Testing Ourselves Levent Sevgl DoOu, Univesiy Eletronis and Communication Eng. Dept. Zeamet Sokak, No 21, Acibadem - KadikOy Istanbul, Turkey Email: lsevgi@dogus.edu.tr. levent.sevgi@ieee.og i htp:/ww3.dogus.edu.trllsevgi O pen-area test site (OATS) calibration is a undamental antenna-engineering task. This issue' s tutorial is on the cali­ bration of an open-area test site. The requirements, the theoretical basis, related standards, and the procedure to be followed are dis­ cussed in this tutorial. The next tutorial, scheduled for the August 2010, issue is on the design of a digital-to-analog converter (DAC); its quiz was in the last issue. interested in this topic are referred to [1-3] (and the references therein) for the initial information and potential content of the tuto­ rial. The problem of earthquake prediction (EP) is an important issue for electromagnetics and society, whether or not individuals are interested in it. Electromagnetic (EM) people are right at the center of these discussions, because the majority of these earth­ quake-prediction claims are based on EM precursors. The part of the community that deals with radars and sensing systems knows very well know that EM-precursor-based earthquake prediction is nothing but a multi-sensor surveillance problem. As mentioned in the last issue, there will be a tutorial on this topic, with a prospec­ tive title of "Prediction: Extrapolation to the Future," which is tentatively scheduled for the October 2010 issue. Readers who are I. "Nature Debate on Earthquake Prediction," http://ww.nature.com/nature/debates/earthguake/eguake framese t.html, February 1999. References 2. Science, http://www.sciencemag.org. 3. L. Sevgi, "Earthquake Early Waning: Prediction vs. Guess," Comment to the "Earthquake Alarm," by Tom Bleier and Friedemann Freund, IEEE Spectrum Online, 42, 12, December 2005, pp. 17-21 (http://www.spectum.ieee.org.csulib.ctstateu.edul dec05/commentsll181). Open-Area Test Site (OATS) Calibration 1 s. Eser and L. Sevgi 2 lESiM EMC Test and Measurement Center TOSa Organized Industrial Region, $ekerpmar - Gebze Kocaeli, Turkey E-mail: sedateser@esim.com.tr 2 Dogu} University, Electronics and Communications Engineering Department Zeamet Sokak 21, AClbadem - Kadlkoy, 34722 Istanbul, Turkey E-mail: Isevgi@dogus.edu.tr Abstract Open-area test site (OATS) calibration is discussed in this tutorial. The calibration is peformed according to the CISPR and ANSI/IEEE standards. First, theoretical calculations of the normalized-site-attenuation (NSA) procedure are presented. Then, normalized-site-attenuation measurements of a newly established open-area test site are given. Finally, calibration is done through comparisons. Keywords: Antennas; Calibration, Open-Area Test Site (OATS), Normalized Site Attenuation (NSA), Antenna Factor (AF), EMC measurement, Emission Measurements. 204 IEEE Antennas and Popagain Mgazine, Vol. 52, No.3, June 2010 1. Introduction O pen-area test sites (OATS) are used in radiated electromag­ netic interference (EMI) measurements, while antenna cali­ bration through antenna-factor (AF) measurements is performed in a calibration test site (CALTS). Both sufer from environmental conditions. Therefore, site calibration is essential before conduct­ ing any measurement and/or test. Note that both open-area test sites and calibration test sites can also be used in immunity tests if calibrated together with ground absorbers. The calibration proce­ dures are presented in CISPR [I, 2], FCC [3 ], and ANSI/IEEE [4, 5] standards. The calibration of an open-area test site can be done according to the procedure given in [I, 3, 4]. On the other hand, calibration test sites are also open-area test sites that are used for determining the free-space antenna factor, and can be cali­ brated/validated according to [2]. Open-area test site calibration is performed through site measurements and comparisons against a reference site. The ideal or reference site is nothing but a perfectly conducting large plane over level ground. In [I], the reference or standard site was deined as "A site comprised of a lat, open-area, devoid of nearby scatter­ ers such as trees, power lines, and fences that has a large metallic ground plane." There are also many papers in the literature on open-area test site calibration (see, e.g., [6-8]). A typical picture of an open-area test site is sketched in Fig­ ure I. The site is an elliptical, open, leveled area, with major and minor axes of 2R and R j , respectively. The distance between the antenna and the equipment under test (EUT) is R. The meas­ urement distance, R, may be 3 m, 10 m, or 30 m. tive or capacItIve energy storage. RA is the combination of RL and Rro which represent the power (thermal) losses of the antenna and EM radiation, respectively. Ideally, RL and XA are zero, and the antenna' s resistance is equal to the radiation resistance (RA = Rr). hen the receiving antenna' s circuit model contains a voltage source (with induced open-circuit voltage Voc)' there is the same antenna impedance, ZA , and the receiver input impedance, Zr. In the ideal (matched) case, all impedances are real and 50 n, and Vr is half of the induced voltage, Voc. The ree-space far-ield electric-ield strength at a distance d rom a transmitting antenna having transmitted (radiated) power � and antenna gain Gt is [9] 2 .G e -jd £ = t I 2 120 r 4 rd _ � £= �30'G t t jkd e__. d (I) The radiated power is determined by the antenna' s current and the 2 radiation resistance, and reduces to � = ] RA for the lossless antenna. Equation (I) may therefore also be written as £=] �30RAGt jd e__. d (2) The received power of a matched (50 n) lossless antenna is the product of the power density, Pd, of the incident ield and the The standard site calibration is based on the measurement of the insertion loss between transmitting and receiving antennas on a large, lat, and unobstructed conducting ground plane. As speciied in clause 5.8.2 of [I], site performance may be validated by the normalized-site-attenuation (NSA) method. One antenna is set to be at a ixed height (e.g., I m or 2 m), while the other antenna is scanned rom I to 4 m in height. The maximum response between the two antennas is recorded. The conducting ground plane is there to ensure the repeatability of the calibration measurements. This tutorial is organized as follows. First, the antenna is dis­ cussed as a transducer in Section 2, and derivations of the receiv­ ing and transmitting antenna factors are given. The ield strength at a distance d over perfectly conducting lat ground is also given in this section. The deinition of the normalized site attenuation is presented in Section 3. The theoretical normalized-site-attenuation calculations mentioned in the standards are included in Section 4, and the measurement procedure is reviewed in Section 5. Section 6 is reserved for the open-area test-site calibration. Finally, the con­ clusions are listed in Section 7. Figure l. A sketch of a typical open-area test site with a turn­ table. 2. An Antenna as a Circuit Element An antenna is a device that can be modeled/examined using wave and circuit theories [9]. Figure 2 pictures one of the simple models of a transmitting/receiving antenna pair. The basic circuit representation of a transmitting antenna consists of a complex impedance, ZA , connected across a voltage source with an intenal impedance Zs. The input impedance, ZA , of the antenna consists of a real part, RA, and an imaginary part, XA' representing inducIEEE Antennas and Popagation Magazine, Vol. 52, No.3, June 2010 Figure 2. Circuit models of transmitting and receiving anten­ nas. 205 maximum effective aperture, Ae, where A is the wavelength of the received signal: P= r PA d e 2 ,2 �� = 120r 4r (3) = The antenna voltage, VA' across the load is (4) The open-circuit voltage, Voc' of the antenna is twice this voltage, and can be given as the product of the effective antenna height and the incident electric ield, so that (5) For a 5 0 n system, both the receiving and transmitting antenna factors - the antenna factor (AF) and the transmitting antenna fac­ tor (TAF), respectively - reduce to (6) (7) The ield strength may therefore be given in terms of the transmit­ ting antenna factor as E= r VAJMHz e-fd (8) d 79.58AFr The ield strength at a distance d over a perfectly conducting ground plane is the sum of the direct, ground-relected, and surface waves. Since the surface waves are negligible above 30 MHz, the ield strength is E= where Vf JMHz 79.58AFr I pl efJ [ e-fdl _ dl I pl efJe-fd2 + ..___ ! _ d2 ] ' The following instruments and components are required for normalized-site-attenuation measurements: A broadband signal generator and a spectrum analyzer, Two 10 dB padding attenuators, ampliiers, or pre­ ampliiers, A mast capable of scanning the receiving antenna rom I m to 4 m in height, and a ixed mast for the transmit­ ting antenna. There are two widely accepted normalized-site-attenuation calculation approaches. The irst one, which considers the ground plane to be ininitely conducting and is proposed by ANSI/IEEE [4], is based only on the far-ield term. The second one also takes near-ield contributions into account, assuming that far-ield radia­ tion in the 30-200 MHz range is inappropriate when the distance is equal to 3 m or 10 m, as the test procedure in an open-area test site imposes. The disadvantages of the ANSI/IEEE models are the elimination of the near-ield terms of the electromagnetic ield, as well as the elimination of the Norton surface-wave component, which can cause discrepancies up to 1.5 dB, especially for vertical polarization [I]. 4. Theoretical NSA Calculations The coniguration for theoretical open-area test site normal­ ized-site-attenuation calculations is pictured in Figure 3. Here, (10) (9) is the relection coeficient of the ground, and dl and d2 correspond to the path lengths of the direct and ground­ relected waves, respectively. 3. Normalized Site Attenuation (NSA) The normalized site attenuation (NSA) has become the stan­ dard parameter for determining the adequacy of an open-area test site for performing electromagnetic emission and immunity meas­ urements. The normalized site attenuation is deined as the ratio of the power input of a matched, balanced, lossless, tuned dipole radiator to that at the output of a similarly matched, balanced, lossless, tuned dipole receiving antenna, for speciied polarization, separation, and heights above a lat electromagnetically relecting surface [I]. It is a measure of the transmission path loss between two antennas. 206 The method proposed by ANSI/IEEE [4] for the evaluation of normalized site attenuation is obtained with a Geometrical­ Optics (GO) approximation, based on the Friis equation, since Norton suface waves are negligible for these antenna heights and requency bands. Standards deine the normalized site attenuation as the input power available to a short dipole for a ield strength of 100 �Vim at a distance of d 3 m over a leveled conducting ground screen [1]. (\ I) are the direct and ground-relected path lengths, r=tan -I (ht +hr d J (12) x Figure 3. A sketch for the theoretical normalized site- attenua­ tion calculations. IEEE Antennas and Popagation Mgazine, Vol. 52, No.3, June 2010 is the grazing angle, and % ----------------------------------------------------------------------% Program : LS_NSAm (Computes theoretical NSA) % ----------------------------------------------------------------------- vP = ( &r - j60T.) sinr- I( &r - j60T.) -COS2 r " ( &r - j60T.) sinr+ �(cr - j60T.) -COS2 r (14) are complex relection coeicients for the horizontal and vertical polarizations, respectively. Here, &r and T [S/m] are the relative penittivity and conductivity of the ground, respectively. Electric ield strengths for both polarizations are then calculated as [1] where J = 2! .' (Phv), ] 'hv, _,[lm() Phv, . - tan Re (17) clear al; clc; ht = I; d = 30; er = 15; sigma = 0.01; k=O; pol=input(' [I] V Pol (TM), [2] H Pol (TE)'); for req=3e7:le7:le9 k=k+I; kO=2*pi*freq/3eS; lambda=3eS/req; Edmax(k)=1e-6; hnax(k)=I; epO=S.S54e-12; n2=erj*60*lambda*sigma; for hr=1:.02:4 dl = sqt(dA2+(ht-hr)"2); d2 = sqrt(dA2+(ht+hr)"2); theta = atan(ht+hr)/d; sq=sqt(n2-cos(theta)"2»; fpol == I rho = (n2*sin(theta)-sq l(n2*sin(theta)+sq); rhom=abs(rho); rhoa=angle(rho); rh(k)=rho; Ed=sqrt(49.2)*dA2*(d2A6+dlA6*rhomA2+2*dIA3*d2A3* rhom*cos(rhoa-kO*(d2-dI»)"(0.5)/(d1A3*d2A3); elsefpol == 2 rho=(sin(theta)-sq)/(sin(theta)+sq); rhom=abs(rho); rhoa=angle(rho); rh(k)=rho; Ed=sqt(49.2)*(d2A2+dIA2*rhomA2+2*d1*d2* rhom*cos(rhoa-kO*(d2-dI »)"(0.5)/(dl*d2); end fEdmax(k) < Ed; Edmax(k)=Ed; Hrmax(k)=hr; end end f(k) = req/le6; nsa(k) = 279.1/(f(k)*Edmax(k»; end igure(I); plot(f,20*log1 O(nsa),'LineWidth',2); grid on; xlabe/('Frequency [MHz]'); ylabel('NSA [dB]'); xim[30,I 000); yim([-20,10); % End Figure 4. A MATLAB code for the theoretical normalized site­ attenuation calculations (ht (�) is the transmitter height, hr ( hr ) is the receiver height, d is the range, er (cr ) is the relative permittivity, and sigma ( T) is the conductivity). . Finally, the nonalized site attenuation is calculated rom 279.1 NSA = ___ _ _ (MHz)E Dmx(lm) __ (IS) Here, EDmx is the maximum electric ield strength (in ..lV/m) received by the receiving antenna, height scanned between I m and 4 m, rom a theoretical half-wave dipole with 1 pW radiated power and 1.64 dB antenna gain, calculated from Equations (15) and (16) for horizontal and vertical polarizations, respectively. A short ALAB-based nonalized-site-attenuation numeri­ cal calculator, prepared for [10], is given in Figure 4. This may be . used to prepare theoretical nonalized site attenuation charts showing requency, nonalized site attenuation, E Dmx' and . -5 � . . � ....... i , -. . . -10 m � i _. .......... . : , IEEE Antennas and Popagation Magazine, Vol. 52, No.3, June 2010 h, = 1m, d = 3m V-l ,/ T=O.OISlm,s, =15 .: / =le7 Slm,s, =15 -15 -20 j , Z hmx data in four columns. In these charts, hmx is the maximum receiving antenna height where E Dmx is recorded. This calculator can be used for any penittivity and conductivity of the leveled ground. Figure 5 shows an example produced with this ALAB­ based nonalized-site-attenuation calculator. Nonalized site attenuation values as a unction of requency for both vertical and horizontal polarizations above perfectly conducting and lossy ground are shown. Figure 6 shows the ront panel of the standalone open-area test site nonalized-site-attenuation calculator devel­ oped in [10]. For a free-space environment and ininitely small , = 1m,d = 3m H-ol T=0.0ISlm,6, =15 = Ie7 S / sr 15 m, = 800 800 -20 100 200 00 40 500 600 Fqueny [Mz] 700 1000 Figure 5. Plots of the results from the MATLAB-based nor­ malized site attenuation calculator: the solid line is for lossy ground, the dashed line is for PEe ground. 207 . ---_ .. O_IIJ --o -- .. __ . --- -.. ...I - .. - .. ���. . ,. , - . �/Edmax . �HmaX --o --o -; �..- .. ___ .. ..-.. -- .. -- ...-� -- I 1 j 0_ . -_ . y- ,,- t_ .. -;- - . - - - - - -- - - ._ - --- ---_> -- .. -- .... -- ..... - - .. --.. -- .. ----- -- .. --o -.. 0_ .. 0_,., --o --o 0." .. 1 -j -- -_ . o; -;; .. Figure 6. The front panel of the normalized site attenuation calculator. Table I lists free-space normalized site attenuation values at some requencies computed rom Equations (18), (19), and (20) in free space. The ree-space normalized site attenuation rom Equa­ tion (18) may be obtained directly by setting the relative permit­ tivity and conductivity values to one and zero, respectively. As observed, using Equation (20) instead of Equation (19) introduces signiicant errors that are not acceptable [ I] for shot-range meas­ urements. For d = 3 m, this error is I dB and maximum at 30 MHz. This error is less than 0.1 dB above 60 MHz at d = 5 m, and above 110 MHz at d = 1 0 m. Therefore, the free-space normalized site atenuation equation with only the 1/ d term can be used for EMC and antenna-factor measurements in the requency range of 30 MHz to I GHz at d = 10 m and d = 30 m, but a correc­ tion factor (CF) should be added at shorter ranges for the lower end of this requency band (i.e., for f < 110 MHz at 3 m and f < 60 MHz at 5 m) [1-4]. 5. NSA Measurements Table 1. The normalized site attenuation as a function of frequency in free space computed from Equation (18) (NSAI), Equation (19) (NSA2), and Equation (20) (NSA3) (V-pol, d = 3m, � = 2m). Frequency )MHz) NSAI NSA2 NSA3 30 12.00 12.98 12.00 9.50 IdBI IdBI IdBI 40 9.50 10.12 50 7.56 7.97 7.56 80 3.48 3.64 3.48 100 1.54 1.65 1.54 120 l.05 0.03 l.04 140 -1.38 -1.33 -1.38 160 -2.54 -2.50 -2.54 180 -3.57 -3.53 -3.56 200 4.48 4.46 4.48 Normalized site attenuation measurements can be performed in accordance with the CISPRI6-1-4 [I] or ANSI/IEEE C63.4 [3] standards. Two antennas are set up on the test site in an appropri­ ate geomety mentioned in the standards. The single-point normalized site attenuation measurement is valid for an open-area test site. To be sure, normalized site atenuation measurements of an anechoic chamber might be pre­ fered. These require a maximum of 20 separate measurements (inside a test volume deined in [ID, i.e., ive positions in the hori­ zontal plane (center, let, right, ront, and rear, measured with respect to the center and a line drawn rom the center to the posi­ tion of the measuring antenna), for two polarizations (horizontal and vertical), and for two heights (I m and 2 m, horizontal, I m and 1.5 m, vetical). The normalized site attenuation procedure requires two diferent measurements of the voltage, Vr, at the receiver. The irst Vr reading belongs to the case where the two antennas are antennas (and 20 system), normalized site attenuation can be calculated rom [I] NSA - -[ ::l [ � J ( H d == I= + I_ = -1 antennas are connected to their coaxial cable, and the maximum signal is scanned in height. For both of these measurements, the height of the signal source is kept constant. The irst reading of Vr (19) (Jd)2 (Jdt in 1/ d2 and 1/ d4. In the far-ield, this expression can be reduced to _ ( ) [ (Mz) ]. 52 0 2. d (20) For 20 = 50 n and in decibels, Equation (II) can be written as NSA = 32+2010g(dm)-2010g(JMHz). 208 is called VDirect' and the second is VSite. Measured normalized site attenuation is then extracted rom NSA = VDirect - VSite - A, - AFr - A,ot, The ree-space normalized site attenuation equality contains teris NSA = removed, and coaxial cables are directly connected to each other via an adapter. The second Vr reading belongs to the case where ( 21) (22) where A, and AFr are the antenna factors of the transmitting and receiving antennas (in dBm-J), and A,ot is the mutual imped­ ance correction factor (in dB). Note that VDirect includes cable losses of both the transmitting and receiving antennas, and the diference VDirect - VSite is equal to the classic site attenuation. These measurements are carried out with broadband anten­ nas. The distance d is measured from the center of the transmit­ ting antenna to the center of the receiving antenna. This distance should be maintained for all measurements, which requires that the receiving antenna be moved along the line in the directions shown IEEE Antennas and Popagation Mgazine, Vol. 52. No.3, June 2010 in the igure. Also, the transmitting and receiving antennas should be aligned with the antennas' elements parallel to each other and orthogonal to the measurement axis. Furthermore, the lower tip of the antenna should be at a distance greater than 25 cm rom the loor, which may require the center of the antenna to be slightly higher than I m for the lowest height measurement. Both CISPR and ANSI/IEEE normalized site attenuation necessitate that normalized site attenuation of an open-area test site must be within ±4 dB of the theoretical values [II] (±I dB for antenna uncertainties, ±I dB for antenna-factor uncertainties, ±I dB receiver uncertainties, and ±I dB for site uncertainties). There are curently no open-area test site validation requirements above I GHz. However, facilities suitable for measurements in the requency range of 30 MHz to 1000 MHz are considered suitable for the frequency range I GHz to 40 GHz, including the presence of the reference (metal) ground plane. 3. Record the maximum signal level. This value is VSite in Equation (22). 4. Disconnect the transmitting and receiving cables rom their antennas. Directly connect these cables with a straight-through adapter. 5. Record the signal level with the transmitting and receiving cables connected. This value is VDirect in Equation (22). 6. OATS Calibration A new open-area test site was constructed according to [12]. The photos of the open-area test site to be calibrated are given in Figure 7. As shown there, there were nearby buildings, trees, and a hill with a long I m high sidewall. A sketch showing the location, nearby obstacles, and their distances is shown in Figure 8. Two broadband antennas can be used in normalized site attenuation measurements. The transmitting antenna has its refer­ ence point at the measurement positions of the test volume, and the receiving antenna is outside this test volume at a prescribed orien­ tation and position. The transmitting antenna should have an approximately omnidirectional H-plane patten. Typical receiving antennas are hybrid antennas (bi-conical/log-periodic dipole com­ bination) for 30 to 1000 MHz, or separate biconical antennas (for 30 MHz to 200 MHz) and log-periodic dipole antennas (for 200 to 1000 MHz). The equipment used during the normalized site attenuation measurements is listed in Table 2. Due to their large dimensions and phase-center problems, hybrid antennas are not recommended for normalized site attenuation measurements. Figure 7. Photographs of the open-area test site to be cali­ brated. As described in [I], the normalized site attenuation method is used for the calibration of an open-area test site having an inter­ antenna distance greater than 5 m. The ree-space antenna factors of the antennas used in the calibration are needed in this method. For each requency site, calibration is performed with the follow­ ing steps: I. 2. Adjust the output level of the signal generator to give a received voltage display well above the ambient and measured receiver or spectum-analyzer noise. oer ild Fot -( Floes 1, ... Control rom School lab KOSGEB building Raise the receiving antenna on the mast through a scan of 1-4 m. Figure 8. A sketch of the open-area test site to be calibrated. Table 2. A list of the equipment used in the calibration. No 1 Equipment Biconical ntenna 2 Biconical ntenna 3 Log-Periodic ntenna Producer Model Spec Schwarzbeck VHA 9103 30-300 MHz EMCO 3109 30-300 MHz Schwarzbeck VUSPL 9111 200-2000 MHz Schauer 30-2000 MHz HP 8560E 30 Hz-2900 MHz RG213U 18m,4m,IOm 4 BiLog antenna Spectm analyzer 6 Coxial cable 7 Sinal generator - CBL 6141A 5 HP 8657B 0.1-2000 MHz 8 Signal generator Anritsu MG3633A O.oJ-2700 MHz IEEE Antennas and Popagation Magazine, Vol. 52, No.3, June 2010 209 6. At each requency and for each polarization, enter the values in Steps 3 and 5 into Equation (22). 7. Insert the transmitting and receiving antenna factors at the measurement frequency as shown in Equation (22). 8. Insert the mutual-impedance correction factor M;ot, which applies only for the speciic geometry of hori­ zontal polarization using tunable dipoles separated by 3 m. M;ot 0 for all other geometries. = 9. Solve Equation (22) for AN' which is the normalized site attenuation for the measurement frequency and polarization used. 10. Subtract the value in Step 9 rom the appropriate nor­ malized site attenuation value. I I. If the results in Step 10 are less then ±4 dB, the site is considered validated at that requency and polarization. 12. Repeat these steps for the next frequency and polariza­ tion combination. The measurements are performed as follows: I. The spectrum analyzer is located in the control room. 2. A signal generator is placed at the urthest point behind the transmitting antenna. The signal generator and spectrum analyzer are directly connected via coaxial cables and suitable connectors. 3. The signal generator is set to 120 dBlV and the re­ quency is set to 30 MHz. 4. A coaxial cable is connected to the F input of the spectrum analyzer with the center requency set to 30 MHz, with 200 kHz SPAN, RB 10 kHz, VB 100 Hz, and sweep time of 200 ms. The reference level is set to a suitable level according to the signal strength to be measured. = = 5. 7. 8. 210 The signal generator is set to 120 dBlV and the re­ quency is set to 30 MHz. MAX HOLD is selected at the spectrum analyzer. 10. The receiving antenna is moved vertically between I m and 4 m by scanning rom the control room. I I. MARK PEAK is used to record the maximum ield value. 12. The procedure is repeated for other requencies. Measurements are repeated for 3 m and 10 m for both polarizations. A typical measurement chart is given in Table 3. The results of the open-area test site calibration are given in Figures 912. These igures belong to 3 m and 10 m calibration measure­ ments for both horizontal and vertical polarizations. The dashed lines belong to theoretical normalized site attenuation calculations, � ---�. l d=10m, Hor Pol t4 d81imits i' 10 � � Z 5 o 5 .. · 10 ·15 ....... .... . ... ..... ....... . . .... . . .. . .. .� -������_���=-����� 1� � � D � � � � � � Frqueny [MHz) Figure 9. The normalized site attenuation as a function of fre­ quency of the open-area test site, and uncertainty limits dic­ tated in the standards (H-Pol, d 10 m). = t4 d81imits The value at the spectrum analyzer is recorded as VDirect at 30 MHz. This procedure is repeated for every measurement frequency, without changing the level at the signal generator. 6. 9. The coaxial cables are then disconnected. The signal generator and its cable is connected to the transmitting antenna through a 10 dB attenuator; the spectrum ana­ lyzer' s cable is connected to the receiving antenna through another 10 dB attenuator, via an adaptor. The transmitting antenna is located on the mast and ixed at I m height above the ground plane. The coaxial cable is ixed horizontally, and extended by at least 2 m behind the antenna before dropping to the ground and connecting to the signal generator. The receiving antenna is located 3 m or 10 m away from the transmitting antenna. 5 i' � � 0 n z ·5 ·10 . . .. .. .. ... .. . .. . ". . .... ·15 . . .. .... ... . . . I ••• • • • . .. ... ·�-�l 0 ����-D--��-��- � ������ �����l� Frqueny [MHz) Figure 10. The normalized site attenuation as a function of fre­ quency of the open-area test site, and uncertainty limits dic­ tated in the standards (V-Pol, d 10 m). = IEEE Antennas and Popagation Magazine, Vol. 52, No.3, June 2010 D � « D � .o � z � -10 -15 ' -10 -15 . ...... .. ..... ... . . ... .. .. . . ·5 l . ·0 .. .. .... -5 . · . t. -o:-O0=)o. Fqueny [MHz] = l ��-::�-:�-�:�:�:�-�-�� : 10 00 l 00 00 00 00 m )) 100 Figure 12. The normalized site attenuation as a function of fre­ quency of the open-area test site, and uncertainty limits dic­ tated in the standards (V-Pol, d 3 m). = Table 3. A normalized site-attenuation measurement chart (d [MHz] VDirect [dBlV] VSite [dBlV] Measured AFTx Site VHA9103 Attenuation CBL6141A [dB) ........... . . ... . Frqueny [MHz] Figure 11. The normalized site attenuation as a function of fre­ quency of the open-area test site, and uncertainty limits dic­ tated in the standards (H-Pol, d 3 m). Freq ±4 dB limits d=10m, Ver Pol ±4dB limits d=3m, Hor Pol 15 [dBm-l) AFx = 10m, Hor-Pol). EMC03109 AN AN VUSLP Measured 9111 [dBm-2 ] Calculated [dBm-l] [dBm-2 ) Deviation [dB] Lower Upper Boundary Boundary [dBm-2 ) [dBm-2 ] 30 98.67 38.67 60.00 18.67 12.35 28.98 29.80 0.82 25.80 33.80 35 98.67 43.67 55.00 17.72 11.91 25.38 27.10 1.72 23.10 31.10 40 98.67 46.83 51.84 14.92 11.46 25.46 24.90 0.56 20.90 28.90 45 98.33 51.17 47.16 13.52 10.80 22.84 22.90 0.06 18.90 26.90 50 98.33 56.00 42.33 11.61 10.14 20.58 21.10 0.52 17.10 25.10 60 98.50 63.67 34.83 8.82 8.36 17.65 18.00 0.35 14.00 22.00 70 98.17 67.67 30.50 6.47 7.50 16.53 15.50 -1.03 11.50 19.50 80 97.83 69.50 28.33 6.71 8.37 13.25 13.30 0.05 9.30 17.30 90 97.83 68.00 29.83 8.49 7.74 13.60 11.40 -2.20 7.40 15.40 100 98.00 65.83 32.17 11.47 9.37 11.33 9.70 -1.63 5.70 13.70 120 97.83 65.33 32.50 13.62 11.60 7.28 7.00 -0.28 3.00 11.00 140 96.67 64.17 32.50 14.57 12.16 5.77 4.80 -0.97 0.80 8.80 160 96.50 64.67 31.83 16.65 12.60 2.59 3.10 0.52 -0.90 7.10 5.70 180 96.33 64.83 31.50 17.79 12.93 0.78 1.70 0.92 -2.30 200 96.00 62.17 33.83 18.55 14.34 0.94 0.60 -0.34 -3.40 4.60 250 95.67 71.83 23.84 13.26 12.28 -1.70 -1.60 0.10 -5.60 2.40 300 95.17 71.83 23.34 13.97 12.86 -3.49 -3.30 0.19 -7.30 0.70 350 94.83 69.83 25.00 15.31 14.33 -4.64 -4.66 -0.02 -8.66 -0.66 400 94.50 68.50 26.00 16.65 15.79 -6.44 -5.90 0.54 -9.90 -1.90 450 94.17 68.00 26.17 17.40 16.73 -7.96 -6.92 1.04 -10.92 -2.92 500 93.67 66.83 26.84 18.15 17.67 -8.98 -7.90 1.08 -11.90 -3.90 550 93.67 64.00 29.67 19.64 18.17 -8.14 -8.71 -0.57 -12.71 -4.71 600 93.50 64.33 29.17 21.13 18.66 -10.62 -9.50 1.12 -13.50 -5.50 650 93.33 63.50 29.83 22.20 19.35 -11.72 -10.18 1.54 -14.18 -6.18 700 93.00 62.50 30.50 23.26 20.04 -12.80 -10.80 2.00 -14.80 -6.80 750 93.33 62.50 30.83 23.49 20.30 -12.96 -11.44 1.52 -15.44 -7.44 800 94.00 63.33 30.67 23.71 20.56 -13.60 -12.00 1.60 -16.00 -8.00 850 93.00 61.50 31.50 23.79 21.07 -13.36 -12.36 1.00 -16.36 -8.36 900 92.50 61.17 31.33 23.86 21.57 -14.10 -12.80 1.30 -16.80 -8.80 950 92.00 59.17 32.83 24.63 22.09 -13.89 -13.29 0.60 -17.29 -9.29 1000 92.17 59.33 32.84 25.40 22.61 -15.17 -13.80 1.37 -17.80 -9.80 IEEE Antennas and Popagation Magazine, Vol. 52, No.3, June 2010 211 the solid lines are the real open-area test site measurement results, and the dots show the ±4 dB margins. As observed, the newly established open-area test site successully completed the calibra­ tion procedure. 4. ANSIIIEEE C63.4-1992, "American National Standard Guide for Methods of Measurement of Radio-Noise Emissions rom Low-Voltage Electrical and Electronic Equipment in the Range of 9 kHz to 40 GHz." Note that the antennas used in open-rea test site calibration were calibrated by the National Metrology Institute (UME) http://www.ume.tubitak.gov.tr) according to [S]. The antenna fac­ tor values as a unction of requency listed in Table 3 therefore belong to this calibration data. In fact, open-area test site calibra­ tion indirectly veriies the antenna calibration performed in UME. S. ANSI/IEEE C63.S-2006 (Revision of C63.5-2003) "American National Standard Guide for Electromagnetic Compatibility-Radi­ ated Emission Measurements in Electromagnetic Interference (EMI) Control-Calibration of Antennas (9 kHz to 40 GHz)." 7. Conclusions Open-area test site (OATS) calibration is an important engi­ neering task. The keywords for open-area test site calibration are traceability, accreditation, repeatability (precision), and accuracy. Both CISPR and ANSI/IEEE standards present every detail of the calibration procedure. Normalized site attenuation (NSA) is an important parameter for open-area test site calibration. 8. References 1. CISPR 16-1-4: 2003, "Radio Disturbance and Immunity Meas­ uring Apparatus - Ancillary Equipment - Radiated Disturbances," Geneva, Switzerland, Comite Intenational Special des Perturba­ tions Radioelectriques. 2. CISPRI6-I-S: 2003, "Antenna Calibration Test Sites (CALTS) for 30 MHz to 1000 MHz," Geneva, Switzerland, Comite Intena­ tional Special des Peturbations Radioelectriques. 3. Federal Communications Commission, "Calibration of a Radia­ tion Measurement Site - Site Attenuation," Bull. aCE 44. Wash­ ington, DC, US Govenment Printing Oice, September 1977, Docket 21371. 212 6. A. A. Smith, R. F. German, and J. B. Pate, "Calculation of Site Attenuation rom Antenna Factors," IEEE Transactions on Elec­ tromagnetic Compatibiliy, 24, 3, August 1982, pp. 301-316. 7. P. T. Trakadas nd C. N. Capsalis, "A Mixed Model for the Determination of Normalized Site Attenuation in OATS," IEEE Transactions on Electromagnetic Compatibiliy, 43, I, February 2001, pp. 29-36. 8. A. Asri, C. Vollaire, L. Nicolas, and D. Prebet, "Normalized Site Attenuation Standard Correction rom Numerical Computing," IEEE Transactions on Electromagnetic Compatibiliy, 38, 2, March 2002, pp. 693-696. 9. L. Sevgi, "The Antenna as a Transducer: Simple Circuit and Electromagnetic Models," IEEE Antennas and Propagation Maga­ zine, 49, 6, December 2007, pp. 211-218. 10. L. Sevgi, S. ;aklr, and G. ;aklr, "Antenna Calibration for EMC Tests and Measurements," IEEE Antenns and Propagation Magzine, 50, 3, June 2008, pp. 2IS-224. II. ANSI C63.6 - 1996 (Revision of C63.6-1988), "American National Standard Guide for the Computation of Errors in Open Area Test Site Measurements." 12. ANSI C63.7-1992 (Revision of C63.7-1988), "American National Standard Guide for Construction of Open Area Test Sites for Perfoming Radiated Emission Measurements." D IEEE Anennas and Popagaton Mgazine, Vol. 52, No. 3, June 2010