AgilentLow–Level RF Leakage MeasurementApplication Note 1204

2

3

IntroductionThere are many advancements intwo-way radios and pagers includinggreater sensitivities, and higher fre-quencies. A radio or pager whichcan receive very low level signals farfrom the transmitter is consideredmore valuable and can command ahigher price. Regulatory agenciesare opening channels to higher fre-quencies. In order to design and pro-duce these new products, compo-nent, device and product manufac-turers are requiring signal sourceswhich can accurately simulate thelow-level signals in the right fre-quency range which their new prod-ucts are being designed to workwith.

The impact of greater receiver sensi-tivity is subtle. The obvious need isfor sufficient output level attenua-tion to obtain the new lower test sig-nal levels. The less obvious need isfor reduced RF leakage from the sig-nal generator. The test signal shouldleave the signal generator by the RFoutput connector only. But the sig-nal may also leave the signal genera-tor by a radiated leakage path. If theradio under test is not well shielded,or if the radio's shielding covers areoff for purposes of test, the leakagesignal can enter the radio along withdesired signal from the RF outputconnector. The two signals can addeither constructively or destruc-tively, resulting in a total signal ofunknown amplitude. This reducesthe accuracy of the test.

Higher frequencies also demandmore of a signal generator. Onceagain, there is an obvious need—testsignals at higher frequencies. Buthigher frequencies also make leak-age from the signal generatorsworse. For example, a cabinet seambecomes a larger fraction of a wave-length as the frequency increases,and its shielding effectiveness goesdown. Thus, increasing leakage withincreasing frequency.

The task for the manufacturer of thesignal generator is to measure leak-age at higher frequencies and lowerlevels, while improving the accuracyof those measurements. This paperaddresses the measurement prob-lems, and briefly discusses some ofthe work done to reduce leakagefrom signal generators.

Many are familiar with the problemsof radiated RF leakage from com-puter equipment and its effects onradio and television reception.Reception can be degraded even atsome distance away, say 10 metersor more. The radio regulatory agen-cies in various countries, for exam-ple the Federal CommunicationsCommission in the United States,have set limits on the electric fieldemissions from electronic equip-ment. Agilent has generated a fieldstrength specification that assurescompliance in practically any coun-try by combining the most stringentrequirements of all these countriesinto one field strength curve.

Agilent signal generators have beendesigned to meet these same regula-tory limits. How do these regulatoryfield strength limits compare toradio sensitivities? Let's look at theportion of the regulatory field

strength curves from 100 MHz to1000 MHz, and compare it to typicalVHF and UHF radio sensitivities(.25 m V) in this same frequencyrange. Assume the radios use a reso-nant dipole antenna.

Figure 1 shows that even at 10 metersaway, emissions from equipmentpassing the regulatory requirementsfor RF leakage can be detected byradios quite easily. Typically, radiotests are done at distances much lessthis, often times within one meter ofthe signal generator.

What is this field, and how do wecompare field strength to radio sen-sitivity? How do we quantify this?

Figure 1. Regulatory field strength limits vs. typical radio sensitivities

4

The electromagnetic fields we arediscussing are made up of two com-ponents—an electric field (E), and amagnetic field (H). The strength ofthe E field is measured in volts permeter, and the H field is measuredin units of amps per meter. The ori-gin of these fields may be uninten-tional radiation, such as RF leakagefrom a signal generator or computer,or may be intentional radiation,such as from a radio transmitter. Ata distance of several wavelengthsfrom an electromagnetic fieldsource, perhaps a transmittingantenna, the magnitudes of the Eand H fields are found to be in afixed ratio:

(1)

and 377 ohms is defined as theimpedance of free space. Multiplyingthe magnitude of the E field with themagnitude of the H field yields thepower flux density in watts/sq.meter:

(2)

(3)

(4)

These are the fundamental units ofelectromagnetic fields. They allow usto relate RF leakage, regulatoryrequirements, and radio sensitivitiesto each other as we shall see.

Radio sensitivity is often specifiedas microvolts into a given inputimpedance) for example .5µV into60 ohms. It can also be specified ininput power, for example 5 fem-towatts. These sensitivity figures areusually given for 12 dB SINAD out ofthe radio. How do we convert fromfield strength to radio sensitivity?Antennas convert electromagneticfields to power (or voltage) into theradio input terminals. Let us reviewsome antenna basics. Let's start witha transmitting antenna. If we assumean isotropic antenna, power will beradiated uniformly in all directions.The power flux density (Pf) at a dis-tance r from the antenna is then;

(5)

Most antennas are not isotropic, andwill radiate more power in somedirections then others. The fieldstrength will be increased in thosedirections by the amount of the gainover isotropic. This number is theantenna gain. The power flux den-sity in the preferred direction willthen be;

(6)

where Gt = transmit antenna gain.

A resonant half wave dipole has again of 1.64, or 2.15 dB. This gainfactor holds for both transmit andreceive antennas. A transmittingdipole will radiate fields in its pre-ferred direction 1.64 time as strongas an isotropic antenna. Similarly, areceiving dipole will provide a signalat its terminals 1.64 times as strongas the signal from an isotropicantenna.

Figure 2. Dipole field strength, pattern

5

To calculate the relationship betweenfield strength and antenna terminalvoltage, we use the power flux den-sity equation (6) and the equationfor path loss between two antennas.Path loss is the reduction in power ofthe signal at the receive antenna ter-minals compared to the transmitantenna terminals. It is due to thespreading of the signal power over alarger area as the distance from thesource increases. A given receiveantenna will then intercept lesspower as this distance increases. Theequation for path loss is;

(7)

where:

Aer = Effective area, receive antennal= Wavelength in meters.

Aet = Effective area, transmit antennaGr = Receive antenna gainGt = Transmit antenna gainPf = Power flux densityPr = Received powerPt = Power transmitted by transmit

antenna

Rearranging equation (7) gives:

(8)

Note that antenna gain and effectivearea can be related by:

(9)

To calculate the amount of powerreceived by the antenna in a fieldwith a given power flux density, wetake the ratio of the received powerto the power flux density, Pr/ Pf

[ratio of equation (8) to equation (6)].

Thus:

(10)

Or equivalently:

(11)

We see that the received power doesnot depend on the nature of thetransmit antenna, but only the fieldstrength. For a resonant half wavedipole, Zin = 73 ohms (pure resis-tive).

Thus:

(12)

(13)

For a half wave resonant dipole, Gr =1.64.

Thus:

(14)

assuming a 73 ohm impedance load.We now have a simple conversionbetween field strength and voltageinto a 73 ohms load assuming adipole antenna.

Well then, are we all ready to go nowand make measurements? No! Wehave to decide at what distance weare going to measure, and what levelis acceptable. As discussed earlier,receiver tests are often done close tothe signal generator, usually lessthan one meter away. As equation(5) shows,

(5)

Higher power flux density is foundcloser to the source. Thus, maximumleakage signal is obtained close tothe signal generator. Let us assumethat receiver tests are to be con-ducted as close as 25 mm away fromthe signal generator. Can we repeatthe regulatory requirements testswith a new lower specification lineand extrapolate to our 25 mmreceiver test position? We might beable to do an extrapolation usingequation (3),

(3)

and equation (6),

(6)

Combining the two, we find:

(15)

or:

(16)

6

This analysis indicates that a magni-tude of the electric field strength |E|varies as l/r. We know from equation(14) that electric field strength andantenna terminal voltage (and thusreceiver terminal voltage) varydirectly with each other. If we desireno more than 0.25 µV of antennaterminal voltage at 25 mm away, byour analysis, what must we measureat 10 meters away to insure this?The l/r electric field strength rela-tionship indicates that we wouldfind 0.625 nV of signal at theantenna terminals. A very low noisereceiver with a bandwidth of lessthan 1 Hz would be needed for thismeasurement. To make things easier,why not just make the measurementat 25 mm away?

We must look again at our analysis.The assumptions made for the analy-sis thus far hold true only at dis-tances several wavelengths from thesource of radiation, that is, the farfield region. Several nice field prop-erties hold in the far field region;

1. The ratio of |E| to |H| is 377 ohms,independent of source.

2. The magnitude of E and H vary asl/r along a given radius line wherer is the radial distance from the source.

3. The fields are plane waves.

What happens closer in, at distancesof a fraction of a wavelength? Thisregion is called the near field region,and things get more complicated.First of all, to simplify the discus-sion, let us assume that the sourcesof radiation are point sources. Alsowe must now be concerned with twotypes of sources—an electric pointsource and a magnetic point source.

An infinitesimal dipole is an electricfield point source. An infinitesimalloop is a magnetic field point source.

The equations for the fields gener-ated by an electric point source arebelow:

These are the fields generated by theinfinitesimal dipole current elementIo The terms in parentheses are forthe fields generated by an infinitesi-mal current loop. The term Im is animaginary magnetic “current" ele-ment, perpendicular to the currentloop, whose magnitude is propor-tional to the magnitude of the cur-rent times the area of the loop.

Looking at the equations, we nowfind that the field intensity varies assomething other than just l/r—thereare field components which vary as1/r2 and 1/r2. We find field compo-nents in all three orthoginal direc-tions: r, q, and f.

These equations hold in general,both near field and far field. In thefar field, the 1/r2 and 1/r3 termsbecome negligible. The radial fieldcomponent therefore also becomesnegligible. Thus, m the far field, thefields vary only as l/r, and the ratioof |E|/|H| is 377 ohms.

The q and f terms are perpendicularto the direction of wave travel. Thedirection of wave travel is alongradial lines. At large enough radius,the curvature in the q and fdirec-tions is negligible with respect to awavelength, and so we have planewaves.

Figure 3. Electrical field source and currentsystem

Figure 4. Magnetic current element

7

Dipoles are electric field antennas.Loops are magnetic field antennas.In the far field, where the E fieldand H field are in constant ratio toeach other (377 ohms), the relativesensitivity ofdipoles versus loops isalways the same. In the near field,because of the additional l/r2 andl/r3 terms, this is no longer true. Letus look at the ratio of |E| and |H|in the near field so see what the dif-ference in sensitivity betweendipoles and loopswould be.

The ratio of |E| and |H| is calledthe wave impedance. This illustra-tion shows how the wave impedancevaries as a function of distance andnature of the source, with the radialcomponent neglected. Note that thedistance is normalized to the wave-length divided by 2π.

Several other factors complicatenear field measurements. As wehave just seen, there is no fixed ratiobetween the E and H fields.

We may measure different fieldstrengths depending on the type ofantenna we use. In general, the wavefronts are not plane waves in nearfield, but are curved. The antennacalibrations which are done in farfield, where the wave fronts areplane waves, therefore will not hold.Very small dipoles can be used tomake accurate near Field measure-ments. They will minimize the fieldcurvature effect. But they sacrifice alot of sensitivity due to large mis-match or loss in the matching net-work. In practical leakage measure-ments, wehave additional complicat-ing factors: The nature of the source,whether electric or magnetic, may beunluiown. And most likely there willbe multiple sources for the leakage.

Figure 5. Wave impedalice vs. distance from source

Figure 6. Sources complete measurement

8

Historically, we have used a two turnloop antenna, to measure rf leakagefrom our signal generators. Due toradios becoming more sensitive andoperating at higher frequencies, ithas become inadequate for rf leak-age measurements. Its sensitivity isno match for a resonant dipoleabout 20 to 30 dB less sensitive infar field measurements. It is hard toassign a number for comparativesensitivity due to its unflatness andlack of consistency among units. Atfrequencies higher than several hun-dred MHz, the handle has beenfound to be a better antenna thanthe loop portion, This make it sensi-tive to small changes in orientationand hand position. Even reflectionsoff the user’s body change theresponse.

To improve our sensisivity and flat-neas for measuring our low leakagesignal generators, we have decidedto use resonant dipoles. The dipoleshave spacers, 25 mm radius, to set astandard measuring distance.Several tests have shown this to be agood choice as weshall see.

Figure 7. Two turn loop far field responses vs. ideal dipole

Figure 8. Photo of dipole

9

Correlating receiversand dipole antennas

Our goal is to build signal generatorswith leakage low enough such that itcaimot be detected by receiversplaced a given distance from the sig-nal generators. The problem is thento generate a field that a receivercan barely detect, and then substi-tute a resonant dipole in this samefield. The signal level out of thedipole is then measured.

This is first done in the far fieldwhere all the “nice” properties hold.The test is repeated in the near field.The signal levels from the dipole infar field and near field are thencompared. If they are equal, we haveestablished equivalency betweendipole field strength measurementsand radio sensitivities.

Figure 9. Calibration test between dipole and receiver.

Figure 10. Leakages simulation of cabinet with seam

10

First, the far field tests were car-ried out. In a semianechoic cham-ber (all surfaces covered with RFabsorber except the floor) a trans-mit antenna was placed on an ele-vator mechanism. Various radiosand pagers, 146 MHz to 1260 MHzwere placed about two metersfromthe transmit antenna, about 1 meterabove the floor. The height of thetransmit antenna was adjusted formaximum received signal in thereceiver. The signal strength wasadjusted at the source so that themodulation tone was just perceptiblefrom the re- ceiver. A dipole wasthen substituted at the location ofthe receiver, and the received signalmeasured.

Next, a circuit fed slot antenna wasconstructed. A receiver was placed25 mm from the slot antenna. Thereceiver was rotated for maximumsignal strength, and the signal strengthwas again adjusted so the modulationtone was just perceptible from thereceiver. A dipole with 25 mm radiusspacers was then scanned over theslot antenna. The received signalstrength was then compared to thatpreviously found in the earlier test,and was with 2 dB.

The tests so far have compared therelative sensitivity of dipoles versusreceivers. While this is sufficient foraccurate testing of RF leakage, wewanted to determine the absolutesensitivity of the dipole antennas.

For the far field, this is given byequation 14:

(14)

How well would the dipole do inabsolute terms in measuring theradiation from a slot antenna in thenear field? A test was done at 500 MHzto find out. A 600 MHz resonant cir-cuit fed slot antenna was drivenwith 0 dBm RF power, A resonantdipole was placed across the slot,25 mm away. The received powerwas measured. The power was -3 dBm,indicating a rather efficient couplingof power. We believe that this is dueto capacitive coupling of the poten-tial difference across the slot to theantenna poles. This unexpectedresult shows that a dipole, an elec-tric field antenna, is good even formeasuring radiation from magneticsources like a slot antenna.

This success inspired a similarexperiment. A 500 MHz resonantloop was constructed. This too is amagnetic radiator. The loop wasdriven with 0 dBm RF power. Whena dipole was placed 26 mm from theloop, the power received was -6 dBm.In this case, the coupling was notquite as good, but still quite accept-able to indicate when leakage shouldoccur.

Thus, experimentally, the dipole hasshown itself to be sensitive, even tomagnetic sources which might befound in signal generators.

Figure 11, Absolute measurementloss – loop antenna to dipole antenna

11

Sources of leakage from asignal generator

If the cabinet containing a signalgenerator were a welded metal box,for all practical purposes therewould be no leakage. Of course, thisis impractical, for we must be able toget signals out of the box, and modu-lation, control, and power into it. Inaddition, in order to service it, thecovers must be removable, and notwelded. These considerations causeus to compromise the ultimateshielded box.

Wires and cables passmg through ashielding enclosure are commonlycalled penetrations. If an insulatedwire is simply passed from theinside to the outside of a shieldedbox through a hole, it will conductout the rf fields present in the boxand then radiate them on the out-side. Even a terminated coax cable,if the outer conductor is notgrounded at the hole, can conductout the fields.

For low leakage along seams, theideal is a continuous low impedancecontact. In practice, this is difficultto achieve. The goal can be approachedwith some reduction in performanceby using closely spaced periodic con-tact points of the sort provided bymultiple screws or contact fingers.

In the critical front seam area of ourlow leakage signal generators, theperiodic contact between the instru-ment cover and a front bulk-head issupplied by screws spaced 26 mmapart. This is about .08 wavelengthapart at 1 GHz.

The rear area was deemed less criti-cal forleakage, as it is further awayfrom the area where radio testingwould be done. Here, a spiral strip isused to make contact between theinstrument cover and rear panelassembly.

The cover is a one piece sleeve, witha spotwelded seam. The air vent holearea in the cover is backed by alu-minum mesh riveted to the cover.The mesh lowers rf leakage throughthe vent holes. As mentioned earlier,behind the front panel, there is abulkhead which supplies the shield-ing in front. The instrument covermakes contact with this bulkhead,forming a shielded box. Through thisbulkhead pass the rf output cableand several modulation input cables.Also passing through the bulkhead isa digital control cable for the frontpanel keyboard and display.

Figure 12. Perfect cabinet

12

The solution for coax cables then isto ground the outer conductor at thebulkhead wall. This is done by usinga bulkhead connector in the case ofthe modulation inputs. The rf outputcable is soldered to a groundingplate which is then bolted to thebulkhead. The digital control cablefor the front panel keyboard and dis-play is brought through the bulkheadwith a shielded connector having aninternal low pass filter for each line.

Figure 15. Detail of cabinet optionion 10 Figure 16. Rf leakage wire passage option 10

Figure 14. Wire causes cabinet leakage

Figure 13. Simulation equivalents to signal generator cabinet

13

Dipole discussion

What are some of the sources ofinaccuracy with dipole measure-ments?

1. If the dipole impedance is not matched to the measurement sys—tem impedance, there will be loss due to mismatch. The magnitude of this loss is:

Mismatch loss in dB = -10 log (1-r2)

where:

r = reflection coefficient magnitudelog = base 10 logarithm

As an example, 73 ohms is the impedance of a dipole in free space, slightly shortened from onehalf wave length. If this is connectedto a 50 system, there will be 0.15 dBmismatch loss.

2. A more general form of mismatch loss is mismatch uncertainty, This problem is often seen m systems with cables, where the angle of thereflection coefficient will rotate with frequency. The limits ofthia uncertainty are:

Mismatch uncertainty = 20 log (lip±)r1r2

r1 = reflection coefficient magnitude of source (antenna)

r2 = reflection coefficient magnitude of load (receiver)

Analysis shows that resistive loss in a half wave dipole is negligible. Let the pole elements be made of copper. Assume all the current flows uniformly on the surface of the pole to a thickness of one skindepth. The surface resistivity is:

Surface resistivity =

where s = conductivityd = skin depth

Skin depth is as follows:

Where f = frequencyπ = permeability

Thus;

Surface resistivity =

For copper, µ = 4πl0-7 henry/meter (permeability of free space),

s= 5.8e7 siemens at room temperature.

Surface resistivity =

Assume the pole diameter is 3.58mm (.141 inch). The circumferenceis 11.25 mm. For poles of circumfer-ence 11.25 mm:

Resistance per mm =

3.9x10-3 = .35

14

15

Agilent Technologies Test and MeasurementSupport, Services, and AssistanceAgilent Technologies aims to maximize the

value you receive, while minimizing your risk and

problems. We strive to ensure that you get the

test and measurement capabilities you paid for

and obtain the support you need. Our extensive

support resources and services can help you

choose the right Agilent products for your appli-

cations and apply them successfully. Every instru-

ment and system we sell has a global warranty.

Support is available for at least five years beyond

the production life of the product. Two concepts

underlie Agilent’s overall support policy: “Our

Promise” and “Your Advantage.”

Our PromiseOur Promise means your Agilent test and mea-surement equipment will meet its advertisedperformance and functionality. When you arechoosing new equipment, we will help you withproduct information, including realistic perfor-mance specifications and practical recommenda-tions from experienced test engineers. Whenyou use Agilent equipment, we can verify that itworks properly, help with product operation, andprovide basic measurement assistance for theuse of specified capabilities, at no extra costupon request. Many self-help tools are available.

Your AdvantageYour Advantage means that Agilent offers a widerange of additional expert test and measurementservices, which you can purchase according toyour unique technical and business needs.Solve problems efficiently and gain a competitiveedge by contracting with us for calibration,extra-cost upgrades, out-of-warrantyrepairs, and on-site education and training, aswell as design, system integration, projectmanagement, and other professional engineer-ing services. Experienced Agilent engineers andtechnicians worldwide can help you maximizeyour productivity, optimize the return on invest-ment of your Agilent instruments and systems,and obtain dependable measurement accuracyfor the life of those products.

By internet, phone, or fax, get assistance withall your test & measurement needs

Online assistance:www.agilent.com/find/assist

Phone or FaxUnited States:(tel) 1 800 452 4844

Canada:(tel) 1 877 894 4414(fax) (905) 282 6495

China:(tel) 800 810 0189(fax) 1 0800 650 0121

Europe:(tel) (31 20) 547 2323(fax) (31 20) 547 2390

Japan:(tel) (81) 426 56 7832(fax) (81) 426 56 7840

Korea:(tel) (82 2) 2004 5004 (fax) (82 2) 2004 5115

Latin America:(tel) (305) 269 7500(fax) (305) 269 7599

Taiwan:(tel) 080 004 7866 (fax) (886 2) 2545 6723

Other Asia Pacific Countries:(tel) (65) 375 8100 (fax) (65) 836 0252Email: tm_asia@agilent.com

Product specifications and descriptions in this document subject to change without notice.

Copyright © 2002 Agilent Technologies

Printed in USA , February 5, 2002

5952-2789