Consider the transformer of Fig. 5(a).

There are two to use junction analysis it is necessary to replace it by

M
its equivalent T (Fig. (5b)) or wr (Fig. (Sc)) network. The
T equivalent is simply computed but possesses three
(a) L nodes, while the wr equivalent has only two nodes but
more complex constants. The choice of 7r or T would
obviously be a matter of the circuit and the constants
of the transformer. Generally, however, circuits con-
taining transformers are better handled with mesh
(b) currents.
V. CONCLUSION
- jO (L,-M) From the above illustrations it is seen that the repre-
sentation of a vacuum tube by a current generator in
Z3 r JU(L2-M) parallel with the plate impedances and the solution of
the equivalent circuit in terms of its junction voltages
generally leads to a simpler solution than the conven-
(c) ZA tional mesh-current approach. The only important cases
where this may not be true are in transformer-coupled
Fig. 5 A transformer and its T and 7r representations. amplifiers. A unique feature of the junction analysis
allows vacuum tubes to be represented as passive net-
meshes and the mutual inductance between the coils is works composed of a bilinear self impedance and a
treated as any other mutual impedance. If it is desired unilinear mutual impedance.

The Calculation of the Mutual Inductance of Circular

Filaments in Any Desired Positions *
FREDERICK W. GROVERt, SENIOR MEMBER, I.R.E.
Summary-The calculation of the mutual inductance of circular The case of coaxial circular filaments'-' has been quite
coils whose cross-sectional dimensions are small, compared with thoroughly treated by a number of authors since the
their distance apart, depends upon basic formulas for the calculation
of the mutual inductance of circular filaments. time of Maxwell and an accuracy exceeding anything
The case of coaxial circular filaments has been quite thoroughly required in practice is possible. Tables by Curtis and
treated and any desired accuracy is attainable. However, the avail- Sparks6 and by Grover7 make routine calculations quite
able formulas for more general cases such as circles with parallel or simple and rapid. Calculations for more general cases
inclined axes are slowly convergent for considerable ranges of the have, however, received relatively little attention and
parameters and for such cases admit of only rough accuracy.
The present paper is concerned with providing formulas and only in special cases are satisfactory formulas available.
tables capable of giving a moderate accuracy, with a moderate A formula for two circles whose axes intersect was
amount of labor, even in complicated cases. The tables apply to given by Maxwell.' Formulas for circles with parallel
calculations for equal circles with parallel axes, and to circles of axes have been given by Butterworth8'9 and Snow.10
nearly equal radii having inclined axes which intersect at the center Snow has also treated" the case of circles with inclined
of one of the circles. This latter case enters in the design of a variom-
eter composed of two coils of small cross sections; the former is axes with the center of one circle on the axis of the other.
applicable to the calculation of a variometer with coils having ec- Only in the case of Butterworth's formulas have any nu-
centric axes. merical data appeared. Unfortunately, existing formulas
A general formula is derived, which allows of a fairly simple are slowly convergent, are useful for rather limited
method of computation for any desired case, by the employment of
a chart of the flux distribution about a circular filament carrying
current. An accuracy of a few parts in a thousand is readily attain- ' Maxwell, "Electricity and Magnetism," vol. 2, section 701.
2 E. B. Rosa and F. W. Grover, Bull. Bur. Stand., vol. 8, pp.
able. 1-237; January, 1912. Scientific Paper 169.
Butterworth, Phil. Mag., vol. 31, p. 276; 1916.
Tp HE calculation of the mutual inductance of circu- F. W. Grover, Bur. of Stand. Scientific Paper 320, 1918.
lar coils whose cross-sectional dimensions are 'F. W. Grover, Bur. Stand. Jour. Res., vol. 1, pp. 487; 1928.
6 Curtis and Sparks, Bull. Bur. Stand., vol. 19. Scientific Paper
small, compared with their distance apart, de- 492; 1924.
7 F. W. Grover, Bull. Bur. Stand., vol. 20. Scientific Paper 498;
pends upon basic formulas for the calculation of the 1924.
mutual inductance of circular filaments. See section 697 of footnote reference 1.
Butterworth, Phil. Mag., vol. 31, p. 443; 1916.
* Decimal classification: R232. Original manuscript received by 9 Scientific Paper 320, Bur. of Stand. 1918.
the Institute, April 17, 1944. "Chester Snow, Bur. Stand. Jour. Res., vol. 3, p. 255; 1929.
t Electrical Engineering Department, Union College, Schenec- 11 Chester Snow, Bur. Stand. Jour. Res., vol. 1, p. 531; 1928.
tady, New York. Especially formula (23a).

620 Proceedings of the I.R.E. October. 1044

 Grover: Mutual Inductance of Circular Filaments 621

ranges of the parameters, and fail completely in other There results for the mutual inductance
cases. Furthermore, numerical computations are labori-
ous because of the nature of the calculations and the M = aAf dos,f
o o ~R
dO2.
number of terms which have to be calculated.
The present paper is concerned especially with the Performing the integration with respect to 0)2 there is
problem of providing methods and material suitable for found'2
routine calculations of an accuracy sufficient for practi- M = 47rVAa -rVI (1- p/a ol)N
dk, cos
(1)
cal purposes. Tables have been calculated for the more J r V3/2
important special cases and a general method is de- in which
N = (2/k - k)K - (2/k)E
veloped capable of giving results accurate to a few parts (2)
in one thousand. V = 1 - 2(p/a) cos 4)' + p2/a2
K and E are the complete elliptic integrals whose modu-
(CIRCLES WITH PARALLEL AXES lus k is given by the relation
Let the axes of the circles be a and A (Fig. 1) and the 1 - k-2k - (1 aV)2 + 62
-
(3)
distance between centers r. Butterworth's formula9"10 (1 + a1V)2 + 62
for distant circles r> (a +A) converges slowly unless r ca= a/A, 6=d/A.
is appreciably greater than 1.5(a+A). Buitterworth's The writer has not succeeded in performing the inte-
gration indicated in (1): attempts to develop the inte-
TP A
grand in series form are blocked by the fact that both
f--K.P
__ large and small values of k" occur in the range of
' * m ^ integration. However, it is entirely feasible to evaluate
M by mechanical integration. The quantity 47rwVAa N
is Maxwell's formula' for the mutual inductance of two
coaxial circles and tables7 give this in the form f-v'Aa
for values of f as a function of k". Thus equation (1)
may be written in the form
M = \Aa V(
-
d4) microhenries. (4)
The modulus k'2 is calculated by (3) for the selected
Fig. 1-Circular filaments with parallel axes. values of 4) and the corresponding values of f taken from
the tables7 of the Bureau of Standards Scientific Paper
formula for equal circles near together is useful only 498. In performing the mechanical quadrature, the fol-
when r is less than the radius. Snow'0 shows that Butter- lowing procedure has been found effective. To apply
worth's formula for unequal circles near together is in- Simpson's rule the interval of integration has to be
complete. He provides the correct formula for the divided into an even number of equal intervals. Weddle's
general case (A -a) <r < (A +a), but in such a form that rule assumes that the number of intervals is a multiple
considerable labor is necessary to put it into a form of six. If, therefore, the points are calculated for every
suitable for computation. All these formulas are series 15 degrees the integration may be performed by both
expressions, each term involving shape parameters and formulas. Simpson's rule ignores the fourth order in a
zonal harmonics of the angle X shown in Fig. 1; they are table of differences calculated from the ordinates, while
unsuited to routine numerical work. in Weddle's formula differences higher than the sixth
The general formula for this case may be obtained by are neglected. The closeness of agreement of the results
writing the Neumann integral ffcos edsjds2/R. Referring found by the two formulas is a measure of the impor-
to Fig. 1 and taking as origin the center of one circle the tance of higher-order differences. If now the interval of
co-ordinates of elements ds, and ds2 of the circles are the ordinates is halved and a calculation is made for
given by 24 points, then the difference between the two Simpson
x1 acos(p x2 =
p+A coss2 rule calculations, divided by 15, gives the correction to
y, a sin i A sin4)2y2 =
be applied to the 24-point value, provided differences
Z, ° Z2 d =
higher than fourth order are unimportant. Furthermore,
ds, ado i ds2 = Ad4)2. a close agreement between this corrected value and the
The angle e between them is given by value calculated by Weddle's rule from the 24 points
dxi dx2 dy, dy2 dz, dz2 confirms the accuracy of the result. The greatest diffi-
Cos e
= - --+ + = COS 02) culty lies in those cases where the curve of the ordinates
ds, ds2 ds1 ds2 ds, ds2
passes steeply through a peak value. Values obtained
and the radius vector R by from (4) using this procedure closely check the zonal
R = (p2+d2+ a2+A2) + 2p(A cos 2- acos )1) 12 Formulas (1) and (8), below were published by the author
- 2aAcos (c1 - 42). in Communications, October, 1938.
622 Proceedings of the I.R.E. October
TABLE I
VALUES OF F IN FORMULA (5) EQUAL CIRCLES
FOR WITH PARALLEL AXES

r/2a =0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 r/2a =1

1.0 1 1 1 1 1 1 1 1.0
0.9 1 1.0267 1.0330 1.0329 1 .0265 1.0146 0.9982 0.9790 0.9584 0.9376 0.9176 0.9
0.8 1 1.0552 1.0692 1 .0699 1 .0568 1.0313 0.9953 0.9527 0.9070 0.8613 0.8180 0.8
0.7 1 1.0857 1.1087 1.1112 1.0919 1.0509 0.9917 0.9200 0.8428 0.7665 0.6959 0.7
0.6 1 1. 1155 1.1517 1.1580 1.1328 1.0750 0.9876 0.8787 0.7619 0.6472 0.5441 0.6
0.5 1 1.1536 1.1997 1.2111 1.1812 1.1052 0.9842 0.8291 0.6585 0.4938 0.3515 0.5
0.4 1 1.1917 1.2524 1.2717 1.2390 1. 1440 0.9836 0.7668 0.5246 0.2914 0.1014 0.4
0.3 1 1.2330 1.3109 1.3411 1.3085 1.1952 0.9897 0.6964 0.3849 0.0137 -0.2378 0.3
0.2 1 1.2780 1.3760 1.4212 1.3929 1.2641 1.0102 0.5850 0.1178 -0.3874 -0. 7240 0.2
0.1 1 1.3274 1.4489 1.5139 1 .4959 1.3577 1.0557 0.5505 -0.1681 -1.0231 -1.5087 0.1
0 1 1. 3820 1.5311 1.6214 1.6220 1.4851 1.1450 0.5253 -0.4672 -
1.953 -4.053 0

A 2a/r =1 0.9 0.8 0.6 0.6 0.5 0.4 0.3 0.2 0.1 2a/e =0 A
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
0.9 0.9176 0.8968 0.8735 0.8481 0.8231 0.7946 0.7693 0.7471 0.7298 0.7188 0.7150 0.9
0.8 0.8180 0.7741 0.7266 0.6771 0.6292 0.5812 0.5398 0.5058 0.4806 0.4652 0.4600 0.8
0.7 0.6959 0.6267 0.5551 0.4848 0.4196 0.3626 0.3157 0.2797 0.2546 0.2399 0.2350 0.7
0.6 0.5441 0.4477 0.3543 0.2700 0.1992 0.1434 0.1019 0.0729 0.0540 0.0434 0.0400 0.6
0.5 0.3515 0.2275 0.1190 0.0336 -0.0287 -0.0697 -0.0956 -0.1109 -0.1195 -0.1237 -0.1250 0.5
0.4 0.1014 -0.0470 -0.1551 -0.2212 -0.2551 -0.2681 -0.2704 -0.2678 -0.2640 -0.2610 -0.2600 0.4
0.3 -0.2378 -0.3939 -0.4670 -0.4809 -0.4704 -0.4410 -0.4156 -0.3941 -0.3780 -0.3683 -0.3650 0.3
0.2 -0.7240 -0.8337 -0.7992 -0.7204 -0.6357 -0.5764 -0.5250 -0.4867 -0.4604 -0.4451 -0.4400 0.2
0.1 -1.5087 -1.3509 -1.0900 -0.8992 -0.7586 -0.6632 -0.5930 -0.5434 -0.5102 -0.4912 -0.4850 0.1
0- -4.053 -1.677 -1.2154 -0.9636 -0.8030 -0.6931 -0.6160 -0.5624 -0.5269 -0.5066 -0.5000 0

harmonic series formulas in cases where those are also Example 2:

applicable. For the case of two circles each of 48 inches diame-
Table I has been calculated to facilitate routine calcu- ter, distance between planes 15 inches, and distance
lations of the mutual inductance of circles with equal between centers r =50 inches, the parameters are
radii and axes parallel. The chosen parameters are 2a/r=48/50=0.96 and ,=15/50=0.3. Table I gives
r/2a = distance between centers/diameter (or 2a/r, F= -0.3103. From Table 6, Scientific Paper 498, with
whichever is less than unity), and ,u=d/r=cos X, (see diameter/distance = 0.96, there is found f= 0.0012982,
Fig. 1). Writing Mo for the mutual inductance of two so that
coaxial circles having a spacing d equal to the value of M = 24(2.54)(0.0012982)(- 0.3103)
r for the given circles, the required mutual inductance is
0.02456 microhenries
-

M = MoF (5) 24.56 abhenries.

-

the quantity F being obtained from the table for The negative sign signifies that the electromotive force
the given values of the parameters, and Mo from induced in one circle by a change of current in the other
Table 4 or 6 of Scientific Paper 498 with the argument is opposite in direction to the electromotive force re-
distance/diameter = r/2a (or diameter/distance = 2a/r). sulting from the same change of current with the circles
The values of F in Table I for r/2a=0.1 to 0.5, in- arranged in the coaxial position.
clusive, and for 2a/r = 0 to 0.5 inclusive, were calculated An inspection of Table I brings out some interesting
by Butterworth's9'l0 zonal harmonics formulas. The re- facts. With a given pair of equal circles, if one circle is
maining values had to be calculated by the quadrature moved, keeping the distance between centers constant
formula, a time-consuming piece of work. and the axes parallel, the mutual inductance varies
through a large range of values and in a manner depend-
Example 1: ing upon the parameter r/2a. For circles near together,
As an example, consider the case of two circles of r/2a small, the mutual inductance increases continuously
equal radii a= 15 centimeters with a distance between from the coaxial to the coplanar position, owing to the
centers r =20 centimeters and the distance between decrease in distance between the planes. For the case
their planes d = 16 centimeters so that A = cos X = 0.8. r/2a-0.6, the effect of decreasing distance between
The value of r/2a is 2/3. From Table 4 of Scientific planes is compensated by the opposite effect of increas-
Paper 498 there is found for equal circles for which ing distance between axes. For greater values of r/2a,
distance/diameter = 2/3, the value f = 0.0031239, so that the mutual inductance decreases continuously with in-
creasing X, that is, decreasing A, passes through zero and
mo = fa = 0.0031239(15) = 0.046858 microhenry.
becomes negative. Table I is useful in placing two circles
From Table 1, for r/2a=2/3 and M=0.8 the ratio or coils with parallel axes so as to have zero mutual
F=0.9928 is interpolated, so that inductance. For values of r/2a less than about 0.76, the
M = 0.9928(0.04686) = 0.04652microhenry
mutual inductance does not become zero for any value
= 46.52 abhenries.
of the angle X. Table II shows values of X for zero mutual
inductance for different spacings.
If each circle formed the center filament of a coil of 100 Table I suffices for all cases of equal circles with paral-
turns of very small cross section, the mutual inductance lel axes where values may be interpolated with the re-
of the coils would be 100 times 100 as great as this value quired accuracy. Otherwise, the solution may be made
or 465.2 microhenries. by direct calculation by (5) or by the general method to
1944 Grover: Mutual Inductance of Circular Filaments 623
TABLE II N do
ANGU l AR POSITION FOR ZERO ML TI AL INDUCTANCE, PARALLEL EQUAL CIRCLES M = \ Aa cos0jrT p3/(
2a/r u N\ 2a'r L N r/2a So
Degrees Degrees Degrees
in which again N is to be identified with the quantity
0
0.1
0.58
0.58
54.6
54.6
0.5
0.6
0.53
0.51
58.0
59.3
1.0
0.9
0.37
0.30
68.3
72.5
1000 f in Table 1 of Scientific Paper 498 for coaxial
0.2
0.3
0.575
0.565
54.9
55.6
0.7
0.8
0.485
0.455
61.0
62.9
0.8 0.16 80.8 circles, but the argument now is
0.4 0.55 56.6 0.9 0.42 65.2 1 + a2 + 62 + 2a5 cos 4 sin 0 - 2aP
0.5 0.53 58.0 1.0 0.37 68.3
k (9)
1 + al + 62 + 2a( cos ) sin 0 + 2acP
be described later. An exception is made of the case of with
equal coplanar circles. These will require special treat- P = -\/1 - cos2 4 sin2 0) (10)
ment, since interpolation for them in Table I is uncer- Formula (8) may be evaluated by mechanical quadra-
tain. ture as in the previous case. It gives results in agree-
For equal intersecting coplanar circles write
ment with Snow's formula in the region where the latter
M = Ca microhenries (6) converges rapidly.
and for distant coplanar circles
M = - 0.0O1(7r2/8)(2a/r)3D microhenries. (7) ,\ 5(-
Values of C and D are given in Table III. 5
TABLE III C.l
CONSTANIS IN FORMUt AS (6) AND (7). EQUAL COPLANAR CIRCLES

r /2a C Diff. logio C 2a/r L C D Diff.

0.1 0.029766 2.47372 1.0 -0.005749 4.6604
-9085
0.2 0.020681 2.31557 0.9 -0.001886 2.0969
-5608 -4487 A A

0.3 0.015073 2.17821 0.8 -0.001041 1.6482

-4233 -2350
0.4 0.010840 2.03505 0.7 -0.0006160 1.4132
-3506 -1463
0.5 0.007334 3.86537 0.6 -0.0003376 1.2669 Fig. 2-Circular filaments with inclined axes.
-3062 983
0.6 0.004272 3.63068 0.5 -0.0001802 1.1686
-2766 680
0.7 +0.001506 17776 0.4 1.1006
-2551
3.
468 Table IV has been computed to facilitate routine
0.8 -0.001045 n3.01915 0.3 1.0538
-2412 306 calculations. This covers the more difficult but impor-
0.9 -0.003457 .53869 0.2 1.0232
tant cases of nearly equal circles, values of the ratio of
113
-2292 175
-0.005749 n3 .75964 0.1 1.0057
1.0
57 the radii from 0.5 to 0.9 being included. The param-
a

1
eters used are u = cos 0 and d/A (or 1 / 6, whichever
0
3=

is less than unity). For a=0.8 and a=0.9 and the

Tables for- unequal circles with parallel axes would smaller values of the mechanical quadrature formula
have to be based on three parameters and would be (8) had to be employed. In the rest of the cases Snow's
voluminous. It seems better to calculate individual formula was employed, but even here the labor was
cases as they arise from (5) or, where possible, from an quite heavy, since terms of the order of those involving
appropriate series formula. For routine purposes where P15(1.) were included where necessary.
a moderate accuracy only is required the general method In Table IV, as in Table I, the calculation of the
described below is to be recommended. mutual inductance is based on the corresponding case of
coaxial circles. The quantity tabulated is the factor F1
CIRCLES WITH INCLINED AXES in the formula
M0F, cosM = (11)
The most important case is that in which the inter-
6.

section of the axes is at the center of one of the circles. This form of expression takes account of the fact that if
The nomenclature is shown in Fig. 2. For circles where the field were uniform, the mutual inductance would! be
the ratio of the radii a = a/A is nearly unity and the proportional to the cosine of the angle of inclination of
distance ratio 3=d/A is small, the series expression of the axes. Mo, as before, is the mutual inductance of
Snow,11 which involves zonal harnionics Pm(.u) and the coaxial circles having the same radii and the same
derivatives Pm' (3/V1t +62), converges slowly. Such cases centers as the given circles. F, is a factor which takes
may be treated by writing the Neumann integral for into account the nonuniformity of the field. For snmaller
inclined circles, following the method described for the values of not included in the table this factor may be
a

previous case. The resultant expression is calculated by the following formula derived from
Snow's1' expression
1 .-
(1/4)032P3'(Y)P3(g)/g + (1/8)(34)P5'(,y)P5(,1)/U (5/64)36P7'(Y)P7(,)/A + * *
-

1 - (1/4)32P3'(Y) + (1/8)fl4P5'(y) -
(5/64)36P7'(Y) + * (12)-
(
624 Proceedings of the I.R.E. October

TABLE IV
VALUES OF F, IN FORMULA (11) FOR CIRCLES WITH INCLINED AXES
RATIO OF RADII =0.5

0.2 0.3 0.4 0.6 0.7 0.9 1.0

-1--
pA 1/5 =0 0.1 0.5 0.8

0
I!I
iI 1.0092
-1- 1.0349
-1- 1.0719 I_I-
1.1129
-7, 1. 1496
-I
1.1965 1. 1988 1.1852 0

0.1 1.0091 1.0346 1.0712 1.1489 1.1944 1.1969 1.1837 0.1

0.2 I 1.0088 1.0335 1.0688 1.1439 1.1882 1.1938 1.1912 1.1790 0.2
0.3 I 1.0084 1.0317 1.0650 1.1357 1.1618 1.1778 1.1837 1.1816 1.1710 0.3
0.4 1.0077 1.0292 1.0598 1.1243 1.1634 1.1694 1.1681 1.1596 0.4
0.5 I 1.0069 1.0260 1.0530 1.1099 1. 1450 1.1507 1.1442 0.5
0.6 I 1.0059 1.0221 1.0449 1.0927 1.1227 1.1286 1.1292 1.1247 0.6
0.7 1.0047 1.0175 1.0355 1.0728 1.0968 1.1034 1.1007 0.7
0.8 1 .0033 1.0123 1 .0248 1.0506 1.0674 1.0714 1.0734 1.0719 0.8
0.9 1.0017 1.0065 1.0130 1.0262 1.0350 1.0390 1.0382 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1.0
p 51 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 50 A
0 1.1852 1.1660 1.1376 1.0993 1.0522 0.9966 0.9407 0.8835 0.8361 0.8045 0.7934 0
0.1 1.1837 1.1648 1.1368 1.0989 1.0523 0.9972 0.9414 0.8841 0.8373 0.8058 0.7947 0.1
0.2 1.1790 1.1612 1.1343 1.0979 1.0526 0.9989 0.9434 0.8873 0.8410 0.8097 0.7987 0.2
0.3 1.1710 1.1548 1.1300 1.0958 1.0529 1.0016 0.9468 0.8928 0.8472 0.8164 0.8054 0.3
0.4 1.1596 1.1455 1.1234 1.0925 1.0529 1.0050 0.9518 0.9007 0.8563 0.8262 0.8154 0.4
0.5 1.1442 1.1328 1.1143 1.0874 1.0523 1.0088 0.9582 0.9111 0.8686 0.8394 0.8289 0.5
0.6 1.1247 1.1161 1.1016 1.0797 1.0502 1.0125 0.9663 0.9242 0.8844 0.8568 0.8468 0.6
0.7 1.1007 1.0950 1.0847 1.0685 1.0457 1.0153 0.9757 0.9401 0.9046 0.8794 0.8701 0.7
0.8 1.0719 1.0688 1.0627 1.0524 1.0372 1.0159 0.9858 0.9587 0.9299 0.9087 0.9008 0.8
0.9 1.0382 1.0372 1.0346 1.0300 1.0227 1.0121 0.9944 0.9795 0.9614 0.9474 0.9420 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
RATIO OF RADII =0.6

JA 1/5 =0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1/5-1 pA
0 1 1.0133 1.0507 1.1052 1.1665 1.2234 1.2670 1.2925 1.2998 1.2942 1.2722 0
0.1 1 1.0131 1.0502 1.1040 1.1646 1.2208 1.2640 1.2894 1.2968 1.2913 1.2702 0.1
0.2 1 1.0127 1.0486 1.1006 1.1590 1.2131 1.2549 1.2800 1.2880 1.2826 1.2641 0.2
0.3 1 1.0121 1.0459 1.0949 1.1496 1.2005 1.2400 1.2643 1.2731 1.2690 1.2535 0.3
0.4 1 1.0111 1.0422 1.0870 1.1368 1.1830 1.2194 1.2425 1.2521 1.2505 1.2378 0.4
0.5 1 1.0100 1.0375 1.0770 1.1206 1.1651 1.1934 1.2147 1.2247 1.2253 1.2163 0.5
0.6 1 1.0085 1.0319 1.0650 1.1014 1.1351 1.1624 1.1810 1.1910 1.1933 1.1882 0.6
0.7 1 1.0067 1.0252 1.0510 1.0794 1.1055 1.1268 1.1420 1.1510 1.1537 1.1526 0.7
0.8 1 1.0047 1.0177 1.0356 1.0549 1.0727 1.0874 1.0983 1.1053 1.1087 1.1090 0.8
0.9 1 1.0025 1.0092 1.0185 1.0283 1.0373 1.0449 1.0506 1.0546 1.0570 1.0579 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
5A6= 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 =0 A
0 1.2722 1.2425 1.1996 1.1431 1.0754 0.9910 0.9166 0.8404 0.7746 0.7306 0.7149 0
0.1 1.2702 1.2411 1.1988 1.1430 1.0758 0.9979 0.9179 0.8418 0.7761 0.7321 0.7164 0.1
0.2 1.2641 1.2366 1.1963 1.1425 1.0771 1.0006 0.9216 0.8461 0.7806 0.7366 0.7209 0.2
0.3 1.2535 1.2288 1.1918 1.1413 1.0790 1.0051 0.9278 0.8533 0.7883 0.7444 0.7287 0.3
0.4 1.2378 1.2169 1.1844 1.1388 1.0810 1.0111 0.9366 0.8639 0.7996 0.7560 0.7403 0.4
0.5 1.2163 1.2000 1.1732 1.1340 1.0826 1.0183 0.9480 0.8779 0.8150 0.7718 0.7563 0.5
0.6 1.1882 1.1768 1.1566 1.1253 1.0823 1.0258 0.9619 0.8960 0.8354 0.7932 0.7779 0.6
0.7 1.1526 1.1460 1.1327 1.1106 1.0779 1.0323 0.9777 0.9187 0.8621 0.8218 0.8071 0.7
0.8 1.1090 1.1064 1.0995 1.0866 1.0660 1.0344 0.9936 0.9460 0.8972 0.8608 0.8472 0.8
0.9 1.0579 1.0575 1.0554 1.0505 1.0417 1.0260 1.0048 0.9764 0.9429 0.9160 0.9055 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0

RATIO oF RADII -0. 7

IA 1/5=0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1/6=1
0 1 1.0181 1.0696 1.1459 1.2331 1.3150 1.3777 1.4136 1.4221 1.4080 1.3788 0
0.1 1 1.0179 1.0689 1.1442 1.2303 1.3112 1.3733 1.4091 1.4182 1.4049 1.3765 0.1
0.2 1 1.0174 1.0667 1.1393 1.2220 1.2999 1.3601 1.3958 1.4062 1.3951 1.3693 0.2
0.3 1 1.0165 1.0630 1.1311 1.2084 1.2813 1.3383 1.3735 1.3858 1.3783 1.3565 0.3
0.4 1 1.0152 1.0578 1.1199 1.1899 1.2558 1.3083 1.3422 1.3566 1.3534 1.3370 0.4
0.5 1 1.0135 1.0513 1.1058 1.1666 1.2240 1.2704' 1.3021 1.3180 1.3193 1.3091 0.5
0.6 1 1.0115 1.0435 1.0889 1.1393 1.1866 1.2256 1.2536 1.2698 1.2749 1.2708 0.6
0.7 1 1.0092 1.0343 1.0697 1.1083 1.1446 1.1749 1.1976 1.2125 1.2197 1.2205 0.7
0.8 1 1.0064 1.0240 1.0482 1.0744 1.0988 1.1194 1.1354 1.1469 1.1540 1.1574 0.8
0.9 1 1.0034 1.0125 1.0249 1.0381 1.0502 1.0606 1.0689 1.0752 1.0795 1.0827 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
5=1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 5-=0 p
0 1.3788 1.3354 1.2745 1.1967 1.1055 1.0028 0.8979 0.7974 0.7104 0.6497 0.6276 0
0.1 1.3765 1.3339 1.2740 1.1970 1.1066 1.0043 0.8995 0.7990 0.7121 0.6513 0.6291 0.1
0.2 1.3693 1.3293 1.2721 1.1977 1.1097 1.0088 0.9044 0.8041 0.7172 0.6562 0.6337 0.2
0.3 1.3565 1.3210 1.2684 1.1985 1.1140 1.0159 0.9127 0.8129 0.7259 0.6645 0.6419 0.3
0.4 1.3370 1.3076 1.2618 1.1985 1.1197 1.0255 0.9247 0.8258 0.7387 0.6769 0.6542 0.4
0.5 1.3091 1.2872 1.2503 1.1961 1.1253 1.0375 0.9406 0.8434 0.7566 0.6943 0.6714 0.5
0.6 1.2708 1.2571 1.2308 1.1886 1.1293 1.0511 0.9607 0.8667 0.7807 0.7182 0.6950 0.6
0.7 1.2205 1.2144 1.1992 1.1712 1.1275 1.0641 0.9846 0.8971 0.8135 0.7512 0.7279 0.7
0.8 1.1574 1.1568 1.1512 1.1375 1.1125 1.0702 1.0097 0.9354 0.8585 0.7984 0.7753 0.8
0.9 1.0827 1.0842 1.0840 1.0808 1.0729 1.0557 1.0251 0.9790 0.9217 0.8709 0.8503 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
RATIO OF RADII =0.8

p 1/5=0 0.1 0.2 0.3 Oi0. 0.5 0.6 0.7 0.8 0.9 1/5 =1 p
0 1 1.0237 1.0919 1.1947 1.3139 1.4287 1.5156 1.5633 1.5718 1.5492 1.5063 0
0.1 1 1.0235 1.0909 1.1924 1.3100 1.4232 1.5093 1.5573 1.5668 1.5455 1.5039 0.1
0.1 1 1.0228 1.0879 1.1856 1.2983 1.4070 1.4908 1.5393 1.5517 1.5342 1.4964 0.2
0.3 1 1.0216 1.0830 1.1743 1.2791 1.3805 1.4600 1.5089 1.5254 1.5142 1.4827 0.3
0.4 1 1.0199 1.0761 1.1589 1.2531 1.3444 1.4177 1.4658 1.4868 1.4833 1.4608 0.4
0.5 1 1.0177 1.0674 1.1396 1.2209 1.2997 1.3645 1.4101 1.4346 1.4393 1.4272 0.5
0.6 1 1.0151 1.0569 1.1169 1.1833 1.2477 1.3018 1.3424 1.3682 1.3795 1.3777 0.6
0.7 1 1.0120 1.0449 1.0911 1.1414 1.1901 1.2317 1.2645 1.2881 1.3026 1.3085 0.7
0.8 1 1.0084 1.0313 1.0627 1.0961 1.1285' 1.1562 1.1791 1.1969 1.2101 1.2187 0.8
0.9 1 1.0044 1.0162 1.0322 1.0485 1.0646 1.0782 1.0897 1.0991 1.1067 1.1128 0.9
1.0 11111111111 1.0
 1944 Grover: Mutual Inductance of Circular Filaments 625
TABLE IV-Continued

a-= 1 0.9 0.8 0. 7 0.6 0.5 0.4 0.3 0.2 0.1 5I=0 A

0 1.5063 1.4454 1.3637 1.2627 1.1475 1.0189 0.8911 0.768 0.655 0.566 0.530 0
0.1 1.5039 1.4443 1.3640 1.2636 1.1492 1.0210 0.8928 0.767 0.654 0.568 0.533 0.1
0.2 1.4964 1.4408 1.3636 1.2664 1.1540 1.0271 0.8985 0.773 0.658 0.573 0.536 0.2
0.3 1.4827 1.4337 1.3627 1.2706 1.1617 1.0368 0.9090 0.782 0.667 0.578 0.543 0.3
0.4 1.4608 1.4209 1.3593 1.2752 1.1721 1.0508 0.9245 0.796 0.680 0.592 0.556 0.4
0.5 1.4272 1.3989 1.3503 1.2782 1.1845 1.0693 0.9453 0.819 0.700 0.611 0.573 0.5
0.6 1.3777 1.3627 1.3304 1.2754 1. 1963 1.0916 0.9729 0.847 0.729 0.633 0.597 0.6
0.7 1.3085 1.3059 1.2913 1.2583 1.2015 1. 1153 1.0077 0.886 0.765 0.668 0.627 0.7
0.8 1.2187 1.2235 1.2231 1.2126 1.1855 1. 1309 1.0475 0.928 0.818 0.719 0.679 0.8
0.9 1.1128 1.1178 1.1222 1.1242 1.1215 1.1069 1.07 19 1.0075 0.905 0.807 0.764 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
RATIO OF RADII =0.9

1/3=0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1/5=1 #

0 1 1.0305 1.1178 1.2526 1.4133 1.5681 1.6850 1.7454 1.7477 1.7158 1.6564 0
0.1 1 1.0301 1.1165 1.2495 1.4079 1.5605 1.6765 1.7379 1.7420 1.7120 1.6541 0.1
0.1 1 1.0292 1.1126 1.2403 1.3916 1.5379 1.6514 1.7148 1.7242 1.7003 1.6470 0.2
0.3 1 1.0276 1.1061 1.2251 1.3651 1.5010 1.6091 1.6751 1.6926 1.6787 1.6343 0.3
0.4 1 1.0255 1.0971 1.2045 1.3294 1.4510 1.5508 1.6178 1.6443 1.6437 1.6131 0.4
0.5 1 1.0227 1.0858 1.1789 1.2857 1.3897 1.4777 1.5425 1.5765 1.5899 1.5765 0.5
0.6 1 1.0193 1.0723 1.1490 1.2353 1.3192 1.3922 1.4501 1.4875 1.5118 1.5155 0.6
0.7 1 1.0153 1.0568 1.1154 1.1800 1.2423 1.2975 1.3438 1.3783 1.4066 1.4221 0.7
0.8 1 1.0107 1.0394 1.0789 1.1213 1.1618 1.1977 1.2291 1.2546 1.2781 1.2955 0.8
0.9 1 1.0056 1.0204 1.0402 1.0609 1.0802 100972 1.1125 1.1248 1.1374 1.1478 0.9
1.0 1 1 1 1 1 1 1 1 1 1 1 1.0
5 =1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 5=0 As
.I I. 0.9025 0.7511 0.6078
I-
0 1 .6564 1.5737 1.4700 1.3451 1.2034 1.0588 0.4741 0.4099 0
0.1 1. 6541 1.5734 1.4705 1.3466 1.2057 1.0580 0.9050 0.7520 0.6095 0.4763 0.4114 0.1
0.2 1.6470 1.5722 1.4725 1.3517 1.2130 1.0644 0.9109 0.7596 0.6150 0.4807 0.4161 0.2
0.3 1.6343 1.5686 1.4762 1.3606 1.2260 1.0808 0.9273 0. 7725 0.6246 0.4890 0.4229 0.3
0.4 1.f6131 1.5598 1.4793 1.3723 1.2427 1.0973 0.9432 0.7887 0.6389 0.5012 0.4311 0.4
0.5 1.5765 1.5403 1.4771 1.3845 1.2665 1.1207 0.9693 0.8121 0.6593 0.5185 0.4472 0. 5
0.6 1. 5 155 1.5008 1.4620 1.3921 1.2887 1.1564 1.0067 0.8473 0.6886 0.5431 0.4673 0.6
0.7 1.4221 1.4275 1.4178 1.3820 1.3091 1.2007 1.0567 0.8960 0.7308 0.5794 0.4969 0.7
0.8 1.2955 1.3100 1.3214 1.3225 1.2993 1.2331 1.1190 0.9648 0.7950 0.6383 0.5433 0.8
0.9 1.1478 1.1579 1.1698 1.1829 1.1971 1.1918 1.1694 1.0658 0.9064 0.7313 0.6278 0.9
1.0 1 1 1 1 1 1 1 1 1 I 1 1.0

in which ductance of a variometer consisting of two circular coils

a2 (1/6)
a of small cross section, with the movable coil centered on
32 = _ =
1 + 62 1 + (1/6)2 the axis of the larger, would have a mutual inductance
32 1 closely proportional to the cosine of the angular in-
2
7=
1 +32 1 + (1/6)2
clination, even with quite a range of values of the ratio
a of the radii, provided the space ratio parameter 6 is of
The general term of the denominator is the order of 0.5 or a little smaller. To obtain a more
1 . 3 .5 * * (2n - 1) 'nI+( Y) (14) nearly linear relation, the coils should be placed still
2 46 ..*. 2n(n + 1) 2 more nearly concentric. (It is, of course, impossible to
obtain a linear range over the whole 90 degrees of rota-
and P2n+1'(,y) denotes the differential coefficient of the tion.) To obtain larger values of mutual inductance, the
corresponding zonal harmonic with respect to its argu- radii of the coils should approach equality. However, in
ment. The zonal harmonic factors may be taken from a
such cases the effect of the cross sectional area would
table. The following series, however, may be useful
become more important.
P3'(Y) = 3/2(5y2 - 1),
P51(,y) = 15/8(2lJy4 - 14y2 + 1) GENERAL METHOD OF CALCULATION
P7/(Y) = 7/16(429y6 - 495,y4 + 135y2 5) Tables to cover all possible cases of inclined circles
P3(A)IA = 1/2(5 - 3), (15) and unequal circles with parallel axes and with intervals
P5(A)/, = 1/8(63A4 - 70gU2 + 15) of the argument small enough to allow of accurate inter-
NA) /11 1/16(429M6 - 693 +
=
A

315A2 35). polation, would be very unwieldy and would enta;I a

In (11), F1 is quite closely equal to unity for very vast amount of calculation for their preparation. Fur-
distant circles, where the magnetic field is approxi- thermore, other more general cases would remain un-
mately uniform over the area of the circle. For nearer provided for. The following general method for making
circles may
it be either greater or less than unity. An routine calculations in any desired case is offered as a
examination of Fig. 2 shows that, for the region P of the practical solution of the problem. It leads to an accuracy
smaller circle, the magnetic flux density greater than
is sufficient for most practical work and with no excessive
at the center R, and at Q the flux density is less than at amount of labor.
R. These differences nearly compensate for values of 6 A study of (4) and (8) shows that in each the mutual
of the order of 0.5. For the concentric case 6 =0 both inductance is given as a line integral, each element of
regions P and Q are in much weaker portions of the field which may be evaluated in terms of the mutual induct.
than that at the center and F1 is considerably smaller ance of a pair of coaxial circles. Consider, for greater
than unity. definiteness, the case of the inclined circles in Fig. 2.
A study of Table IV shows that the mutual in- Taking the origin at the center of the larger circle,
 _~~~~~~~~~~~~~~~~~~~~~~~~~)f to _z _if (O 44 f L.+

A I
1X 11IIISIII.IIIIIIIIMI_1151iI /I I IA I I /I I VI LI I YI I I I 'i I I LI I

I3S1 I0- ISW

-
4;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ W4 AII
1-1
ri

VIF
I
Z/_1-t1 1/I-t/L1Il1LI I L 1 1 II111111I
C -111111I111111IT1111111IIII
IS G tiGVItI1 -FlW 0100 1 4iW W1 11 +It _A
aL~~1-11 IA1III 11II11 1 __1IIIII IIIIIIIIIL1 1111111111111111111111111111111111111111111 111111111111

2X1$(; lk1
tXg||t4|T|Y-|2g4t|ttlttit1-11 bt
W A / ID 11II W VI i14 _l'III HI II II tnOFIt
2#t2S1llu
gt1X111~~~~~~f1111;IlnIml4tlLlMlil#lLSiiiiiiIIIIIMIrlII

+ IA1
e1/14~~~~~~~~~~ I
1'I1 1111
Af
I/-r!v II
rl'1 1 _
,4RT IIIIIIIIII||tIIII I IILIIIIIIIIIIgr l1111111111O ,

A 4
1114AX 2 + 1 = v1
XI H IIIIIIIIIIILl 00S L U IIIII > 111111111
a~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
[email protected] - S-1 0tS1N't [lUlMtlQll I t0INII3l1a,
1, l t l -A I-
I1/RIVIt0 LA' It k Ek 1('I111 411 S
IE

]/ VI II 00 J

zig.~~~~~~~~~O
O-uvsso
'A1
-igterto1/ oftemtainutneo oiaciceofterduoftelrecic.Abissae
ratio of th radi othci rlsa/;odntsaertooftedsacd e-enpaeadthlrgraiuA.
 - 0I 03 _aA 4 0
° O.( 0 0.' - 0.9 1.0 [2. f.3
_.
- I - -I I r-Iw'L 4-
-Ko
* i I t

_o
t t I

wi,y

4,
4s

06'6 1 c43]

07 ,0 °-7 C'
,_

5. t o.,

SVA ;t

;t
as5
m

0.4. 't

.3 p
o3
;t
C-1.

o0.2
t

0.I 0.1

0., 0.t .3 (Xjt 0~ 0O47 Al O.l 0.9 I 1.A §J! 13 )

Fig. 4-Curves of l/A of coaxial circles. Detail of part of the region covered in Fig. 3 to an enlarged scale.
628 Proceedings of the I.R.E. October

and the axis of Z along its axis, pass through any The calculation of the ordinates used in the mechani-
element ad) of the smaller circle a circle C coaxial cal quadrature may then be arranged as shown in Table
with the larger circle A. The radius of this circle is V. For each point (a', 3') the value of m/A is obtained
a,=a-/I-COS 24) sin2 O=aP, and its plane is at a dis- TABLE V
tance zi=d +a cos sin from that of the circle A. The
mutual inductance m of the pair of circles A C may be De- (1/4) p2 (1/4) cosI '1+(1 /4) m Ordinate
_1 /P2(m /A) =yn
cos' a' =aP
calculated by the formulas and tables for coaxial circles. grees cos A

Resolving the element adq of the given circle a into two 0

15
0.2500
0.2333
0.7500
0.7667
0.4330
0.4378
0.2500
0.2415
1.2500
1.2415
0.89 1.187
0.90 1.174
components, one equal to aldq$ along the circumference 30
45
0.1875
0.1250
0.8125
0.8750
0.4507
0.4677
0.2165 1.2165 0.95 1.169
0. 1768 1.1768 1.10 1.257
of circle C and the other along its radius, the flux link- 60
75
0.0625
0.0167
0.9375
0.9833
0.4841
0.4958
0.1250
0.0647
1.1250
1.0647
1.26 1.344
1.45 1.475
ages of the latter with the field of circle A are equal to 90 0 1 0.5000 0 1.0000 1.60 1 .600
105 0.0167 0.9833 0.4958 -0.0647 0.9353 1.77 1.800
zero. Thus the flux linkages per ampere of ado with the 120 0.0625 0.9375 0.4841 -0.1250 0.8750 1. 86 1 .984
135 0.1250 0.8750 0.4677 -0.1768 0.8232 1.89 2.160
circle A are the same as those of the element a1d4) of 150 0.1875 0.8125 0.4507 -0.2 165 0.7835 1.88 2.314
165 0.2333 0.7667 0.4378 -0.2415 0.7585 1.85 2.413
circle C with circle A. From symmetry, the mutual in- 180 0.2500 0.7500 0.4330 -0.2500 0.7500 1.83 2.440
ductance of element aid4) and circle A is ajd4)1/2r times
the total mutual inductance m of the circles A and C. by interpolation between the curves. Denoting the
It is found that do, = 1/P2 cos Od4, so that (8) may be ordinates as yo, yi, Y2, , Y12, Simpson's formula gives
written for the integral the formula
cOS r 1 /m\
S = 1/3(r/12) [2(yo + Y2 + y4 + * + Y12)
M

A 7r J P2tA)4) (16)
+ 4(y, + y3 + y5±**+ * + Yll)-(YO +Y12)] (17)
This expression may be calculated by mechanical quad- and the corresponding formula yielded by the Weddle
rature by calculating values of m for coaxial circles of rule is
radii A and aP with spacing d +a sin cos o. Referring
dimensions to the larger radius A, the parameters of m W = 3/10(r/12) [5(yi + ys + Y7 + yll) + 6(y3 + Y9)
are a' =aP and 3'= +a sin 0 cos 4. + (YO + y2 + y4 + * ±+YY12) + Y6]- (18)
To avoid the necessity of calculating the modulus k' in Applying these formulas to the calculated ordinates
order to find f so as to calculate m, the following graphi- the numerical work may be systematized in some such
cal method of procedure is recommended. The accuracy manner as shown in Table VI.
attainable is a few parts in a thousand. TABLE VI
The curves of Figs. 3 and 4 are extensions of those
given by Curtis and Sparks6 in their paper on coaxial Y2n Y2n+l YO +Yi2 y, = 1. 174 5 X6.862 =34.310
1.187 1.174 1.187 =1.475 6 X3.417 =20.502
circles. Ordinates are values of the spacing parameter 3' 1.169 1.257 2.440
y

y7=1.800 2Y2n = 12.038

1.344 1.475 yll =2.413 Y6 = 1.600
and abscissas are a'. The circle of radius A has for its 1.600 1.800 3.627
1.984 2.160 6.862 68.450
axis the Y axis and it passes through the points (1, 0) 2.314 2.413
2.440 y3 =1. 257
and (-1, 0). The curves are the loci of the co-ordinate 10.279 S 1 y9 =2. 160 W 1
12.038 (61.565) (68.450)
pairs (a', 3') of circles which have the same mutual in- -=-
7r 36 3.417
-=-
r 40
24.076 =2y.Y2n =1.711
ductance with circle A. Actually, the parameters of the 41.116 =42;Y2n + =1.710
curves are values of m/A, that is, of the ratio of the 65.192
3.627 =yo+yI2
mutual inductance in abhenries to the radius of the
61.565
larger circle A in centimeters. It is evident that the
curves give a picture of the flux distribution around a
circular filament which cuts the axial plane (plane of the Using only six equally spaced ordinates, the value
paper) in the point (1, 0) and whose axis is the Y axis. given by Simpson's rule is 1.701. Applying the correction
Fig. 3 is a general extended view of which Fig. 4 is one 1/15(1.710-1.701) to the twelve point value the result
suitable for points near the circular filament. is S/ir= 1.711.
Example 3: Therefore, M/A=1.711 cos 6=1.482 and M=20
As an example of the use of formula and the curves, (1.482) 29.64 abhenries. The value found from (4)
=

the mutual inductance will be calculated for a pair of using the tables for coaxial circles is M/A = 1.6994 cos 6.

circles of radii a= 10 centimeters and A =20 centi- Formulas for applying the graphical method to other
cases follow.
meters with the center of the smaller circle on the axis of
the larger and with a distance between centers of 20 Parallel Circles, Fig. 1:
centimeters. The axes will be assumed to have an in- M I , (1-(p/a) r
cos 4))
clination of 30 degrees. A~~~~~d (19)
Here
a = 10, A = 20, d= 20 where V2 = (1 -2 (p/a) cos 4)+p2/a2) and the co-ordinates
a = 1/2, = 1, 0 30 degrees,
=
sin 0 = 1/2 to be used in obtaining m/A from the curves are x ab- =

p2 = 1 (1/4) cos2
- 4, 3' = 1 + (1/4) cos q5. scissas = a V and ordinates y = a = d/A.
1944 Grover: Mutual Inductance of Circular Filaments 629

Circles with Axes Intersecting at a Point x = p, z = d. Fig. 2: 41 = 90 degrees (inclined axes but with the axes in
The mutual inductance is given by parallel planes, separated by a distance p.)
M 1 r r (Cos a-(pla)
MA J~T (cos6-(p/a) 42
cos) /m\
cos .(-)d4 abhenries (20)
0=90 degrees (axis of one circle in a plane perpendicu-
lar to the axis of the other).
For these cases the general formula is readily modified
with Q2= (1-COs2 q5 sin2 0-2 (p/a) cos X cos 6+p2/a2) to take these conditions into account.
and the curves are to be used in finding m/A graphically
with the co-ordinates x =aQ and y-= =3-a sin 0 cos 4.
The formula to be used if the tables for coaxial circles
are to be employed is
Aa r T (cos0a-(p/a) cos+)) fd4 microhenries (21)
k /
O 0

M= J\IA_~_ (cos 0 (pl a) cos q5) \/l

with the argument
k/2k =
(1 - aQ)2
(1 + aQ)2 + P2
+ t2
T)~~~~~
Most General Case. Fig. 5:
Take the origin at the center of the larger circle, with A 0

the axis of Z along the axis of this circle. The XZ plane

is passed through the center O' of the smaller circle of Fig. 5-Co-ordinates of circular filaments pl'aced
radius a. The co-ordinates of O' are x=p, y=0, z =d. To in any desired positions.
orient the axis of the circle of radius a, imagine a sphere All these formulas apply, strictly, only to circular
taken with O' as center. The trace P of this axis is given filaments, and accurately enough to coils whose cross-
by the longitude i1, reckoned clockwise from the XZ sectional dimensions are negligible, compared with the
plane and a colatitude 0, reckoned from the axis O'Z' other dimensions. For coils of appreciable cross sections,
taken parallel with OZ. a first approximation is obtained by calculating the
The general formula for the graphical method is mutual inductance of the filaments passing through the
found to be
M 1 j 2 (m) [cos 0- (p/a) (cos cos 4 - sin 4 sin 4) cos0)]
R2
Gdo abhenries (22)
A 27r J A
in which centers of the cross sections and multiplying this value
R2 = (1 - cos2 4 sin2 0) by the product of the numbers of turns in the two coils.
+ 2(p/la) (sin ,6 sin 4-cos cos 4) cos 0) + p2/a2 (23) A better approximation could be obtained by replacing
and the co-ordinates used with the curves are each coil by more than one filament, calculating the
x = aR, y = 6-a sin 0 cos 0 =. mutual inductances of the different pairs by the formulas
of this article, and obtaining an average by some such
If the tables for coaxial circles are to be used, the expres- expression as the Rayleigh quadrature formula.'3'14 In
sion is
-v/Aa2f
a 2r2[COS 0 -(p/a)(cos At cos4 - sin 4A sin 4 cos 0)]
o 6 o R'12 f dq microhenries (24)

and f is to be taken from the tables with the argument general, however, the difficulty of measuring the actual
- (1 - aR)2 + ¢2 (25) dimensions of the coils would hardly justify the labor of
(1 + aR)2 + ¢2 such a refinement.
The general case includes in additon to the special 13 Gray, "Absolute Measurements," vol. 2, part 2, p. 403.
14 Bull. Bur. Stand., vol. 8, p. 34; 1912, or Scientific Paper 169,
cases already considered, two others; namely, p. 34.