]\gt ; On Standard of Mutual Inductance .
By ALBERT CAMPBELL , B.A. ( Communicated by Dr. R. T. Glazebrook , F.R.S. Received June 5 , \mdash ; Read June 27 , 1907 .
) ( From the National Physical Laboratory .
) CONTEN TS .
PAGE S1 .
Introductory 428 S 2 .
Practical Conditions in the Problem 428 S 3 .
Variation of with Dimensions and Positions of Coils 429 S4 .
Practical Case of Two Helixes and Circle 432 S 5 .
Actual Construction of Standard 435 S1 .
Introductory.\mdash ; In many electrical measurements , such as those of capacity and , as well as in the netic testing of iron , an accurately known standard of mutual inductance is of great .
It is sometimes convenient to derive such a standard from the standard unib of resistance , and this may be done in several ways , for example , by the wellknown method of the ballistic galvanometer ; or by Carey Foster 's method the mutual inductions may be tested against a condenser whose capacity has been found in terms of resistance and frequency by Maxwell 's commutator method ; or it may be obtained directly in similar terms by the help of an unknown inductance by the Hughes-Rayleigh method .
In the National Physical Laboratory I have used both of these latter methods ( with the help of a vibration galvanometer ) to obtain a working standard of mutual inductance .
* But this procedure is somewhat seeing that the unit of resistance has been itself commonly determined by the aid of mutual inductances calculated from the dimension of the coils or other conductors used ; thus for the highest accuracy it is desirable to revert to a standard whose value can be determined solely from the geometrical dimensions .
Accordingly , some eighteen months , I took in hand the investigation of a suitable design for such a standard , and I proceed describe the result at which I arrived .
S2 .
Practical Conditions.\mdash ; The most important conditions governing the design were that the standard must be\mdash ; ( a ) accurately calculable from the geometrical dimensions ; ( b ) as permanent as possible ; and 'Phys .
Soc May , 1907 .
On of ( c ) of a value sufficiently large to give high sensitivity when used in comparison methods , such as those mentioned above .
In addition to these , it was desirable to keep the resistances of the parts as low as possible and to avoid ( as far as possible ) the occurrence of eddy currents , and also of capacity effects between the and secondary oircuits .
As is so often the case in designs , a compromise had to be effected between the various conditions , and so it was decided to make the value approximately henry .
In order to carry out some of the above conditions , it is clearly desirable that the distance from the primary circuit to the secondary should , for all points , be relatively as great as possible .
For convenience , let us call the circuit with the smaller number of turns the primary , the secondary having turns , and let be the mutual inductance in millihenries .
Since , for a given geometrical disposition of the circuits , is proportional to , a little consideration showed that for millihenries it old be desirable to make of order of 100,000 .
Now it seems to be generally recognised that , for a coil whose dimensions have to be accurately measured , the satisfactory construction is of bare wire wound in an accurate screw-cut on a marble cylinder .
If the above conditions be kept in mind , it is out of the question , with , to attempt to construct both the primary and secondar ) circuits of single-layer coils .
As will be shown later , the solution of the problem consists in making the primary an accurately measured single-layer coil or coils , while the is a many-layered coil , so designed its dimensions position do equire to very accurately .
The possibility of such a and the method of carrying it out were found by an examination of the manner in which varies with the dimensions and positions of the primary and secondary coils .
I proceed to give some of the more interesting results of examination .
S3 .
Variation of Dimensions and Positions of the Coils.\mdash ; First , let us consider the simple case of two circular co-axial coils with ible section , assuming for convenience of calculation Let their radii be A and , and let be the distance between their Let , and Then* ( in henlies ) , ( 1 ) .
where and are complete elliptic integrals to modulus * Maxwell 's ( Elect. and Mag vol. 2 , S 701 .
430 Mr. A. Campbell .
[ June 9- , Since and , it can be shown that .
( 2 ) From equation ( 1 ) , mainly by the help of Lord Rayleigh 's table , in some cases by 's formula and tables , were calculated sets of values of for a fixed value of ( viz. , 10 cm and values of and From these calculations the families of curves shown in figs. 1 and 2 were drawn .
Each curve in fig. 1 corresponds to a constant value of , and in FIO .
1 .
*Ibid .
, 2nd edition .
TOky5 S\amp ; ubar ; g , ' Bntsu .
Kiji.gaiyo , ' vo12 , No. 17 .
1907 .
] On Standard of Mutual Inducta to a constant value of .
It will be noticed that in is always Illaximunl and only when , while in fig. 2 the nlaximum value occurs for a different value of not for each .
Thus when only two coils are used , should either be zero or relatively .
When is zero , is zero only when or the coils coincide ; when is relatively large , the whole construction has to be very bulky in order to obtain a enough M. the case of two coils is not sufficient for our purpose .
If , however , the primary consists of two equal coils arranged with the secondary between them as in fig. 3 , all three coaxial , is a maximmn minimum for axial displacements when If , then , for any desired value of we choose from the proper curve in the value of which gives a maximum , the mutual inductance thus obtained varies only very slightly for small variations of VOL. LXXIX.\mdash ; A. 2 Mr. A. Campbell .
[ June 5 , or small axial displacements of the centre coil ; in fact , we have placed the secondary coil in such a position that all round its mean circumference the due to tlu primary coils is zero .
FIG. 3 .
If be chosen of such a value that the corresponding curve in fig. 2 is reasonably flat at its highest point , then the mutual inductance per turn will be practically constant over the whole section of a secondary coil whose axial and radial depths are both small ; and the secondary may consist of a many-layered coil whose dimensions and position need not be known with any high accuracy .
By doing this we throw all the burden of accuracy on the two primary coils which we have assumed to be mere circles .
Clearly these must be replaced by single-layer coils of accurately known dimensiOns and relative position , and so we must extend the above investiga- tion to the more complicated case where the ) rimary consists of two coaxial helixes , of equal and finite length , with the secondary coil midway between them as in fig. 4 .
Here at the points and ( and all round the mean circumference of the field due to the two primaries should be zero , the component lines of force due to the upper and lower helixes bein , tangential to one another , and in opposite directions as shown .
S4 .
Helix Cirde.\mdash ; Several cases of the system shown in fig. 4 were investigated by means of the curves of fig. 2 .
The mutual inductance between the helix LN and the circle was taken as approximately* , where , and refer to circles at the ends and middle of the helix ; a similar approximation to was also made and a series of curves were * Merrifield 's 'B.A .
Report , ' 1880 .
907 .
] On Standard of Inductance .
lrawn .
The and the curves for the particular dimensions which roved satisfactory are shown in fig. 5 .
The relative dimensions ( fig. 4 ) Q. Mr. A. Campbell .
[ June 5 , , OH , OK , and OP , this last being the value of for which It will be noticed that , near the value shows very little variation with .
( I should mention that the values for shown in fig. 5 for a single helix and a circle are the same for the two helixes and circle , provided the total number of primary turns be kept constant .
) After the proper relative dimensions had been found by the above method , it was thought desirable to check the results by means of the much more laborious accurate formula of Viriamu Jones .
* The formula is applied to the helix LN in fig. 4 by calculating for a helix of height OK , and then for one of height OH , and taking the difference .
If , ' and then the Viriamu Jones formula may be reduced to , ( 3 ) where , also .
( 4 ) From ( 3 ) were calculated the values of for , helix to 10 , with , and cm .
respectively .
The results which are given in the following table entirely corroborate those obtained by the less exact method .
Similarly , formula ( 4 ) gave for Table .
ries Millihenries Finally the variation in dne to a small axial displacement of the secondary c , oil ( from the mid position ) was estimated .
It was found that a displacement of cm .
reduced by less than 1 in 10,000 .
'Roy .
Soc. Proc p. 192 , December 9 , 1897 .
1907 .
] On Standard of Mutual It will be seen that , with the above proportions , if the radius of the secondary coil is cm .
, we may make it a coil of many layers and of appreciable cross-section .
If , for example , the cross-section be cm .
cm .
, then the nlaximum variation from the mean value , over the whole section , of the inductance per turn will be the same within a few parts in 1000 , and thus we can with perfect safety obtain an accurate result by a method of averaging , such as the Purkiss formula used by Lord Rayleigh .
* S Actual of standard of the described above is at present being constructed at the National Laboratory .
In this the two primary helixes are of bare wire ( turns each ) wound on one nlarble cylinder of 30 cm .
diameter , while secondary coil consists of 488 turns in a channel of 1 sq .
cm .
section with a mean diameter of cm .
, the nominal value , therefore , close to 10 millihenries .
I may remark that the principle here employed will , without doubt , be of value in ) problems where accurately known mutual inductances are required .
In conclusion , I would express thanks to Dr. R. T. Glazebrook for valued and helpful criticism of the design , and to Mr. F. E. Smith for kind advice with regard to the material construction of it .
Maxwell 's ' Elect. and Mag vol. 2 , p. 350 , 3rd edition .

