ELECTROMAGNETIC COMPATIBILITY

Electromagnetic Compatibility (EMC) - the ability of a system tooperate in its intended environment without

(1) suffering unacceptable degradation in performance due to coupling from other systems or

(2) causing unacceptable degradation in the performance of othersystems via coupling.

Electromagnetic Interference (EMI) - undesirable signals coupledfrom one system (emitter) to another (receptor) which degrade theperformance of the receptor. The emitter-receptor systems in an EMIproblem are sometimes referred to as the threat-victim systems.

Effective EMC design has become a critical component in the designof most modern electronic devices. The electromagnetic interferenceenvironment is becoming increasingly cluttered as more small high-speed wireless devices are introduced into the marketplace. In order to reduce theamount of interference, EMC standards (commercial and military) havebeen introduced. These standards set prescribed limits on the amount ofelectromagnetic energy that a device can emit at specific frequencies. Some standards also prescribe the susceptibility levels for the operation ofcertain devices.

Manufacturers of electronic devices must certify that these devicesmeet the appropriate standards in order for these products to be marketed. In many cases, unforseen EMC problems are identified in the producttesting phase. This requires modification of the product design (increasingthe design complexity through the addition of components) whichinherently decreases the product reliability. Correcting EMC problemsafter product testing also increases the time-to-market. The implementationof basic EMC design principles in the initial product design phase is amuch more cost-effective approach to meeting EMC standards. Thus, thedesign engineer should be knowledgeable in the basic principles ofeffective EMC design.

The fundamental EMC coupling problem can be decomposed intothree components as shown below in the emitter-path-receptor EMCmodel.

Emitter-Path-Receptor EMC Model

The identification of the three individual components of the emitter-path-receptor model in an EMC problem is not always trivial. The receptoritself may be a subsystem in a complex system. Note that the emitter andthe receptor may be associated with two independent systems or both couldbe subsystems in a larger system (subsystems on a crowded printed circuitboard). Once the receptor is identified based on its inability to functionproperly, the emitter can be located by analyzing the characteristics of theenergy received by the receptor. The properties of the interference signalsproduced in the receptor are affected by the emitter characteristics(amplitude, spectrum, etc.) and the properties of the coupling path (thecoupling path may act like a filter). The problem may be furthercomplicated by the fact that there may be multiple coupling paths in a givenEMC problem.

The three components of the emitter-path-receptor EMC modelsuggest that the effects of EMI can be reduced by

(1) suppressing emissions,(2) reducing the efficiency of the coupling path, or(3) reducing the susceptibility of the receptor.

The effects of EMI can be minimized by applying all three reductiontechniques in concert. Depending on the EMC problem, some of these EMIreduction techniques may not be applicable. For example, the emitter mayassociated with an independent system producing intentional signals forthat system.

ReceptorEmitter Coupling path

The coupling paths encountered in an EMC problem (in the emitter-path-receptor model) can be classified according to the couplingmechanism.

Coupling Path Classifications

(1) Conductive coupling - a conductive path exists between theemitter and the receptor (power cords, ground returns, interfacecables, cases, etc.)

(2) Radiative coupling - no conductive path exists between theemitter and the receptor (electromagnetic coupling), thereceptor lies in the far-field of the emitter, the emitter “radiationfield” decays as 1/R where R is the separation distance betweenthe emitter and the receptor.

(3) Inductive (magnetic) coupling - no conductive path existsbetween the emitter and the receptor (electromagneticcoupling), the receptor lies in the near-field of the emitter wherethe magnetic field is dominant, the proximity of the emitter andreceptor leads to “mutual coupling” (the emitter radiation isaffected by the presence of the receptor).

(4) Capacitive (electric) coupling - no conductive path existsbetween the emitter and the receptor (electromagneticcoupling), the receptor lies in the near-field of the emitter wherethe electric field is dominant, the proximity of the emitter andreceptor leads to “mutual coupling” (the emitter radiation isaffected by the presence of the receptor).

Note that the coupling mechanisms described above can be generalized intotwo simple classifications: conduction or radiation (radiative, inductive andcapacitive coupling are all due to radiated fields, only the emitter-receptorseparation distance and emitter field characteristics are different). Usingthe general coupling classifications of “conducted” and “radiated”, we mayclassify the general EMC problem into one of four subgroups, based onwhether the device under test (DUT) is the emitter of conducted or radiatedcoupling or the receptor of conducted or radiated coupling. The DUT isalso referred to as the equipment under test (EUT).

DUT is the source ofconducted emissions

DUT is the receiver ofconducted emissions

DUT is the source ofradiated emissions

DUT is the receiver ofradiated emissions

Radiation

Radiation

EMC Problem Classifications

1. Conducted Emissions

2. Radiated Emissions

3. Conducted Susceptibility

4. Radiated Susceptibility

Receptor

Receptor(DUT)

Emitter(DUT)

Emitter

Receptor

Receptor(DUT)

Emitter(DUT)

Emitter

Conduction

Conduction

Examples (EMC Problem Classifications)

A switched-mode power supply generates noise signals at the supplyswitching frequency and its harmonics. These noise signals areconducted onto the AC power line. (conducted emissions)

A DC-DC converter must operate in an environment where the signalat the input connection is characterized by a DC signal plus AC noiseat a particular frequency. The DC-DC converter must provide an ACrejection level of 50 dB at the noise frequency. (conductedsusceptibility)

The DC motor of a kitchen blender generates wideband noise due tothe arcing that occurs as the motor brushes make and break contact. (radiated emissions)

A carrier-based military aircraft is illuminated by the high-powersearch radar of the carrier under normal operations. The missilesmounted under the wings of the aircraft must not activate. (radiatedsusceptibility)

PHYSICAL AND ELECTRICAL DIMENSIONS OF

COMPONENTS IN EMC PROBLEMS

The ability of an EMC component to operate as a radiator (emitter)or a receiver (receptor) of electromagnetic energy depends on the electricaldimension of the component. The electrical dimension of an EMCcomponent depends on the physical size of the component and thefrequency of operation (wavelength). Thus, the electrical dimension of acomponent is measured in wavelengths. The wavelength of anelectromagnetic wave actually depends on the type of wave. We choosethe wavelength of a uniform plane wave as the standard measure since itswavelength is representative of most electromagnetic waves .

Uniform Plane Wave

(1.) E and H lie in the plane z to the direction of wave propagation.(2.) E and H are z to each other.(3.) E and H are uniform in the plane z to the direction of wave

propagation.

The wavelength of the uniform plane wave depends on the electricalcharacteristics [ó - conductivity (S/m), ì - permeability (H/m), and å -permittivity (F/m)] of the media through which the wave travels. The wavepropagation characteristics of the wave are defined by the wavepropagation constant (ã). The propagation constant of a uniform planewave at a given frequency f traveling through an arbitrary medium is givenby

where ù = 2ðf is the radian frequency (rad/s) of the wave, á is theattenuation constant of the wave (Np/m), and â is the phase constant of thewave (rad/m). The attenuation constant and the phase constant of the wavemay be written as

The velocity of propagation v and the wavelength ë of the uniform planewave are both defined in terms of the wave phase constant.

The wavelength is the distance (m) the wave travels as the fields of thewave experience 2ð radians of phase shift (one cycle). Note that thevelocity of propagation and the wavelength are both inversely proportionalto the phase constant â. Given uniform plane waves at a particularfrequency propagating through two media (medium 1 and medium 2) with

1 2 1 2phase constants â and â such that â > â , the waves in medium 1 willtravel at a slower velocity of propagation and a shorter wavelength than thewaves in medium 2.

For the special case uniform plane wave propagating in a losslessmedium (ó = 0), the phase constant reduces to

while the velocity of propagation and wavelength reduce to

The overall permeability and permittivity of the medium are defined as

o owhere (ì ,å ) are the permeability and permittivity of free-space (vacuum)

r ro o[ì = 4ð×10 H/m, å = 8.854×10 F/m] and (ì ,å ) are the relative!7 !12

permeability and permittivity (unitless). The velocity of propagation for auniform plane wave in free space is the speed of light.

The velocity of propagation and the wavelength of the uniform plane wavein a general lossless medium can be written in terms of the free spacevelocity of propagation.

r rFor any media with ì > 1 and/or å > 1, uniform plane waves propagate atspeeds slower than the speed of light and wavelengths shorter than thosefound in free space.

The electrical dimension of an EMC component located in aparticular lossless medium is determined by taking the ratio of the largestphysical dimension (�) to the wavelength.

The component is electrically small if the largest physical dimension ismuch less than the wavelength (�/ën1). As a rule of thumb, we choose themaximum value of �/ë to be 0.1 in order for the component to be classifiedas electrically small.

If the EMC component is electrically small, the operation of the componentcan be accurately defined using circuit concepts (Kirchoff’s voltage law,Kirchoff’s current law, etc.). The lumped-element circuit equations aresimply a special-case of Maxwell’s equations based on low-frequencyapproximations for the circuit elements. An electrically small EMCcomponent consisting of current carrying conductors will be an inefficientemitter or receptor of electromagnetic energy. For EMC components whichare not electrically small (�/ë$0.1), we must use field equations(Maxwell’s equations) rather than circuit equations to characterize thecomponent operation.

COMMON EMC UNITS

The quantities most often encountered in EMC applications are circuitvalues of voltage (V) and current (A) as seen in conducted emissionsproblems or field values of electric field (V/m) and magnetic field (A/m)as seen in radiated emissions problems. In addition to these quantities, weare frequently interested in the overall circuit power (W) or overall fieldpower density (W/m ). In a typical EMC problem, these quantities may2

range over several orders of magnitude. For this reason, these quantitiesare normally expressed on a logarithmic scale using decibels.

inGiven a component operating with an input power P and an output

outpower P , the power gain of the device is defined as the simple ratio of theoutput power to the input power (we will assume RMS quantities).

The power gain in decibels is defined as

According to the power gain definition in decibels, every 10 dB of powergain represents an order of magnitude in the actual power ratio. If weassume that the input and output powers are delivered to equivalent

in Lresistances (R = R ), then the voltage and current gains in dB can be madeequal to the power gain in dB by choosing the scaling constant to be 20rather than the value of 10 used for the power gain.

Thus, the general formulas for power, voltage and current gain in dB are

Note that the power, current and voltage gain are always expressedas a ratio of two quantities. The magnitude of EMC quantities such asvoltage, current, power, electric field and magnetic field are commonlyexpressed in units of dB referenced to a convenient base value.

Voltage Current

Power

Electric field Magnetic field

A value of 83 dBìV is expressed as “83 dB above a microvolt” while avalue of !35 dBmA is expressed as “35 dB below a milliamp”. One specialcase is the unit of dBmW which is commonly denoted as dBm.

Examples (EMC units)

dBìV1. Convert v = 250 mV to v .

2. Convert P = 56 dBm to P in watts.

out in3. Determine P for the system shown below if P = 1 ìW.

The output power of the cascaded amplifier/attenuator system can bedetermined using the actual gains (not dB) of the amplifier andattenuator.

Alternatively, we can express the power terms on both sides of theequation above in terms of dB.

Amplifier

1,dBG = 45 dBAttenuator

2,dBG = !20 dBinP

outP

The input and output power terms in the equation above can beexpressed using any appropriate base. There is no need to manipulatethe amplifier and attenuator power gains since these terms are basedon ratios of like units. Using dBìW gives

Using dBmW gives

TRANSMISSION LINE THEORY

The equations defining the phasor (frequency domain) voltage andcurrent as a function of position along the lossy uniform transmission lineof length � shown below are

where ã is propagation constant of the transmission line, is thecharacteristic impedance of the transmission line, and are thevoltage coefficients associated with the forward and reverse travelingwaves on the transmission line [all careted quantities denote complexphasor quantities].

The transmission line propagation constant and characteristic impedanceare defined by

where r is the per-unit-length resistance (Ù/m), l is the per-unit-lengthinductance (H/m), g is the per-unit-length conductance (S/m), and c is theper-unit-length capacitance (F/m).

The real part of the complex propagation constant (ã) is the attenuationconstant á (Np/m) while the imaginary part of the propagation constant isthe phase constant â (rad/m). The corresponding instantaneous (timedomain) voltage and current along the transmission line are found bymultiplying the respective phasor by e and taking the real part of thejùt

result. For convenience, the complex voltage coefficient and characteristicimpedance can be written in terms of magnitude and phase as

The instantaneous transmission line voltage and current are

The transmission line voltage and current equations can be written interms of the reflection coefficient which is the ratio of the reversewave voltage to the forward wave voltage.

The transmission line equations, in terms of the reflection coefficient are

The reflection coefficient at the load (z = �) can be shown to be

The transmission line is matched if there are no reflected waves on the line(the reflection coefficient is zero). According to the equation for thereflection coefficient at the load, the transmission line is matched if the loadimpedance is equal to the characteristic impedance. Otherwise, reflectedwaves exist on the transmission line and the transmission line is said to bemismatched.

The input impedance at any point along the transmission line isdefined as the ratio of the voltage to the current at that point.

Note that the input impedance at any point on a matched transmission lineis equal to the characteristic impedance.

The input impedance at the input terminals of the transmission line (z = 0)is

A low-loss transmission line can be accurately modeled as a losslesstransmission line in many cases. A lossless transmission line ischaracterized by r = g = 0 such that

Given that the propagation constant for a lossless transmission line ispurely imaginary (ã = jâ), there is no attenuation of the forward or reversewaves and the input impedance as a function of position along thetransmission line reduces to

and the input impedance looking into the input terminals of the losslesstransmission line is

Example (Cable loss/attenuation constant)

A certain coaxial cable is specified by the manufacturer to have a lossof 4.5 dB/100 feet at 100 MHz. Determine the attenuation constant for thecable.

The cable loss is defined as the ratio of the average power at thetransmission line input to the average power at the transmission lineoutput given a matched transmission line.

The cable loss defined in dB is

Note that the cable loss can be determined by taking the difference ofthe measured input and output powers in dB referenced to someconvenient level (given a matched cable).

EMC SIGNAL SOURCES

The connection of a source to a termination in an EMC systemcommonly includes a transmission line. Consider the signal sourceconnected to a terminated transmission line as shown below. The signalsource is represented by its Thevenin equivalent circuit (an ideal voltagesource in series with a source resistance).

In order to achieve maximum power transfer from the transmission line tothe termination, the termination impedance should be resistive and equal tothe characteristic impedance of the transmission line. The source deliversmaximum power to the input of the transmission line when the transmissionline input impedance is equal to the source resistance. Thus, a commonsystem impedance is required to obtain maximum power transfer for theoverall system(commonly 50 Ù according to industry standard) such that

In addition to the consideration of maximum power transfer, the matchedsystem requirement allows for simple connection of components withoutconsideration to cable lengths. If the termination and the transmission lineare not matched, a standing wave pattern exists on the line which wouldproduce varying transmission line input impedances depending on thetransmission line length. In addition, for swept frequency measurements,the input impedance of the mismatched transmission line would vary withfrequency as the electrical length of the transmission line would increasewith frequency.

The equivalent circuit for a matched system is shown below where theinput impedance looking into the input terminals of the transmission line

L sis defined by the equivalent load resistance R (which would equal R fora matched system).

The normal convention for EMC signal sources and measurement devicesis to define voltages using RMS values and power levels in dBm. According to the system equivalent circuit, the voltage delivered to theoutput terminals of the source and the power delivered by the source aregiven by

L sFor a matched system (R = R ), source output voltage and power are

LGiven a 50 Ù system that delivers an output voltage of V = 300 ìV (49.54dBìV), we find

It should be noted that the signal source power meter is calibrated based ona 50 Ù system. When the 50 Ù signal source is connected to a mismatchedload, the power meter does not yield the actual power delivered to the load. However, the actual output power can be determined from the meterreading. Consider the example of a 50 Ù signal source connected to a 300Ù load where the meter gives an output power of !25 dBm. (3.162 ìW). Since the meter is calibrated for a matched system, the actual value of thesource voltage can be found from the matched system equations.

LMatched System (R = 50 Ù)

Given the actual source voltage, the load voltage under mismatchedconditions can be determined by analyzing the mismatched system.

LMismatched System (R = 300 Ù)

The manufacturer specification of the cable loss (dB/length) can beused to quickly determine the power delivered to the load in a matchedsystem (source/transmission line/load). The cable gain (the inverse of the

incable loss) can be used to relate the power delivered by the source (P ) to

outthe power delivered to the load (P ).

where we have chosen 1 mW as our power reference level. This equationcan be rewritten as

10Taking 10log of both sides of the previous equation, the input and outputpower (dBm) of a matched transmission line are related by

A similar relationship for the input and output voltages of a matchedtransmission line can be determined according to

L inwhere R = R for a matched system. We may choose any convenientvoltage reference (using 1 ìV as the reference) which gives

or

10Taking 10log of both sides of this equation yields

Note that the equations relating the power levels and voltages on either endof the transmission line can be easily modified for any convenient referencelevel.

Example (Signal source/cable loss/received power)

An antenna is connected to a 50 Ù receiver through 200 m of RG-58U (50 Ù) coaxial cable. The receiver indicates an input powerlevel of !20 dBm at 200 MHz. Determine the voltage (dBìV) andpower (dBm) at the antenna/transmission line connection if the cableloss is 8 dB/100ft at 200 MHz.