ShieldingTheory

Asktheaverageengineeronthestreetaboutcontrollingelectromagneticinterferenceandtheresponsewillprobably
involveshielding.Virtuallyallhighspeedelectronicdevicesemployshieldinginsomeform.Computers,cellphones,
video games, industrial controls, automotive and avionic systems, etc., all typically come packaged in metal (or
metalized)enclosuresorhaveshieldslocateddirectlyoverspecificcomponentsontheirprintedcircuitboards.
Shielded enclosures that are properly designed and installed can be a very effective means of attenuating radiated
emissionsandprotectingproductsfromexternalsourcesofinterference.Infact,ametallicenclosurewithnoapertures,
seamsorcablepenetrationscantypicallyreduceradiatedemissionsandimproveradiatedimmunityby40dBormore.
Inotherwords,evenapoorlydesignedcircuitboardcanmeetEMCrequirementsifitissealedinametalbox.
However, shielded enclosures are a poor substitute for good EMC design at the board level. Effective enclosures can
addsignificantcostandweighttoaproductandasinglebreachoftheenclosure(e.g.anunfilteredcablepenetration)
can completely eliminate any benefit the enclosure would otherwise provide. In many cases, a product in a poorly
designedshieldedenclosurewillradiatemore(orbemoresusceptible)thanthesameproductwithouttheenclosure.
Shields work by reflecting, absorbing or redirecting electric and/or magnetic fields. It is not always necessary for a
shieldtocompletelyencloseaproductinordertobeeffective.Forexample,partialshieldsareoftenutilizedtoredirect
fields on or above a source circuit to isolate it from another circuit or to prevent coupling to cables or other
unintentionalantennas.
Choosingproperlocation,orientationandmaterialforashieldrequiresaknowledgeofthetypeoffieldbeingshielded
and the objectives of the shield. The following sections will describe basic shielding theory and provide several
examplesofgoodshieldinginvarioussituations.

PlaneWaveShieldingTheory
When an electromagnetic wave propagating in one material encounters another material with different electrical
properties,someoftheenergyinthewaveisreflectedandtherestistransmittedintothenewmaterial.Forexample,
considertheelectromagneticplanewave,Einc,incidentuponaninfiniteslabofmaterialasillustratedinFigure1.The
wavepropagatesinfreespaceinthexdirectionuntilitstrikesthematerial,whichhasintrinsicimpedance,s.

Figure1:Planewaveincidentonashieldingmaterial
Themagneticfieldintheplanewaveisperpendiculartotheelectricfieldandhasamplitude,

(1)

where

istheintrinsicimpedanceoffreespace(~377ohms).

Whentheplanewavestrikestheslab,areflectedwave,Eref,andatransmittedwave,Eslab,arecreated.Themagnetic
fieldintheshieldingmaterialisrelatedtotheelectricfield,
(2)

Inaddition,theboundaryconditionsonthesurfaceatx=0requirethat,
(3)
and
(4)
wherethesubscriptsx=0andx=0+indicatethefields,justtotherightorleftofthex=0surface.Inordertosatisfy
Equations(1)through(4),theamplitudeofthereflectedfieldmustsatisfytherelation,
(5)
whereEistheelectricfieldreflectioncoefficient,
(6)

Theamplitudeofthetransmittedfield,Eslab,is
(7)
where
(8)

istheelectricfieldtransmissioncoefficient.
Notethatassgetscloserto0,thetransmissioncoefficientincreasesandthereflectioncoefficientdecreases.Ifs=
0,alloftheincidentfieldistransmitted.
IfthematerialinFigure1islossy,(i.e.

),thetransmittedwavewilldecreaseinamplitudeasitpropagates,
(9)

whereistheskindepthofthematerial.Forhighlossmaterials,
(10)

Figure2:Planewaveincidentonafinitethicknessshieldingmaterial
NowconsiderthefiniteslabofshieldingmaterialillustratedinFigure2.Anincidentfield,Einc,strikesthesurfaceof
theshieldingmaterial.Someofthepowerinthefieldisreflectedandsomecontinuesintothematerial.Thepartthat
penetratesintothematerialisattenuatedbeforeitstrikesthesecondsurfaceatx=t.Atthatpoint,onceagainsomeof
the power is attenuated and some of the power is transmitted. If the attenuation is high, the power reflected at the
secondinterfaceisabsorbedandthefieldtransmittedtotheregionoffreespaceontherightoftheslabisgivenby,
(11)
where
(12)

Combining(7),(8),(9),(11)and(12)weobtainanexpressionforthetransmittedelectricfieldintermsoftheincident
field,
(13)

This expression applies to any shield material that is much thicker than a skin depth. Typically, the best planewave
shieldswillbegoodconductorswithahighconductivity,
.Forgoodconductors,
(14)

Forthesematerials,

andEquation(13)reducesto,
(15)

Ifwedefinetheshieldingeffectivenessoftheslabtobe,
(16)

thentheshieldingeffectivenessofaninfinitesheetofgoodconductorcanbewrittenintheform,
(17)

wherethetotalshieldingeffectivenessisobservedtoconsistoftwoterms.Thereflectionloss,R(dB),istheattenuation
duetothereflectionofpowerattheinterfaces.Theabsorptionloss,A(dB),istheattenuationduetopowerconvertedto
heat as the wave propagates through the material. A webbased calculator for determining the planewave shielding
effectivenessofvariousmaterialscanbefoundat,
http://www.cvel.clemson.edu/emc/calculators/SE_Calculator/index.html.
The reflection loss is independent of the thickness of the shield and depends entirely on the mismatch between the
shield'sintrinsicimpedanceandtheintrinsicimpedanceoffreespace.Theabsorptionlossisdirectlyproportionalto
thethicknessoftheshieldexpressedinskindepths,
(18)

Example1:CalculatingShieldingEffectivenessofCopperFoil
Calculatetheshieldingeffectivenessofasheetof2milcopperfoil,=5.7x107S/m,at100MHz.

Westartbycalculatingtheskindepthincopperat100MHz,
(19)

The material thickness (t = 2 mils = 50.8 m) is clearly much greater than the skin depth so (17) can be used to
calculatetheshieldingeffectiveness.Infact,theabsorptionlosscanbeeasilycalculatedas,
(20)

Tocalculatethereflectionloss,weneedtodeterminetheintrinsicimpedanceofcopperat100MHz,
(21)

Thenthereflectionlossisquicklydeterminedtobe,
(22)

Theoverallshieldingeffectivenessisthesumofthereflectionlossandtheabsorptionloss,
(23)
Notethatvirtuallyalloftheincidentpowerisreflectedbytheshield.154decibelsisaverylargeratio,suggestingthat
thetransmittedpowerissmallerthantheincidentpowerbyafactorof1015.Inpractice,attenuationsofthismagnitude
areneitherrealizablenormeasurable.Thelargestrealizablefieldstrengths(withoutcausingionizationoftheair)are
ontheorderof106V/m.Thesmallestdetectablefieldstrengths(usingsensitivefieldprobes)areontheorderof10
6 V/m.Thisrepresentsapossibledynamicrangeof,

(24)

As a practical matter, most engineering test equipment has a maximum dynamic range of around 80 120 dB.
Therefore any calculated attenuation or shielding effectiveness much higher than 100 dB implies the material is
essentiallyimpenetrable.Amaterialwithacalculatedshieldingeffectivenessof154dBisessentiallynobetterorworse
thanamaterialwithacalculatedvalueof120dB.

IfthematerialinFig.2isnotthickrelativetoaskindepth,someoftheenergythatreflectsoffthesecondinterface(at
x=t)propagatesbackintotheslabandisreflectedofftheinsideofthefirstinterface(atx=0+).Thisenergywillthen
again strike the second interface and some fraction will be transmitted adding to the total energy transmitted and
reducing the shielding effectiveness. The wave may bounce back and forth multiple times before attenuating to the
pointwhereitnolongercontributessignificantlytothetransmittedfield.Iftheabsorptionlosstermin(17)islessthan
about15dB,theaccuracyoftheshieldingeffectivenessestimateiscompromisedbythesemultiplereflections.
Forconductivematerialsthatareelectricallythin(i.e.t<<), we can adjust the expression for shielding effectiveness
(17)byaddingathirdtermtoaccountformultiplereflectionsresultinginageneralexpressionforplanewaveshielding
effectiveness[1,2],
(25)

Note that the multiple reflection loss term has a negative value and reduces the overall shielding effectiveness. This
termissometimesusedtomakeminorcorrectionstotheexpressionin(17),butitmaynotbeaccurateforthinorlow
lossmaterials(i.e.when(t<).Itdoeshoweverprovideanindicationofwhenthehighlossassumptionusedtoderive
(17)hasbeenviolated.Ifthemultiplereflectionlossfactoriscomparabletothereflectionloss,thenneithershielding
effectivenesscalculation[(17)or(25)]isaccurate.

NearFieldShielding
Planewave shielding theory conveniently permits us to calculate a shielding effectiveness value for any shielding
material based on its material properties and thickness. Unfortunately, practical shields are never located in the far
fieldofboththesourceandreceptorcircuits.Becauseofthis,weareveryunlikelytohaveplanewavepropagationon
both sides of the material and the calculated shielding effectiveness will not correspond to anything we are likely to
measure(exceptinspeciallydesignedtestfixtures).
In order to help understand how nearfield shielding differs from planewave shielding, consider the configurations
showninFigure3.InFigure3(a),theincidentplanewavehasbeenreplacedbyasmallelectricdipolesourceandthe
shieldingmaterialislocatedinthenearfieldofthesource.InFigure3(b),thesourceisamagneticdipole,represented
byasmallloopofelectriccurrent.

Figure3:Shieldingelectricandmagneticdipolesources.

Recallthatinthenearfield(r<<),anelectricdipolesourcehasastrongelectricfield.Thewaveimpedanceinthe
nearfieldisapproximately,
(26)

Inthenearfieldofamagneticdipolesource,themagneticfielddominatesandthewaveimpedanceisapproximately,
(27)

WecanestimatetheshieldingeffectivenessoftheslabinFig.3,bysubstitutingthewaveimpedance(
)fortheintrinsicimpedanceoffreespace,0,in(25).Thisyieldsanewexpressionforthereflectionlossterm,
(28)

The expressions for absorption loss and multiple reflection loss are unchanged. Although this type of shielding
effectivenesscalculationisasimpleapproximationthatdoesnotcorrespondtoanyparticularrealizableteststructure,
itcanprovideagreatdealofinsightrelativetotheperformanceofvariousshieldingmaterialsinrealisticsituations.
Thereisanearfieldshieldingeffectivenesscalculatorbasedontheseequationsonthewebat,
http://www.cvel.clemson.edu/emc/calculators/SE_Calculator/nearfieldSE.html.
Example2:ShieldingaLowFrequencyMagneticFieldSource
A transformer generating primarily a magnetic field is located 10 cm from a shielding structure. The shielding
structureismadefroma1cmthicksheetofcopper.Estimatetheshieldingeffectivenessofthisstructureat1.5kHz.
Ifwestartbymodelingthetransformerasamagneticdipolesource,wecanquicklyestimatethewaveimpedanceatthe
positionoftheshieldtobe,
(29)
Theintrinsicimpedanceandskindepthofthecopperare,
(30)
(31)

Thecalculatedshieldingeffectivenessistherefore,
(32)

Notethatinthiscasetheabsorptionlossplaysanimportantroleintheoverallshieldingeffectiveness.Generallyatlow
frequencies close to a magnetic field source, the wave impedance is low and therefore the reflection loss due to
conductive shields is less significant. Absorption loss also decreases as frequencies get lower, but not as quickly as
reflectionloss.

ShieldingEffectivenessMeasurements
Planewaveshieldingeffectiveness
As discussed in the previous section, the concept of planewave shielding effectiveness is convenient because it is a

function of only the material properties and thickness of a shielding material. Attempts to measure the planewave
shieldingeffectivenessgenerallyinvolvelaunchingaguidedTEMwaveinacoaxialtestfixturecontainingasampleof
thematerial,asillustratedinFig.4.

Figure4.ShieldingEffectivenessTestFixture
The transmission line structure has a specific characteristic impedance (usually 50 ohms). The crosssectional
dimensions are scaled up in the midsection of the test fixture in order to accommodate a reasonably sized material
sample,whichisdiskshapedwithaholeinthecenter.Themeasuredshieldingeffectivenessissimplycalculatedas,
(33)
When measurements are made with a network analyzer, the shielding effectiveness can be conveniently express
in
termsofthesparametersas,
(34)
Notethateventhoughthecharacteristicimpedance(ratioof)ofthetestfixtureis50ohms,theratioof|E|to|H|is
stilldeterminedbytheintrinsicimpedanceofthemedium(ohmsinair).
Othershieldingeffectivenessmeasurements
Ofcourse,theeffectivenessofashieldedenclosuremaybeverydifferentfromtheplanewaveshieldingeffectivenessof
the material from which the enclosure is made. Many factors influence the effectiveness of a shielded enclosure
including the size and shape of the enclosure and the type and location of the source. Also, typically power escaping
throughaperturesandseamsinarealenclosureismuchmoresignificantthananypowerpropagatingdirectlythrough
theenclosurewalls.
Forthisreason,itisusuallymorepracticaltodefinetheshieldingeffectivenessofanenclosureasfollows,
(35)
For example, suppose the measured radiated field from an electronic product was measured with no enclosure (or
a
plasticenclosure)andfoundtobe52dB(V/m).Thensupposethatthesameproductweretestedinexactlythesame
mannerwithametallicenclosureandthemeasuredfieldstrengthwas38dB(V/m).Theshieldingeffectivenessofthe
enclosureinthisparticularconfigurationwouldthenbereportedas,
(36)
Thisisprobablyamuchlowervaluethantheplanewaveshieldingeffectiveness,butitaccountsfortheleakagethrough
apertures and seams. It also takes into account the fact that shielded enclosures generally interact with the enclosed
sources and the enclosure itself becomes an integral part of the unintentional antennapathconvertingcurrentsinto
radiatedfields.

QuizQuestion
Theshieldingeffectivenessofanenclosuremadeofamaterialthathasaplanewaveshielding
effectivenessof60dBis,
a.)~60dB
b.)alwayslessthan60dB
c.)usuallygreaterthan60dB
d.)sometimeslessthan0dB
Recallingthepreviousdiscussionofunintentionalradiationsourcesandantennaefficiency,itshouldbeclearthatan

inefficientradiationsource(e.g.anelectricallysmallcircuit)canbecomemanyordersofmagnitudemoreefficientby
couplingtoalargerconductingstructure.Thereforeitisnotonlypossible,butcommon,forashieldingenclosurewith
aperturesorseamstoincreasetheradiatedemissionsduetoinefficientsourcesenclosed.Inotherwords,theshielding
effectiveness of a shielded enclosure can easily be less than 0 dB (i.e. the enclosure amplifies the radiation) at some
frequencies.Hopefully,thesameenclosurealsoreducestheefficiencyofthestrongestsourcessothattheneteffectisa
reductioninthemaximumradiatedemissions.Nevertheless,itisnotsafetoassumethatsomeshieldingisbetterthan
noshielding.AdiscussionofpracticalshieldingtechniquesforsolvingrealworldEMCproblemscanbefoundinthe
nextsection.

References
[1]H.Ott,ElectromagneticCompatibilityEngineering,JohnWiley&Sons,NewYork,2009.
[2]C.R.Paul,IntroductiontoElectromagneticCompatibility,2ndEd.,WileySeriesinMicrowaveandOptical
Engineering,2006.
Updatesorcorrectionstothiswebpageshouldbeemailedtowebmaster@LearnEMC.com.
ReturntoLearnEMCTutorialsPage.