flux through a given area in the field , can hardly be regarded as a method for delineating lines of induction .
The remark is applicable to two other methods ; that in which the field intensity is ascertained by measuring the resistance of a bismuth spiral ; and that in which a conductor conveying a known current is placed in the field , and the pull on the conductor is measured .
The experimental method which the above inyestigation proves to be accurate , is applicable to two-dimensional problems only .
In so far as this is the case , its scope may appear restricted .
But when we consider the fact that in the practical applica- tions of netism , the bulk of the phenomena with which we have to deal are of the two-dimensional order , it will be recognised that it offers a solution of problems whose interest and importance are by no means inconsiderable .
The magnetic field in and around the armature of a dynamo or alternator , and that in and around cylindrical case which is used for oses of magnetic shielding , are examples of important practical problems in two dimensions .
ON LINES OF ) UCTION i ) .
.305 very } ) ortant feature cterises the lnetlod , is the , that it enables us to the shape of the lines of induction not only in , but or nagnetic material itselJ .
This nnot be ccomplished by of the older nethods .
The method been to determine the exact lines duction f lmmbel of of interest , several of which been also ) lotted 1tion .
slnuch connecting the relative thicknesses of two liquid fihns with the ) lues of the permeability , by the above iuvestigatio been determined , it is possible to inyestigate practical case which may be of interest , and examples giyen of the application of lnethod in of intel.est The mathematical portion of the work which been necessary , 11its not the authors \mdash ; hitherto been available in published will be seen the historical statement , } of the ) have ) dealt various , it ) he- account given will found of use in the ftlrtheltion of this ) ject .
The schenle of the .
is as follows:\mdash ; I. The of the ickness film the of flow .
tenlent of results ison Description of used in the 11 .
Account of the test case out ) olled , it , s rison w the actual ) obtained by ) ( b ) ples of of interest , OlIlG o to the electrical 111 .
Brief ] of of the slllject , ( b ) A general letllud of dealing elliptic con-tocal shells .
SECTION I. ( liquid is contained in the cylindel , fig. hich it is ) of , the pressure being by B. lt then the cock fine djustment I by the level ] ) which over a scale ) to the therlnometel box , its ) ture is recorded the thermometer , and from thence to the slide J. The slide consists of two sheets of plate-glass fixed within the il well cut in lower sheet into which the liquid is oduced , channel formed between the by using tin-foil as a border .
The thickness is ined by screwing down the plates the thiokness gauges ) tho V0L .
of in th low , found meant4 lnbc ; the ) for Determining the Viscosity of a ) the method of a liquid is in the box it ]Jasses tough the spout , into the vessel , where it is collected .
When viscous , the liquid was to flow for hxed time into , and then rhed ; when more mobile , it was collected intu it arger of known city , the time it took to fill vesscl a tely made ) tables the esults uf the renlents : ON . .
on Fl of Thin ) veen P ( 1 ) lbs. pressure .
Amount in grammes flowing in 1 Time ffir 1000 gramnles to flow .
Velocity in centilns .
) minute .
017 9 .
xPERIMENTS on Flow of Thin between Parallel Plates .
( 2 ) Water at 21 lbs. pressure .
Amount in grammes flowing through in 10 minutes .
Time for 3850 grammes to flow .
Velocity in centims. per minute .
ON OF INDUCTION IN A Fig. 2 , NOT ER .
Quaentity for time .
for Jantity . .
Quantity for rime for .310 -SHAW ANY ) ) these tables the curves shown in figs. ) ) dralvn , and these ryeR w actually used to determine the relation between different depths of well corresponding perlneabilities .
authors are indebted to Mr. J. C. W. HUMFREY , B.Sc. , Victoria University Scholar in Engineering , for assistance in carrying out the above experiments .
After the results embodied in the curves , figs. 2 and 3 , had been obtained , we thought it would be interesting to see in how far they were in agreement with the deductions from theory .
Let thickness of liquid film .
tangential stress ( per unit area ) due to viscosity at a distance . .
from the middle layer .
velocity of flow at distance from middle layer .
coefficient of viscosity .
Then At a point distant frnm the middle layer the tangential stress will be Hence if we consider a layer of unit length , unit width and thickness Sx , the to lvhich this elementary layer is subject on account of viscosity is given by Sx .
Now , since we suppose that the flow is steady , the elementary layer considered is not undergoing any acceleration .
Hence the backward drag due to viscous resistance must be balanced by the forward push due to the difference of pressure over the ends of the elementary layer .
stand for the gradient of pressure , then S , or , say , where is the fall of pressure per unit length of liquid layer .
rating this equation once , we get whereC is a constant .
Since when ( the velocity in the middle being at its maximum ) , .
Integrating a second time , we get OF . .
to find notice when , the velocity the boundary gives The volume flowing through ' second ) unit width 01 the layer of thickness is ' rate of flow of viscous liquid in a thin layer veen 1 ) plane valls it ; thus , a given of ) the layer , seen be ) to the cube of the of the layer .
An of the obtained } ) shows eenlent with contained iu the curves also enable us to calculate tely the coefficient of viscosity .
atus is not , howevel , well ) for exact enleuts , since , in addition to the difficulty in ately t thickness of the liquid film , it doubtful if the glass ) lates itre eithel suflicielltly true or sufliciently rigid ( their great tbickness ) for refined lnents .
Since the lnula ' for the coeflicient of viscosity involves the cube of the thickness of the liquid .
comparatively small in the determination of will give rise to in .
Taking the water curve ( fig. 2 ) , we find for the coeiiicient of viscosity in C.G.S. units ( using the point on the curye esponding to thickness of ) the value 0092 , a value which is considel.ably too low , which may be ccounted f either by an error iu estimating the thickne ; s , or by slight egularities in the containing walls , or slight bulging of the walls in the of which causes colnbined easily account the The glycerine curve ( fig. 3 ) gives a value of the viscosity equal to ; the density of the glyceriue being .
This value is ent with the results ) previous As , howevel , our )ject w not to out careful bsolute 1 , but tu find relative yalues which would enable us to ) lytheluethod to in laguetiC induction , and , the curves were * It must also ) reluembered that the used in these ) had not been freed from dissolved gases.\mdash ; Jnly 18 , Seo p. 131 of ' HAW AND HAY to be ig.eelnent with the theoretical law , we did not pursue subject ) viscosity nts any further .
A of the general lnethod of providing the flow of coloul bands in fihn eady been givell in ) aper 1 the Institution of , for bas beeIlconsidel.ablyiml ) oved ified , shows in figs. 4 , 5 , , the combination of glass )lates which been devised in COlmectiol l the ) esent papel .
It will be seen that instead of having parallel plates of glass , only two cire used , in o1le of \mdash ; made very thick\mdash ; two wells used for the clear and coloured uids respectively .
The figures show the construction of the apparatus , and it may that with this contrivance it is very easy to use a bordel of ) , cut to employ the opposite plate of glass , the cover-plate , .
the purpose roviding the well essio through which the flows .
The of this well is made to correspond to the nagnetic pmeability , the outline of the well having the form of a cross-section of the body under investiation .
How this well or depression was obtained must now be described , as the ditJiculties expel.ienced in this case proved much greater than when solid obstacle required , as in the ) above mentioned .
In .
to solve the ) oblem , it necessary to obtain il material could be worked to any shal ) , and ) ecenlented to glass of the slide to the ecess o well , ) ondintr iu to the body the lines of induction were required .
Originally , for simple circular forms , sheet of glass such as is used microscopical work , was cemented to the covel , but this was not , as the glass could not eadily be obtained of the thickness required .
Ful .
there , the thickness of the layer of cement between the endered very diflicult to secure , any figul.es othel could not easily be cut in glass .
Celluloid had the advantage that it could be worked to ; but again the difficulty arose that it could not be obt ined of the thicknesses for esistances , and eovel .
could not be cemented to cover plate so as to lie pel.fectly flat .
ious p and tracing cloths , but they all less altel.ed in thickness when moistened , and it tely f that any lnaterial requiring to be cemented to the glass could not used unless it was capable of being brought to true surface afterwards , as 110 cement found ] ) able of holding sufficiently for the object in view .
Plates of white lnetal were of , but the .
involved in them , and the difficulty of hoto 1 of the stream-lines , led to the abandonment of this idea also .
lethod finally adopted was to cast a layer of paraflin wax on the coyer-plate of glass , and lthough this was not transparent enough to enable the stream-lines " " of the of Surface , of Water of Motion under xperiment Couditions 'Trans .
Naval tects , ' vol. 40 , 1898 .
ON LINLS OF INDUCTtON Fig. 4 .
Side view .
Fig. 5 .
view .
Fig. 6 .
l'ront view .
IN A IC FIELD .
Arrangement of Glass Plates and VOL. CXCV .
\mdash ; A. 2 :$14 JILLL-SHAW ANY ) .
ALFRED HAY to be rojected on the screen , it allowed , when illuminated from behind , sufficient light to pass to enable the stream-lines to be photographed .
At first the difficulties of using paraffin wax were numerous , for on casting the wax and allowing it to cool , it became covel.ed with star-like cracks beneath the surface , which were evidently caused by unequal heating cooling , and imprisoned air bubbles .
Casting in a acuum was also tried , but without success .
The difficulty was at length overcome by a systematic method of covering the plates , which was as follows : The plate on which the layer of wax was required , was placed in an iron tray , the wax was shredded on to the plate , and the whole heated gradually and uniformly , a plate of iron 1 inch thick bein inserted between the tray and the Bunsen burner used for heating .
By this means the plate could be covered to any required thickness in a satisfactory manner .
In order to reduce the layer of wax to the thickness required the experiment , the simple machine illustrated in figs. 7 and 8 , was used : Plan .
Planing Instrument for Paraffin Surface .
the gauges ( AA ) for obtaining the required thickness being shown in position .
The llilled-headed screws enabled the cutter to be adjusted so that the desired thickness of wax could be accurately obtained of any dimensions above inch to within inch with the greatest ease .
The method ofilluminating the slide had , however , , be abandoned , as the heat of the arc lamp the at once putting an end ON LINES OF INDUCTION IN A MAGNETIC FIELD .
.315 to the experiment .
Instead of this , the slide was taken to a place where it could be illuminated by strong daylight , and the exposure given to the sensitive photographic plate was of course correspondingly increased .
Zinc templates of the shapes required were made , and by placing the template on the wax , the outline was cut with a sharp knife , and the particular well thus formed .
In cases where the flow in the well was very slow , air bubbles became imprisoned , and could only be removed after great difficulty ; in fact , in one or two cases\mdash ; as will be seen in the photographs , 31 and 32 ( Plate )\mdash ; they could not be removed at all .
SECTTON II .
In order to test the applicability of the stream-line method to the solution of two-dimensional magnetic problems , it was decided to work out mathematically the case of an infinite cylinder of elliptic section in an originally uniform nlagnPtic field with its major axis along the field , to plot the of magnetic corresponding to a permeability of 100 , and to compare the diagram so obtained with a stream-line diagram .
The theoretical diagram is given in fig. 9 ( Plate 14 ) and the corresponding stream-line diagram in fig. 10 .
In making the com- parison , a greatly enlarged photograph of fig. 10 was prepared , and fig. which was actually drawn to a much larger scale , was then superposed on it : the coincidence of the lines in the two diagrams satisfactorily established the soundness of the stream-line method .
It may be noted , howeyer , that slight local divergences along the elliptic boundary are clearly observable .
Instead of actio of the lines as they enter the elliptic cylinder in the diagram , we have in fig. 10 a slight curvature at the ends of the otherwise perfectly straight lines crossing the ellipse .
This feature is noticeable , to a greater or smaller extent , in all the streamline diagrams accompanying the present paper .
It is more marked in those cases where the difference etween the thicknesses of the two liquid layers , , the permeability in the corresponding netic problem , is greater .
It is clear that the presence of the highly permeable cylinder disturbs the originally unifor1n distribution of the lines , and in order to effect a satisfactory comparison between the theoretical and the experimental diagrams , it became necessary to assign to the liquid layer an external boundary whose shape corresponded to that of a particular stream-line in the theoretical diagram .
The boundary of the liquid in fig. 10 is clearly shown by the dark shadows on either side of the diagram , and the shape of their boundaries is the same as that of the outer stream-lines in fig. 9 .
The method by which the solution was obtained for the theoretical case will be found fully explained in the mathematical appendix to the present paper .
( b ) In most of the succeeding diagrams the boundaries of the liquid layer are straight lines .
Interpreted netically , this means that the diagram giyes the * For see Mathematical Section of this solution not for a single cylinder of the given cross-section laced in a magnetic field , but of a whole row or grating of such cylinders , the distance apart of any two neighbouring cylinders being equal to the breadth of the liquid film .
Fig. 11 gives the solution for an elliptic cylinder similar to that in fig. 10 , but having a permeability of 20 .
In fig. 12 we again have a cylinder of the same cross-section , but of permeability equal to 1000 .
On account of the great depth of the elliptic wall , the lines crossing the ellipse appear to be almost entirely obliterated ; notwithstanding this fact , they emerge on the other side , preserving their identity and not mixing with the body of the liquid .
The distortion of the lines close to and on the inner side of the elliptic boundary is in this case very strongly marked , and well illustrates one of the difficulties encountered in attempting to imitate the effects of highly permeable bodies .
An interesting feature , well-known as a result of theoretical deductions , is clearly brought out by a compariso1l of the three diagrams , , the gradual decrease in the angle made by the external lines with the normal to the ellipse as the permeability is increased .
Thus in fig. 11 this angle of incidence of the lines is quite large at certain points of the elliptic boundary ; it is greatly reduced in fig. 10 , and in fig. 12 , for which the permeability is 1000 , it is practically zero .
It will be seen that the refraction of the lines in fig. 11 is very sharp \mdash ; the " " weir effect\ldquo ; being extremely feeble .
No trouble was experienced on this account in any of the diagrams so long as the permeability did not exceed about 100 .
Figs. 13 and 14 ( Plate 15 ) are a set relating to circular cylinders of permeability 2 and 100 respectively .
They illustrate clearly the well-known theoretical result that , on account of the shape of the cylinder , large changes of permeability produce only relatively slight changes in the netic induction through the cylinder .
In figs. 15-20 we have a set of rams which will be found interesting in connection with the important question of magnetic shielding\mdash ; a subject which has recently attracted a good deal of attention .
Fig. 15 is a diagram corresponding to the case of a hollow circular cylindric shield surrounding a solid cylinder , both cylinders having a permeability of 100 .
Fig. 16 relates to a hollow cylinder whose cross-section is bounded by two confocal ellipses .
As shown in the mathematical appendix , the field produced inside such a cylinder is uniform if the original impressed field is uniform .
Fig. 17 ( Plate 16 ) represents the effect produced by a double shield ; it is one of the earliest diagrams obtained by us , and the permeability is so extremely low that , even with the double shield , the field in the innermost space is comparable with the undisturbed field .
It would correspond to a -saturated double cylindric shield in a field of very great intensity .
Figs. 18 , 19 , and 20 have a more practical interest , the permeability being 100 in each case .
In fig. 18 we have a double shield , with a solid cylinder of iron placed concentrically in the innermost space ; the powerful shielding effect on the central cylinder is sufficiently evident .
Fig. 20 shows the effect produced by a triple conON LINES OF INDUCTION IN A FIELD .
.317 centric shield .
It is interesting to compare this latter stream-line diagram with the theoretical diagram of fig. 19 ; the agreement in the shape and general distribution of the lines between the two cases is very striking .
Fig. is a theoretical diagram for an elliptic cylinder of permeability 100 , and the ratio of whose axes is 2 : 1 , placed in a uniform field with the major axis of the elliptic section of the cylinder making an angle of with the impressed field .
Fig. 22 gives a stream-line diagram for a cylinder of same permeability , but the ratio of whose axes is 3 : 1 .
Fig. 23 is a similar diagram for a very thin elliptic plate , and fig. 24 relates to the hollow cylinder of fig. 16 , but here turned through an angle of It will be noticed that fig. 24 confirms the theoretical result that the field in the interior of a hollow elliptic cylinder bounded by two confocal surfaces is uniform if the impressed field be uniform .
We have hitherto dealt with cases which may be treated theoretically as well as experimentally .
But the number of such cases is very limited , and the vast majority of two-dimensional magnetic lems are beyond the powers of analysis .
It is in such cases that the stream-line method employed by us becomes a powerful weapon of.research .
Figs. 25-28 relate to cylinders of rectangular section , figs. 25 and 26 giving the field distributions for cylinders of square section placed with one of their diagonals at to the field and parallel to it respectively .
In fig. 25 , where the width of the cylinder remains constant along the direction of the impressed field , we notice that the lines inside the cylinder are concave outwards ; in fig. 26 , on the other hand , where the cylinder tapers rapidly as we proceed along the field , the curvature of the lines a convexity outwards .
This suggests that an intermediate form between the two might be found for which the lines exhibit neither convexity nor concavity , are straight ; and , a matter of fact , we know of one such intermediate form\mdash ; a circular cylinder .
Fig. 26 closely corresponds to the theoretical case which Dr. C. H. LEES has recently succeeded in working out analytically in connection with a problem in heat conduction .
Figs. 27 and 28 are intended to illustrate the effect of increasing the length of one of the sides of the rectangular section while keeping the other constant .
We know from theoretical considerations that this has the effect of reducing the de-magnetising factor , and thus increasing the flux through the cylinder .
This point is very clearly brought out by a comparison of figs. 25 , 27 , and 28 .
Figs. 29 and 30 ( Plate 19 ) show the magnetic fields corresponding to a cylinder of triangular section in two different positions .
Figs. 31 and 32\mdash ; a solid circular cylinder inside a hollow square one , and a solid square one inside a hollow circular one\mdash ; are interesting in connection with the problem of magnetic shielding .
Both these diagrams are slightly disfigured by air-bubbles .
These latter are extremely difficult to get rid of , once they are allowed to reach a of the field where the thickness of the liquid film varies .
* phil .
Mag February , 1900 , p. 226 .
Fig. 33 ( Plate 20 ) is a field diagram for a hollow square cylinder ; the shielding effect is seen to be very powerful .
Fig. 35 gives approximately the field distribution between the tapered pole-pieces of an electro-magnet ; and fig. 36 is intended to illustrate the pull of an electro-magnet on an armature .
The next few diagrams relate to cases of practical interest and importance , and serve to show how the stream-line method may be made to yield results of great interest to the electrical engineer .
Figs. 37 and 38 ( Plate 21 ) show an ordinary SIEMENS shuttle-wonnd armature\mdash ; such as is used in connection with telephone call apparatus\mdash ; in two positions .
The direction of the torque acting on the armature in its second position may be at once inferred from an inspection of the diagram .
It is also interesting to note the leakage lines outside the armature .
Figs. 34 and 39 relate to a toot , hed-core armature .
Fig. 34 shows the symmetrical field distribution obtained when the teeth are well under cover of the pole-pieces .
The induction in the air-gap consists of alternate maxima minima , and the lines are only slightly curved at the level of the teeth .
Below this level , they curve sharply to enter the flanks of the teeth .
The permeability of the core is assumed to be 100 .
It is evident that a diagram of this description furnishes a means ulculating the line-integral of intensity along a line drawn from the polar surface to the base of a tooth , provided the total flux per tooth and the permeability are known .
In fig. 39 we have the same armature , but now shown emerging from under the polar surface .
The fringe of the netic field is clearly exhibited .
The distribution of the lines in the core below the level of the bases of the teeth is not correct .
In an actual armature , the lines in this region , after passing into the core , would turn rply to the left , proceeding towards the neighbouring pole-piece .
In the streamline diagram , on account of the straight-line boundary on the left , the lines are forced to go on in a downward direction .
Yet it is curious to note the strong twist towards the left which the lines exhibit immediately after leaving the teeth , and which indicates the direction in which the flow would naturally continue if not subject to the artificial constraint just mentioned .
An objection which might be raised in connection with some of the diagrams is the fact that in those cases where the induction varies from point to point , the permeability in any actual magnetic substance is also variable .
In so far as this is the case , of course the stream-line diagrams do not afford a rigid solution of the problem .
But it must be remembered that there are many cases in which the magnetic intensity is either so weak or so strong that the permeability does not vary greatly within certain limits of the induction .
For all such cases , the solution obtained by the stream-line method is a very close approximation .
And even in cases where considerable variations of permeability occur , the stream-line method , at any rate , the first rough approximation to the solution of the problem , ON LINES OF INDUCTION IN A IslAGNETIC Brief History and Solution for Confocal Elliptic Shells .
( a ) The foundation of the modern theory 01 lnagnetic induction was laid by POISSON , between 1821 and 1838 .
He was the first to out in detail the solution for a solid or hollow sphere of paramagnetic material placed in a field of uniform intensity .
* Subsequently , he extended his investigations to the case of ellipsoid placed in a uniform field .
, at a later date , gave an approximate solution for a cylinder of finite length placed with its axis along a uniform field .
In 1848 , J. NEUMANN attacked the problem of an ellipsoid of revolution placed in any given field .
In 1854 , succeeded in solving the same problem for a circular cylinder of infinite length .
In 1881 .
, A. G. GREENHILL considered the case of a hollow ellipsoid .
S Lord KELVI was the first to publish , in 1872 diagrams of lines of induction for spheres of netic and diamagnetic material placed in a uniform field .
On account of their frequent reproduction ( they figure almost every text-book on the ject ) , these diagrams are now very well known .
In MAXWELL'S great treatise are to be found a number of diagrams .
These include the following cases : two circular cylinders rigidly magnetised transversely , and placed with their lnagnetic axes at right angles to each other ( vol. 2 , fig. 14 ) ; ( 2 ) a circular cylinder permanently magnetised in a transverse direction , and placed in a uniform field , so that the direction of magnetisation of the cylinder is coincident with that of the field ; ( 3 ) a cylinder of netic material in a uniform field ( vol. 2 , fig. 15 ) ; ( 4 ) a permanently magnetised cylinder in a uniform field whose direction is at right angles to the direction of magnetisation of the cylinder ( vol. 2 , fig. 16 ) ; ( 5 ) a field disturbed by a current in an infinitely long straight cylindrical conductor normal to the direction of the field ( vol. 2 , fig. 17 ) .
Tn 1882 , published a long paper dealing with the induced magnetisatiou of an infinitely long hollow circular cylinder .
He obtains a solution by assuming the magnetic potential V to be of the form , where is the distance of the point considered from the axis of the cylinder , and the angle between , and the direction of the impressed field ; the constants A and assuming different values for the regions within the cylinder , in its substance , and outside the cylinder respectively .
The function V is an integl.al of the equation , and it is the real part of the function Az of the complex ' vol. 2 , p. 59 .
, p. 66 .
roll , ' Bd .
37 ( 1848 ) .
S de Physiqne ' ( 1881 ) .
'Reprint of apers on Electrostatics Magnetism , ' pp. 493-495 .
' Wicn .
Ber , Part 2 , p. 613 .
variable .
The real factor of the imaginary part of this function is , and this gives the system of lines of induction .
STEFAN further deals with the case of a hollow cylinder along whose axis are placed two wires ( infinitely close together ) , conveying currents in opposite directions .
In 1894 , considered in detail the question of magnetic shielding by shclls , with special reference to the problem of maximum shielding for a given total weight of material .
In the same year PERRY contributed a brief paper to the Physical Society on the magnetic shielding of external space by a hollow iron cylinder enclosing two parallel conductors conveying equal currents in opposite directions in a diametrical plane of the cylinder , each conductor being at the same distance the axis .
In 1897 , fresh interest was given to the problem of magnetic shielding by the discussion which followed the reading of Mr. MORDEY'S paper on " " Dynamos before the Institution of Electrical Engineers .
Towards the close of that 'year , BOISS colnmenced the publication of an elaborate series of articles on magnetic shielding , in which he briefly reviewed the history of the subject , and gave diagrams of magnetic fields the case of hollow circular cylinders of varying thickness in a iform field Almost simultaneously with Du BOIS , published , in January 1898 , paper on the tic field due to a current in a wire placed parallel to the axis of a cylinder of iron .
This paper is accompanied by some extremely interesting diagrams of magnetic fields .
The latest contribution to this subject is a paper by A. P. WILLS , ' ' On the Magnetic Shielding Effect of Trilamellar Spherical and Cylindrical Shells which may be considered as au extension of DU BOIS ' inyestigations to triple shields .
( b ) The problem of the induced magnetisation due to a uniform impressed field in infinite ders of elliptic section or infinite cylindrical shells bounded by confocal elliptic surfaces , may be dealt with by the following method .
The general problem of magnetic induction in space of two dimensions may be regarded as consisting in the determination of a continuous potential function V , i.e. , a function satisfying LAPLACE'S equation for two-dimensional space , which fulfils the condition of continuity of normal induction across any surface of sepal.ation between two media , viz. , 'Phil .
, vol. 37 , p. 95 .
'Proc .
Phys. Soc vol. 13 , p. 227 .
' vol. 26 , p. 564 .
S The Electrician , ' vol. 40 .
Ibid. , vol. 40 , p. 613 .
Ibid. , vol. 40 , p. 453 .
'Physical Review ' [ 9 ] , pp. 193-213 , October , 1899 ON LINES OF ) IN ) .
stand pern ities drawn point the ce into the In vitl ) elliptic ) , it ) ando nse of rectangular ( ] recourse 10 ci ] functions .
take ( 1 ) where is that ) family , hile where is such the ) with ( If no we then ?
constant is equation certain bola , both curves being .
It is obvious that ( which ) vary fi ) ) , to , completely uniquely determine ) ) of point in ) ellipse ( 1 ) .
We use ordinateH .
In order to al.rive at the induced potent inction , ) the following considerations .
) shown ellipsoid placed in a uniform field ) made .
Then we pass ) the ) two-dimensional one of the cylindel see ) is uniforn ] .
) ) ) by imagining ) ) cylinders of the imaginary hnnedensity of sign , )riginally irlcident , ) relatively to other ongh i small distance Ss in direction ) , such that Ss is equal to the actual intensity of of the .
I [ stand potential at due to of the solid mattel , the bined potential of ) S But VOL. CXCV .
\mdash ; A. 2 PROFESSOR -SHAW AND MR ALFRED HAY the intensity-component in the direction of Ss due to one of the cylinders .
the ) finding the netic potential function is reduced to that of the intensity due to a solid cylindel of attracting or repelling matter .
The comp ents of intensity due.to latter cylinder be shown to be We itre led tu ) tentatively the somewhat ) general forms A sin , and 1 ) constants , as the typical magnetic potential functions for gnetised elliptic and confocal elliptic cylindric shells .
In , above ) essions , will all be different for the space included two co1lfocal bounding and for the external space ; to infinity .
And for the space the innel.most bounding , which includes the foci .
assume the ) bility to be stant , then need only consider the standard etisation the two axes of the ellipse ( 1 ) , as any interdirection be obtained by a proper superposition of the two standard We shall in the first place consider the case of solid cylinder whose bounding is ellipse of reference ( 1 ) , the impl.essed field being along the major axis of the ellipse .
Tf V and stand for essed potential , the lnduced potential inside , that outside the cylinder respectively , therl and we ssume 1 order that these satisfy LAPLACL'S equation\mdash ; may afford a solution ) the , they must satisfy the conditions of ( 1 ) continuity of potential itnd ( 2 ) continuity of normal induction\mdash ; two conditio1ls which enable us to determine the constants A and B. The hrst condition gives ( 2 ) .
ON INES OF INDUCTION IN A MAGNETIC FIELD .
32.3 The second gives , or where is the permeability , and ?
is the outward-drawn normal at any point of the ellipse ( 1 ) .
Now ' so that the last equation becomes Carrying out the differentiations , and then we get for the second condition . . . . .
( 3 ) .
Solving ( 2 ) and ( 3 ) for A and , we find .
H. We thus have Hx , cosb ab and the equipotential lines may at once be plotted from these equations .
equation to the lines of induction inside the cylinder is obyiously constant .
In order to find the equation to the external lines , we have to determine the function which is conjugate to .
Since the function which is con.jugate to is , and that conjugate to is we have ab for the equation to the lines of the external field , being a constant which varies from one line to another .
If we next assume that the direction of the impressed field is along the minor axis of the ellipse , we similarly finci ) the equation to the lines of the external field being as before a parameter which varies from one line to another .
The treatment given above for the case of a solid elliptic cylinder may be extended -SHAIV ANI ) HAY to of confocal elliptic cylinders of different permeabilities .
For if ( taking direction of the major axis of the ellipse ) we assume that the induced ) otential V the space is of the form the potential in the substance of any one of the hollow confocal is finally the induced in external space is , , then ( the tions V. . . . . .
V , . . .
satisfy two-dimensional the function V , ishes at infinite distance .
If , therefore , determine various constants , . . . . . .
, so as to satisfy ( 1 ) nnit y continuity of normal induction , the only possible of the problem will ) obtained .
Let there hollow ders .
The number of constants in the assumed ) essions f the otential will in that case be .
There are , and at such surface the two conditions of continuity of continuity of normal induction lllust be fulfilled , thus giving two for each ) onndary .
There will therefore be equations , these will completely determine the values of the constants .
By way of further illustration : shall work out the case of a hollow elliptic cylinder of laced i uniform field .
Let the internal bounding surface be the ellipse , the external bounding surface the ellipse or , say , where ?
the impressed potential be nding ) to a direction of the field along the major axis of the cylinder .
We .
the induced magnetic potential functions:\mdash ; ON LINFS OF INDUCTION IN A IstAGNETIC When ?
, , and when ; this gives . . . . .
( 4 ) and sin . . . . .
( 5 ) .
Again , when , we must have and when These two conditions give . . . . .
( 6 ) , ( 7 ) olving equations1 ( 4 ) , ( 5 ) , ( 6 ) , and ( 7 ) for , and , we ab ab rosh , These expressions may be reduced to a somewhat simpler form by putting , so that stand for the semi-axes of the external elliptic bounding surface .
Remembering that , we get -SHAW AND MR. HAY to the lines of induction For the central space inside the shell , constant .
For the space occupied by the substance of the shell , and for the external \mdash ; in each case a variable parameter .
If we suppose that the impressed field is along the minor axis of the ellipse , so th then , and V. , we find , proceeding as ) efore Or , in terms of where ) The equations to the lines of induction ON LINES OF INDUCTION IN A NETIC F ,327 constant , in the innermost space , , in the substance of the shell , and , in external space .
In working out numerical examples with the view of plotting curves , it is convenient to use a table of hyperbolic sines cosines , the one compiled by .
H. BLAKESLEY and published by the Physical Society of london .
) values of and having been found , it is ( easy step to ) to rectangular -ordinates , more convenient for the method given above , as ah.eady mentioned , be extended to of confocal elliptic cylindric shells .
In , however , when numerical available , it is best to substitute these in the instead of trying to obtain a solution in general terms .
The case of concentric circular shells may be regarded as a limiting case of elliptic shells , two axes of the ellipse equal .
If in ssioJlH obtained for the constants , \amp ; c. , in the case of a hollow elliptic shell regard constant and l ) as variable , and then proceed to the limit , we find that the values so obtained in agreement with those deduced by BOIS for circular cylindric shell in his rticles on " " Magnetic Shielding .
, But although an interesting verification of the method for a special case is thus obtained , it is much simpler , when dealing with circular cylindric shells , to follow the method developed by STEFAN in the paper already referred to .
'The Electrician , ' vol. 40 .
'Wien .
Akad .
Sitz .
Ber vol. 85 , Section 2 , p. 613 .
Theoretical Diagram for Infinite Elliptic Cylinder placed in uniform field .
Ratio of axes 3 : 1 .
Permeability Infinite Elliptic Cylinder in uniforn ] fieldRatio of axes 3 : 1 .
Permeability Stream-line Diagram to theoretical Diagram of Fig. 9 .
Infinite EUiptic Cylinder in for ) Ratio of axes 3 : 1 .
) ] Infinite Circular Cylinder in uniform field .
Permeability Solid Circular Cylinder shielded by hollow one .
Permeability Infinite Circular Cylinder in uniform field .
Permeability Infinite Hollow Elliptic Cylinder in uniforn ] field .
Permeability Double Cylindric Shield of very low permeability .
Triple Circular Cylindric Shield .
Theoretical Diagram .
Permeability Doul ) Cylindric Shield enclosing Solid Cylinder .
Permeability Triple Circular Cylindric ield .
Stream-line Diagram .
Pcrmeability haw Hay .
iptic Cylinder .
Ratio oi axes 2 : 1 , major axis inclined at to field .
Theoretical Diagram .
Permeability Flat Elliptic Plate , inclined to field .
Permeability .
Trans. , A. , Vol. , Plate 17 .
Elliptic Cylinder .
Ratio of axes 3 : 1 , major axis inclined at to field .
Stream-line Diagram .
Permeability Hollow Elliptic Cylinder .
Iajor axis inclined to field .
Permeability mite Cylinder of Square Section .
Permeability Infinite Cylinder of Rectangular Section .
Permeability Infinite Cylinder of luare Section .
Permeability Infinite Cylinder of ltecanguIar Section .
Permeability Infinite Cylinder of Triangular Section .
Square Cylinder enclosing Circular Solid one .
Permeability : ; 0 Infinite Cylinder of Triangular Section .
Permeability Hollow Circular Cylinder enclosing Solid Square one .
Hollow Square Shield .
Permeability Field between Tapered Pole-pieces 01 Electromagnet .
Induction in Air-gap and Teeth of Toothedcore.lrmature .
Permeability :36 lttraction of Armature by Pole-piece .
iemens Shuttle-wound Armature in two positions .
Fringe of Magnetic ield near edge of Pole-piece in Dynamo with Toothed-core Armature .

