52 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. EMC-10, NO.

1, MARCH 1968

Review of Circuit Approach to Calculate

Shielding Effectiveness
DONALD A. 1MIILLER, MEMBER, IEEE, AND JACK E. BRIDGES, SENIOR MEMBER, IEEE

Abstract-The shielding effectiveness of an enclosure at low fre- when compared to a wavelength. Thus, the circuit ap-
quencies can be readily computed using a circuit approach. Not only proach is especially useful for low frequency shielding prob-
does this technique include the effects of the properties of the shield
material, but it also includes the details of the geometry of the lems. By similar reasoning, the transmission line approach
enclosure. Furthermore, this approach allows a nonempirical can be used when propagation effects become important.
consideration of mesh enclosures and the effects of resistive seams It is the purpose of this paper to review the circuit
in enclosure walls. By working with the circuit analogue, penetration approach to electromagnietic shielding problems. Low-
by transient fields can also be computed. frequency electric and magnetic field shielding are dis-
Essentially the enclosure is viewed as an antenna. In the case of
magnetic shielding effectiveness, the enclosure is viewed as a short cussed qualitatively to give physical insight inito the
circuited loop antenna. In the case of electric field penetration, the shielding problem. Then the circuit approach is developed
enclosure is viewed as a fat electric dipole. Using this characteriza- for the electric and magnetic field shielding effectiveness
tion and exact solutions where available, the current distribution on of a sphere. The results are compared to exact solutions.
the outside of the enclosure is first determined. Then, based on the The applicatioin of this a)proach to the shielding of
current distribution, the penetrating fields are computed. The
equations are developed in such a way as to preserve a lumped transient fields is discussed, and results are given and com-
circuit analogue for the low-frequency region. pared to the results of exact solutions. Finally, an example
The basic circuit equations for magnetic field penetration are for the magnetic shielding effectiveness of a rectangular
rederived from a rigorous solution. Rules to estimate the rise- meshwall enclosure is outliined.
time, fall-time, and peak magnitudes of transient penetrating
fields are developed. The electric shielding effectiveness is developed II. QUALITATIVE DISCUSSION OF THE LOW-FREQUENCY
in a similar manner. SHIELDING PROBLEMI
In both cases the results of the circuit approach agree well with
those based on rigorous solutions of the electromagnetic boundary While the exact solution for electromagnetic field
conditions. The results also agree with published experimental data
on both large and small enclosures. Calculations are outlined for
pelnetration into an ideal enclosure of simple geometry is
possible by direct solution of the field problem with
both solid wall and mesh wall enclosures for the purpose of dem-
onstrating this approach to shielding problems.
appropriate bounidary conditions, the procedures involved
are often complicated and are generally of limited value
to the practicing engineer. Some simplification is possible,
I. INTRODUCTION however, by considering limiting cases and then comparing
THE usefulness of approximationis to solve certaill the results where applicable to those obtained from
classes of electromagnetic field problems is well known. scattering theory. In genieral, it would be necessary to
Conventional lumped circuit theory and transmission consider an arbitrary electromagnetic wave impinging on
line theory are just two examples of such approximations. the enclosure. However, the problem is usually greatly
While it may be desirable to solve electromagnetic simplified by assuming that the impinging wave is a
shielding problems exactly, exact solutions are available uniform plane wave.
only for certain idealized geometries, for example, One limiting case applies when the structure is quite
the infinite plane, the infinite cylinder, and the sphere. small compared to a wavelength, and the impinging waves
Furthermore, even when exact solutions are available, are essentially scattering in the Rayleigh region. Here the
they may be so complex as to be of little value to the effect of a high-impedance electric field and a low-imped-
practicing engineer. Two approximate techniques that ance magnetic field may be considered separately. Solu-
apply to shielding problems are the transmission line tions for either the electric or magnetic field distributionis
approach and the circuit approach. Just as conventional upon the surface of the enclosure may be drawn either
circuit theory is applied to circuits that have small di- from scattering theory or from the simpler case of the
mensions when compared to a wavelength, the circuit quasi-static fields. Once the external field distributions
approach to shielding can be applied to small structures are known, the internal fields may be developed.
First, the effects of a low-frequency electric field will
be considered and compared to those of a low-frequency
Manuscript received August 21, 1967. Major portions of this magnetic field to determine the relative importance of the
work were sponsored by the U. S. Naval Radiological Defense Labo-
ratory, San Francisco, Calif., under Contract N 228 (62479) 68443, fields in penetrating a typical metallic enclosure. Con-
with the direction of H. Zagorites. sider the effect of a nearly static electric field impinging
The authors are with IIT Research Institute, Illinois Institute of
Technology, Chicago, Ill. 60616 on a box as indicated in Fig. 1. The external field is
MILLER AND BRIDGES: CIRCUIT APPROACH FOR SHIELDING EFFECTIVENESS 53

v~~~~
t
a-09
0 0

Arrows Indicate Directions

And Regions Of Maximum
Current Flow

Fig. 1. Quasi-static electric field distribution an-d induced currents.

assumed to be uniform. A charge is induced on the surface enclosure. From experience it has been observed that a
of the box and collects the electric flux as shown. Each low-frequency magnetic field readily penetrates the
field line terminates on an induced charge. Since there are enclosure, whereas the effects of the electric field tend
no unnieutralized charges within the box, no field pene- to disappear as the frequency is decreased. These con-
trates the enclosure. siderations suggest that to the electric field the shield
Let the applied electric field vary as a function of time. behaves as a capacitor in series with a resistor. As the
Energy will now be coupled into the enclosure, for the frequency approaches zero, the reactance of the equiv-
charges induced will redistribute themselves and cause a alent capacitor becomes large, and the shielding also
current flow between the upper and lower edges of the box. becomes large. As the frequency increases, the current
The current flow causes a resistive drop to appear be- through the resistor and, thus, the voltage drop across
tween the upper and lower portions of the enclosure. the resistor increase, and the shielding effectiveness de-
Therefore, both electric and magnetic fields appear inside creases until the skin effect becomes important. Then the
the enclosure, with the magnetic fields arising from the fields appearing inside of the field are attenuated by an ex-
current flow. The charge induced is proportional to the ponential factor.
applied electric field and the current will be proportional Consider the effect of a low-frequency magnetic field on
to the time derivative of the charge and, therefore, a thin walled enclosure constructed of a high-conductivity
proportional to the rate of change of the electric field. material as illustrated in Fig. 2. Note that the current
As a result, the magnitude of the current flowing on the flows around the box or near the edges. This occurs be-
box will be proportional to the frequency of the applied cause adjacent eddy current elements cancel and leave a
electric field. Thus, at very low frequencies, the currents concentrated current flow near the edges, an observation
induced on the structure by the time-varying electric again confirmed by the results of scattering theory. For
field will be small but will increase directly with an all practical purposes, the small area in the center of the
increase in frequency. faces perpendicular to the magnetic field has no effect
Note that the sharp corners of the enclosure cause the since little current flows in this area.
charge to concentrate, and this tends to cause the current The structure may be considered to be a shorted turn or
to concentrate at the edges of the structure. This result is loop antenna having an inductance L and a resistance R.
suggested by the results of an analysis made of a plane The time-varying magnetic field induces a voltage in the
wave scattered by an infinite cylinder of rectangular cross shorted turn which is proportional to the frequency of the
section. [1] It also agrees with the experimental results re- applied magnetic field. At very low frequencies, the loop
ported by Kaden[2] in the investigation of the fields within current is proportional to the applied voltage divided by
a screened room. the resistance of the loop. This current and the field
In effect, the energy associated with the high-impedance associated with it are 90° out of phase with the applied
electric field is first converted into current flowing on the field. Therefore, little cancellation or reflection of the
surface of the structure. This current can then give rise to applied field can occur. As the frequency or time rate-of-
both electric and magnetic fields in the interior of the change of the applied field is increased, the inductive
54 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, MARCH 1968

Dashed- Line Indicates Induced

Currents On Structure.

H
Fig. 2. Induced currents on box from time-varyiiig magnetic field.

reactance of the turn tends to dominate, and the current In the followinig sectionis circuit anialogs will be de-
flow in the loop slowly comes into phase with the applied veloped for the electric aind magnetic field shieldiing of
field. From experience we know that some cancellation spherical enclosures using static or quasi-static solutions
of the field must occur within the shorted turn, with to the problem of a conductive sphere in a uniiform electric
augmentation of the magnetic field on the exterior of the or uniform magnietic field.
structure.
As the frequency of the applied magnetic field is further A. Electric Field Case
increased, a significant amount of energy may be absorbed Assume that a thiin-walled conductiing sphere is placed
in the walls of the enclosure by the skin-effect mechanism.
This causes the fields which appear on the outside of the in a uniform static electric field. After the charges have
enclosure to be attenuated exponentially before appearing oriented themselves as shown in Fig. 3, no electric field
on the interior surfaces for solid-wall enclosures. In addi- appears in the interior, since no unneutralized charges
tion, the skin effect causes the series impedance of the can be sustained within the box. Using the coinvenitioii of
surface to be increased somewhat, which tends to reduce Fig. 3, the induced surface charge denisity is given by [8]
the rate of increase of shielding effectiveness with increas- q(6) = 3EoEo coso (1)
ing frequency, especially for mesh or screen-wall enclosures.
Since current concentrations are somewhat localized. where 80 = 10-9/36r and Eo is the zero frequeincy out-
the magnetic shielding problem can be simplified by as- side field iintenisity. Integrating over each hemisphere, the
suming a uniform current distribution oii the exterior of magnitude of the charge on each hemisphere is founid to be
the enclosure. Under this assumption, the circuit approach q = 3-xEoEoa2 (2)
for computing the shielding effectiveness can be applied
in the low-frequency region. where a is the radius of the sphere.
To cause electric-field penetratioin, a currenit must
III. A CIRCUIT APPROACH TO SHIELDING flow anid cause atn I-Z drop. To do this, the applied field
Wheeler, [3] Sunde, [4] and Mfiller and Bridges [5] have must vary so that the charges distribute themselves as
developed expressions for the magnetic field shielding follows:
effectiveness of an enclosure using a circuit analog. q(t) = 3rE0 aIEoeJ t. (3)
Hays [6] experimentally verified Sunde's calculations.
iA,iller and Bridges compared their approximate expres- A niet charge defined by (3) flows through the equator.
sions for a sphere to exact solutions for transient penetra- By defiiition, the curreint is the time derivative of the
tion of a sphere by Harrison and Papas. [7] Bridges time-varying charge, so that the currenit through the
et al. [8] presented a circuit analog for the electric field equator becomes
shielding of a sphere and compared their results with those io(t) = j337Eroa EoeJ t.
of Harrison and Papas. They also achieved a satisfactory (4)
agreement with measurements. M\Jiedzinski [9] investigated The voltage drop for very very low frequenicies across ani
mesh-walled enclosures and developed a circuit analog equatorial hoop of thickness d and height y is equal to the
which he verified by experiment. I-R drop across this hoop. The resistanlce of this hoop is
MILLER AND BRIDGES: CIRCUIT APPROACH FOR SHIELDING EFFECTIVENESS 55

Fig. 3. Charge distribution on a conducting sphere.

(5) The outside equatorial voltage drop per ampere is also

Uad27ra (5) increased by this factor. This outside voltage drop is
attenuated before it appears on the inside, provided
where oa is the conductivity of the thin-walled sph ere. d >> 6, approximately by the factor
The voltage drop across the hoop now becomes
2e -d .
(9)
V (t) =jw3E0aE0
20rad
ej,ty (6) For the high-frequency case where d >> 6, 2ira/X << 1,
and a >> d, the interior voltage at the center as modified
Because of symmetry, the equipotential lines near the by (8) and (9) becomes
equator are parallel to the equatorial plane. Thus the elec-
tric-field initensity at the center of the sphere is the 3 V2wcoEoa d/6
same as that on a surface very close to the equator. Ei = jg~ 0-aa
e- . (10)
Therefore, the center electric-field intensity is simply the
equatorial hoop I-R drop divided by the hoop height y, The electric shielding effectiveness SE is defined as[8]:
where
E V0 .3wE0aE0 (7)
20 logolJSEE_E| = 20 logEoElJ
y 2 0rad
If the frequency is sufficiently high so that the skin depth = 20 I"lolo 3-V/2 wEoae d/6 (11)
6 is small compared to the wall thickness d, the bulk of the
current flows on the outer surface and encounters an for high frequencies where d >> 5. For low frequencies
increased impedance due to skin effect. The ratio of the where d << a
magnitude of the high-frequency to low-frequency sur-
face hoop impedance can be shown, provided (I >> a and 20 log SEE-El 20 log1o [32ooa]. (12)
=

a >> c4, to be approximately, [5], [8]

ZHF V2d (8) It is now worthwhile by means of (11) and (12) to
ZLF compute the steady-state shielding characteristics for a
typical enclosure. Since an 18-inch radius, 1/16-inch
where wall thickness aluminum sphere was used as one of the
examples where a circuit approach was developed and also
1rft
for rigorous solution based on scattering theory and
numerical integration, this example will also be employed
a

andju,U is the permeability of the wall material. to determine the steady-state shielding characteristics.
56 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, MARCH 1968

S4Ur _________
X_______
320

300

0 280
a en
>
,
w
-.

0
2 260
w- 0
C*
cn c

.-0
'a)
240

220

2n O) _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _

100 10I 102 103 104

Frequency, Hz
Fig. 4. Steady-state shieldinig effectiveness (18-inch radius aluminum sphere, '/16-inch wall thickniess) for
ulliform impinging electric field in air.

A plot of this is shown in Fig. 4. The dashed portioni of In the case of a highly conducting sphere, an effective
the curve occurs when d 5 and is sketched in between
- height is choseni to equal onie-half the diameter, the radius
the values determined by (11) or (12). If this curve a. The openI circuit potential then becomes
is compared with the graphical data based oin numerical Vo-c (t) = aEoejwt. (13)
iitegration of an exact analysis, the curves appear iden-
tical to an accuracy of about 41 dB. [7] Equation (4) may then be rewritten as:
In a circuit approach suitable for spheres, spheroids, aE Vo_c
long thin rods, and related structures, the current flowing Io =
1 -
--a- aE0
(14)
perpendicular to an equipotential p)laIle of symmetry _S -i Zsource
due to a field parallel to the major axis must first be (3wEoa) (w) Cw
determined. To do this, it is most convenienit to use an For the effective height chosenl,
analytical technique developed to determine the response
of long antennias to low-frequency transient waveforms C = 3wrEoa. (15)
having spectral compollents wholly within the Rayleigh On the basis of (15), (13), aIld (6), a low-frequency
scattering region (2ira/X << 1 or 2L7rL/X << 1 where L is circuit characterization canl be devised as shown in Fig. 5
the maximum dimenision). for 6 >> cl.
In this case, the structure or antellila is regarded as a Using the parameters chosenl for the previous example
black box Iletwork which relates the uniform impiniging and the circuit characterization of Fig. 5, the time history
field Eo(t) to an open-circuit voltage of the network. The of the interior fields to a Gaussian pulse of onie-volt-per-
ope1circuit voltage of the network is equal to the applied meter was computed and is presented in Fig. 6. For
field intensity Eo times an effective height he of the comparison, the time history based on scattering theory
structure. and numerical initegration is also plotted. Note that good
Next, the network source impedance is determiined and agreemenit exists, except that the numerical results are
used to compute the current into the equipotential plane. delayed in time and are of slightly smaller amplitude.
In the case of high-conductivity shielding structures, this This is because the simple circuit characterizationi nie-
current is essentially the short-circuit current of the net- glected the skin-effect absorption of the higher frequency
work. This current may then be used in conjunction with components. The circuit characterization could, of course,
the surface impedance to compute the illterior fields. be modified to include the skin effect.
MILLER AND BRIDGES: CIRCUlT APPROACH FOR SHIELDING EFFECTIVENESS 57

C =3ir0 a

for 8>>d and 2ad >

3e0
Fig. 5. Very low-frequency electric field penetration characterization for a sphere.

I Ar.
9..

1.2 A.. ,__....... LI(inside i8eradius,1X16"wall

\B thickness. Aluminum Sphere)
1.0 , ;_\_ for Ivolt/meter Gaussian
E(t)xe?
0.8
/ t1 48 x16lr sec.
0.6 /i t \ -~- A-Calculated usinq
approximation
C-R
0.4 / - -B-Exact
----- - ScatteringTheory
--\----
N

0 0.2
x

1$0 Ig0 80 41 1 4.0 8 12

0.2 -e t---t - -v -- --- -- - t- - -- --
-0
0
0.4

0.6

0.8

*1.0 1
-----
- --- --- - ---- ------ w

1.2

1.4..
Fig. 6. Time history of the electric-field intensity at the center of a sphere.

The above technique can be used to compute the the sphere, Hi is the internal field, and Ho is the applied
approximate electric shielding effectiveness of nearly field. The low-frequency and the high-frequency limits
cubical structures. Long structures may be characterized of this equation will be developed and shown to be iden-
as spheroids or as thick antennas. The antenna approach tical to the circuit forms. Consider the asymptotic low-
isfoften useful since the effective height and input imped- frequeincy form as follows:
ance have been developed for a number of antenna
0cyd <<
0*, 1
shapes.
sirbh yd -yd
B. Magnetic Field Case
cosh oyd 1.
It is useful to rederive the circuit approach expressions
for the shielding of a sphere. Harrison[10] cites a shielding Using these relationships and noting that a ; b for thin
relationship rigorously developed by King[11' for a spherical shells, the following is true:
shield: H1(w) 3
(17)
H(w) Ho(cw)
-yb (1 +
3
-yd +
3

Ho(w&) (y b) %I y
y2b2 1 ( +) sinh yd + cosh -ydj
3
(18)
where a is the outer radius, b is the inner radius, d is the lylbd + 3 (1 + b)
wall thickness, 'y is the propagation constant of the wall of
15X IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, MARCH 1968

Fig. 7. Analog circuit characterizing shielding effectiveniess of spheres.

Since d/b << 1,

1 1 Hi () 3do-
(19) Ho(co) - 27ruan2 2i-n2 (26)
y2bd jbdwp,c + 1 9 3d±
+ 1
3
1
For the high-frequency case the following relationships
hold: 3dor
(27)
jcua + 1
lyd >» 1, 3/-yb << 1 9 3do
sinh yd cosh -yd MIultiplying ilumerator and denominator by 3do, the
circuit form for low frequencies is found to be idenitical to
Hi~(co)_
(co)
3ya
(20) the low-frequenicy limit of King's[ll rigorous expression
Iro(co) (-yb)2Fb 1sn d in3 3
I sinh yd + -sinh ydI
L jb 20 logiolSEH H!| = 20 logod| - (28)
J 3 +1
3
(21)
yb sinh -yd ln the high-frequenicy limit, d >> 6, aind jcoLs >> Rs.
Then usinig (8) anid (9) the circuit form becomes
1Hi(o) 3V/2 - d/6 3V62bed (22)
IHo(w) bVA/ rfu a b 2 2e Hi(w) _ R?s5 dI/ (29)
Ho(co)j Ls 6
L -
The shielding effectiveness of an enclosure to low
frequency magnetic fields is given by['] Substitute (24) and (25) for Rs and Ls and note that
Hj(co) Rs a b
(23)
Ho()- Rs + jowLs and
where Rs and Ls are the resistance and inductance of the 2
shield. The equivalent circuit is shown in Fig. 7. Wheeler[I] 1J2 = (30)
gives the resistance and inductance of a spherical coil:

(24) Hi(co) 3do_ d l

3dcr
Ho()I 2 a2e
-
(31)
_

Ls =
2wr/an2 (25) 9
9
3 2e-d/ 2
(32)
where d is the wall thickness, a- is the conductivity, a is the b6 cog
radius of the sphere, and n is an equivalent number of
turns. Therefore, the magnetic field shielding effective- 3 V26e -dl/
(33)
ness of a sphere is given by b
MILLER AND BRIDGES: CIRCUIT APPROACH FOR SHIELDING EFFECTIVENESS 59

(A) Based On Results From ScatteringTheory

netic field Ho(t) and the output voltage i(t)R is propor-
O O Calculated From "Shorted - Turn" Analysis tional to the interior field Hi(t) at the center of the sphere.
(B) Form Assumed For Simplitied Tronsient AnalysisThe series inductance and resistance Ls and Rs, respec-
120 tively, are as previously defined. This characterization
omits skin effect, but could be modified to include it.
This circuit is amenable to the most elementary transient
rjj 100 analysis.
I~ -Ls lL(A) / An analytical comparison of this simplified transient
N 8
approach with the exact solutions is given in Figs. 9 and 10
for two Gaussian waveforms of equal amplitude but dif-
*60
ferent pulse width. The inicident magnetic field amplitude
is that associated with a plane wave with electric field of 1
Z;
40 volt per meter. Fig. 9 shows a fairly long time history for
c
which the exact results were available for only the first
n-
few hundred microseconds. In this figure, the results of
the two approaches appear to be virtually identical.
20
The results appear to be different only during the early
Ieb
pulse period as shown in Fig. 10. Here as might be ex-
pected, the rise time is slightly lengthened by the higher
attenuation alid longer time delays caused by the skin
effect, which is taken into account in the exact analysis.
In either case, the same peak internal magnetic field is
FreaQusp,c hr
realized before transient decay.
An approximate approach to estimate the time history
Fig. 8. Sinusoidal waveform uniform magnetic-field shielding of magnetic field penetration into a spherical shield will
effectiveness for aluminuim sphere 18-inch radius aRid '/16-inch
thick. now be developed. It is assumed that the transient mag-
netic field pulse uniformly illuminates the shield.
and If the shield is to be effective, the duration of the
external magnetic field must be short compared to the
time constant L/IR of the shield. For this condition and
logol4SEH-HI
20 = 20 3og0
boi lV2&e -=d2 neglecting skin effects, the maximum internal magnetic
field Hi(max) can easily be shown to be
This result is identical to the high-frequency rigorous
form. Hi(max) = LL Ho(t)dt (35)
The particular equation attributed to King does not
I

account for shells with permeability greater than 1. In the where T2-T1 is the approximate duration of the applied
circuit characterization, this is considered in the skin field. The rise time t4 (in this case, a > d) would be the
depth parameter &. A calculation of magnetic shielding pulse duration T2-T1. An exponential decay describes the
effectiveness was made for one of the cases considered by response after the rise as follows:
Harrison and Papas. The steady state shielding effective-
ness based oni the circuit approximation for an aluminum Hi(t > tr) =[L Ho(t)dt]exp[- L (t tr)]- (36)
-

sphere with 18-inch radius and 1/16-inch walls was

compared to the results for the same case from the exact The effect of skin effect must be considered. Skin effect
solutionis. The impinging field is assumed to be uniform. both absorbs energy and tends to stretch out the pulse
Wheni exact values are plotted as in Fig. 8, curve A, the duration. The stretching process tends to preserve the volt-
points using the circuit approach and curve A appear seconds or ampere-seconds of a narrow pulse. The stretch-
identical except for a small departure at f = 104, where ing is roughly inversely proportional to the cut-off fre-
dc ~5. Equation (23) was used when a > d and (34) for quency of a low-pass filter. If the additional attenuation
d >> a for the calculated points. introduced by skin effect as shown in (29) is viewed as a
Since transient field penetration is of interest, it is low-pass filter, then the low-pass filter has the charac-
worthwhile to compare the time histories of the magnetic teristic response FI(f) as follows:
fields appearing at the center of the sphere. While curve A
of Fig. 8 defines a transfer function, significant simplifica- F I (f) = 2 V 2 dVw 7rfg o [exp ( VV'rfAg- d) ]. (37)
tion: at a small loss in accuracy can be realized for a wide Defining the cut-off frequency at the half-amplitude point
class of pulse waveforms and shields by assuming curve B and solving for fC, the cut-off frequency
of Fig. S. An analog circuit which characterizes this trans-
fer function is the one shown in Fig. 7 where the voltage fc (2 7)2 1 (38)
source is defined as proportional to the impinging mag- d W -i
6() IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, MARCH 1968

3.0
(A) a nd (B1 2
2.8 *x p.-(t/2!L)
H0 12 0 TT
2.6
(A) tf= 48,Fs L-R Approximatior
2.4 (B) tl= 48p s Exoct Scatter-ng Theory
(C) t= 1 2 u s L -R App oximatior.
2.2
(,) t = 12 us ExacttSctterirg Theory
2.0 (A) ArJ (B) or (C) And (Di
" - (A) Only or (81 Orly
1.8

1.6

[.4
10
x 1.2 N.

1.0
(C) and (D)
.8 -
(C!
.6

.4
-~~~~~~~~
.2

IlI _ _I 2
0 5,000 10,000 5,0-io 2 0,000

Microseccnds Tirre

Fig. 9. Lonig term history of the magnetic field intensity at the center of an 18-inch aluminum sphere
with a 1/16-inch wall.

soep-(t/2t,)2
ol
O
(20 TT
(A) t a 48n si L - R opproximation
25 (B) t, * 48p.es Exact Scattering Theory
(C) t, a 12 gi e L - R approximnat ion
(D) t 12 A s Exoct Scattering Theory

20 _

1.5
*0
x

I.0

-180 -150 -100 - 50 50 200

Time ptsec

Fig. 10. Expanded early time history of the magnetic field intensity at the center of ani 18-inich radius aluminum
sphere with a '/16-inch wall to a Gaussion pulse.
MILLER AND ]BRIDGES: CIRCUIT APPROACH FOR SHIELDING EFFECTIVENESS 61

* I ~~~~~~~~~~~(A)
0 E1

C f 8>>d--,wL*> R,--
-O"

en S~~~~~~~~~~~~
RS> L (aLs>
U,

Log w
Fig. 11. Qulalitative comparison of (a) solid-wall and (b) mesh-wall enclosures having the same low-frequenicy
shielding effectiveness.

The rise and fall time or stretch-out time t, for a low pass inductance of a solenoid of round CFoss section with the
filter is roughly equal to same circumference as the rectangular enclosure. The
internal impedance per turn is also calculated. The skin
ts effect only increases the internal impedance of a turn of the
equivalent solenoid anid reduces the shielding effective-
For most ,u, ellelosure designs, and materials, espe- ness. When the leakage reactance becomes larger than the
cially those having high a, low A,, anid large radii a, the internal impedance, the shielding effectiveness becomes
following holds: indepeindent of frequency.
The circuit approximation may also be applied to a
L solid-wall enclosure (of less simple geometry than the
< -

spherical enclosure previously considered) by dividing it

Where this is the case, the rise time tr of Hi(t) is deter- into n turns of wire of diameter equal to the thickness of
mined either by t, or T2-T1, whichever is the larger. Thus, the walls, where n = fld with f the length of the enclosure
three parameters which approximately describe the rise and d the thickness of the walls. Theni the procedure is the
time, maximum interior fields, and the fall time have been same as the mesh enclosure except the attenuation due to
developed. skin effect must be included as a factor. If the enclosure
This approximate approach was checked by comparing has walls much thinner than the skin depth, the main
the results to those based on rigorous solution of the field contribution to magnetic shielding is the geometry, which
problem by Harrison and Papas and a rigorous but determines the inductance and the resistance per turn.
approximate solution by Harrison. Recently the IIT On the other hand, as the frequency is raised, the skin
Research Institute has programmed the field penetration factor soon becomes as large as the geometric factor. In
based on scattering theory. Interior fields for ferromagnetic either event, the shielding effectiveness muay then be
shells have also been developed there. In all cases, for both limited by some constructional detail such as a bad bond
ferromagnetic and nonferromagnetic shells, the results which may not only be resistive but may be shunted by a
based on the approximate equations were within a factor capacitance formed by oxides in the bond. This effect can
of two of the results obtained by numerical integration of be readily analyzed by the circuit approach. It always
the scattering theory resolution. Since the actual penetra- tends to degrade the shielding effectiveness of the en-
tion is governed by numerous nonideal parameters, there closure.
is justification to use the approximate relationships. Fig. 11 gives a qualitative comparison between the mesh
and solid-wall, steady-state shielding effectiveness.
Equation (23) gives the magnetic field shielding ef-
C. Magnetic Field Penetration of Mesh Enclosures fectiveness of a solid-wall sphere. For mesh-wall enclosures
The circuit approximation is usually applied to a mesh there will be some leakage flux which may be important at
enclosure with nearly square cross section by finding the higher frequencies. In order to include skin effect, the
62 IEEE TR ANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, MARCH 1968

equation will also be rewritten in terms of the internal IV. SUMMARY

impedance of the wires of the mesh. Equation (23) becomes A circuit approach to predict the shielding effectiveness
of an enclosure has been presented for both the electric
SEH = Z +± jLe (40) and magnetic field cases. Since the circuit equations for a
Z1 + jwLs spherical enclosure have been developed by considering
where Zi is the internal impedance of n turns of wire, limiting cases of quasi-static solutions of the electro-
Ls is the external inductance of the shield, and Le is the given to field
magnetic
this
problem, a rigorous justification has been
technique. Furthermore, the excellent
leakage inductance. At low frequencies the internal
impedance is replaced by the dc resistance of the wire. agreement between the results of the approximate tech-
Grover[12] gives expressions for the inductance and leak- nique and the results of exact scattering theory for both
age inductance of an air-filled solenoid conmposed of n shielding effectiveness and transient field penetration
turns (for rationalized units): further substantiate the validity of the circuit approx-
imation. The physical insight gained and the relative
simplicity of application commend this approximation to
Ls ;z:~KiLon2 ,
D2
(41) the working engineer. However, it should be noted that
the basic assumptions that the applied field is uniform
and and that the structure has dimensions small compared to a
wavelength were used in the development of this approach.
Le - K/uOnD (G + H) (42)
7r REFERENCES
[1] K. Mei and J. Van Bladel, "Low-frequency scattering by
where D is the diamneter, L is the length, and K, G, and H rectangular cylinders," IEEE Trans. Antennas and Propagation,
vol. AP-li, pp. 52-56, January 1963.
are tabulated correction factors. In order to apply these [2] H. Kaden, Wirbelstr?ime und Schirmung in der Nachrichten-
formulas to a square solenoid, the diameter D is replaced technik. Berlin: Springer, 1959, pp. 93-107.
[I] H. A. Wheeler, "The spherical coil as an inductor, shield,
by the diameter of a solenoid with the same perimeter as or antenna," Proc. IRE, vol. 46, pp. 1595-1602, September 1958.
the square solenoid. For low frequencies the internal [4] E. G. Sunde, "Switching center shielding against atmospheric
induction," Bell Telephone Labs. (unpublished memorandum).
impedance of the solenoid wires can be approximated by [51 D. A. M1iller and J. E. Bridges, "Geometrical effects on
shielding effectiveness at low frequencies" IEEE Trans. Electro-
Z~i nRo (43) magnetic Compatibility, vol. EMC-8, pp. 174-185, December 1966.
[61 J. B. Hays, "Protecting communications systems from EM\IP
effects of nuclear explosions," IEEE Spectrutm, vol. 1, pp. 115-122,
where Ro is the dc resistance per turn and ni is the number Mav 1964.
[71 C. W. Harrison, Jr., and C. H. Papas, "On the attenuation of
of turnis. In the high frequency case, the internal imped- transient fields by imperfectly conducting spherical shells," IEEE
ance of a round wire can be approximated byv13' Trans. Antennas and Propagation, vol. AP-13, pp. 960-966, Novem-
ber 1965.
[81 J. E. Bridges, R. G. Htuenemauin, and H. R. Hegner, "Electric
o field shielding and measurement," IEEE Electromagnetic Com-
Zi k
nRo (44) patibility Symp. Rec., pp. 173-177, July 1967.
[9] J. Miedzinski, "Electromagnetic screening theory and prac-
tice," British Electrical and Allied Research Assn., Tech. Rept.
where ro is the radius of the wire and 6 is the skiln depth. M/T 135, 1959.
]10] C. W. Harrison, Jr., "Transient electromagnetic field prop-
Then, to calculate the low-frequency magnetic shielding agation through infinite sheets, into spherical shells, and into hollow
cylinders," IEEE Trans. Antennas and Propagation, vol. AP-12,
effectiveness, (41), (42), and the appropriate form for pp. 319-334, May 1964.
the internal impedance are substituted into (40). The [ll] L. V. King, "Electromagnetic shielding at radio frequencies,"
Phil. Mag. VII, vol. 15, pp. 201-223, February, 1933.
effect of a resistive seam can be included by adding a [I] F. W. Grover, Inductance Calculations. New York: Dover,
term, which represents the seam, to the interiial imped- 1962, pp. 142-162.
113] S. Ramo, J. R. Whinnery, and T. Van Duzer, Fields and Waves
ance. in Communication Electronics. New York: Wiley, 1965, p. 294.