Radiation Protection Dosimetry Advance Access published September 8, 2014 Radiation Protection Dosimetry (2014), pp. 1–9
doi:10.1093/rpd/ncu275
AN INTERLABORATORY COMPARISON PROGRAMME ON RADIO FREQUENCY ELECTROMAGNETIC FIELD MEASUREMENTS: THE SECOND ROUND OF THE SCHEME E. P. Nicolopoulou1, I. N. Ztoupis1, E. Karabetsos2, I. F. Gonos1,* and I. A. Stathopulos1 1 High Voltage Laboratory, School of Electrical and Computer Engineering, National Technical University of Athens, 9 Iroon Politechniou Street, Zografou Campus, Athens GR 15780, Greece 2 Non Ionizing Radiation Office, Greek Atomic Energy Commission, Patriarchou Grigoriou and Neapoleos Street, Agia Paraskevi, Athens GR 15310, Greece *Corresponding author: [email protected] Received 18 June 2014; revised 1 August 2014; accepted 3 August 2014 The second round of an interlaboratory comparison scheme on radio frequency electromagnetic field measurements has been conducted in order to evaluate the overall performance of laboratories that perform measurements in the vicinity of mobile phone base stations and broadcast antenna facilities. The participants recorded the electric field strength produced by two high frequency signal generators inside an anechoic chamber in three measurement scenarios with the antennas transmitting each time different signals at the FM, VHF, UHF and GSM frequency bands. In each measurement scenario, the participants also used their measurements in order to calculate the relative exposure ratios. The results were evaluated in each test level calculating performance statistics (z-scores and En numbers). Subsequently, possible sources of errors for each participating laboratory were discussed, and the overall evaluation of their performances was determined by using an aggregated performance statistic. A comparison between the two rounds proves the necessity of the scheme.
INTRODUCTION Serious public concerns are expressed about the possible adverse effects on human health as a consequence of exposure to electromagnetic fields (EMFs). In the case of radio frequency (RF) fields, attention is focused on mobile phone base stations due to their proximity to dwellings and on broadcast antenna facilities due to their large amount of radiated power. According to the Greek legislation(1, 2), measurements of the EMF levels carried out by accredited laboratories in the vicinity of such facilities are necessary in order to evaluate the exposure levels of the general public. The relevant metric used to assess whether the reference levels for human exposure to EMFs are exceeded is the total exposure ratio (ER)(3 – 5). This is defined as the sum of the individual ERs concerning the same quantity (electric or magnetic field) and the same effect (thermal or electrical stimulation) at a measurement location at a specific time slot. For frequencies .10 MHz and ,10 GHz and for measurements in the far field of the radiating antenna, the ER for a specific frequency (or frequency band) and measurement position is calculated as follows: ER ¼
E2 H2 ¼ LE 2 LH 2
ð1Þ
where E (or H ) is the average value of the measured electric (or magnetic) field strength at a certain frequency and measurement position; LE (or LH) the
corresponding reference level for the electric (or magnetic) field at this frequency. Within this growing need for reliable and comparable measurement results, proficiency testing schemes have proved to be an efficient method for assessing the adequacy and improving the performance of a laboratory. According to the standard ISO/IEC 17025(6), every accredited test laboratory should demonstrate traceability and proficiency by participating in interlaboratory comparison (ILC) programmes that cover all measurements lying within its scope of accreditation. The term ILC describes the organisation, performance and evaluation of calibration/tests on the same or similar calibration/test items by two or more laboratories under predetermined conditions(7). The type of ILC applicable to EMF measurements is a results comparison programme, where all participants measure the same EMF source, usually in normal working conditions(8, 9). The objective of a laboratory should not be just a single successful participation in an ILC scheme that would certify its technical competence for a limited time, but a constant monitoring of its performance. Following the relevant policy of the Hellenic Accreditation System S.A. (ESYD) for the participation of laboratories in proficiency testing schemes every 4 y, if possible(10), the Greek Atomic Energy Commission organised the second round of an ILC scheme on EMF measurements in May 2013(8, 9). The scheme, which involved RF EMF measurements
# The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
E. P. NICOLOPOULOU ET AL.
and calculations of ERs in a controlled environment in different measurement scenarios, was set up to comply with the international requirements of the ILAC G13: 2000 Guidelines(11) and the ILAC Policy(12, 13). PARTICIPANTS-MEASURING EQUIPMENT Nineteen laboratories with a total of 30 measuring groups participated in this ILC procedure. Each participating group (Laboratories 1–30) was assigned with a unique identification code number which is kept secret, thus guaranteeing data confidentiality. It must be noted that a measurement is defined as the unique combination of personnel handling of the instruments, measuring equipment and measurement procedure. The participating teams used different types of measuring equipment as presented in Table 1.
axis in order to remain in the far field of the antennas. Also, the first antenna (R&S HE200) was vertically oriented with respect to the floor and the second one (Teseq CBL-6111D BiLog Antenna) was placed at an angle of 458 in order to examine the capacity of the laboratories to detect the three spatial components of the fields. The location and orientation of the antennas are depicted in Figure 1b. The broadcast frequencies of each generator in each of the three scenarios are presented in Table 2. It should be noted that the laboratories did not know the characteristics of the incident field that they were asked to measure in any of the three scenarios.
MEASUREMENT TASK DELIVERABLES The participants were asked at each measurement scenario to deliver the following measurands: †
MEASUREMENT PROCEDURE
†
The high-frequency EMF measurements were conducted in the facilities of the High Voltage Laboratory of the National Technical University of Athens inside the shielded anechoic chamber as shown in Figure 1a(9). The measured field within the anechoic chamber was produced by two signal generators operating at the frequency ranges of 250 kHz –4 GHz and 9 kHz –6 GHz connected via two power amplifiers with two proper antennas (Teseq CBL-6111D BiLog Antenna and R&S HE200, respectively). Adjusting the characteristics of the generator waveforms (frequency, power level) at different combinations enabled controlling the field and forming three measurement scenarios. In all the three scenarios, the measurements were conducted at a single predefined position, i.e. at a height of 1.5 m and at a horizontal distance of 4 m from the antenna
†
A 6-min average value of the electric field strength (ETOT) across the total measured spectral range. The maximum values of the electric field strength (EFM, EVHF, EUHF and EGSM) in the two frequency bands at which they detected the maximum emission level. The two frequency values ( fMAX1 and fMAX2) (in MHz) at which the maximum emission level is detected.
For each laboratory, a time slot of 30 min was allocated to execute the measurements. After completing the measurements, the participating teams had to calculate for each measurement scenario: † †
The total ER (ERTOT) throughout the whole spectral range (applying reduction factor limits of 60 %, as imposed by the Greek legislation (1)). The ERs (ERFM, ERVHF, ERUHF and ERGSM) in both frequency bands with the highest measured value of electric field strength.
Table 1. The measuring equipment used by the participating teams in the present round of the ILC. Type of equipment Frequency-selective field strength measuring a device combined with isotropic probe Broadband survey field meter combined with isotropic probe Portable spectrum analyser in conjunction with antenna
Model
Laboratories
NARDA/SRM-3000 NARDA/SRM-3006 NARDA/EMR-300
18 teams 7 teams 1 team
NARDA/EMR-300 Rohde & Schwartz/ESMI Receiver, SCHAFFNER/6143 Bilog Antenna Rohde & Schwartz/FSH4 Spectrum Analyzer, Rohde & Schwartz Isotropic Antenna Rohde & Schwartz/FSH4 Spectrum Analyzer, Rohde & Schwartz HE200 Antenna GW Instek/GSP-830 Spectrum Analyzer (3 kHz–3 GHz), Aaronia/Bicolog 30100 Antenna Z Technology/Log Periodic AA8 Antenna
1 team
Page 2 of 9
1 team 1 team 1 team
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
Figure 1. (a) Schematic of the exposure set-up for the ILC. (b) Location and orientation of the antennas inside the anechoic chamber.
Table
2. Broadcast frequencies for the scenarios of the ILC scheme.
measurement
Emission level (dBm)
Hewlett Packard E444338 Signal Generator 217
Rohde & Schwarz SMB 100A Signal Generator 228
Scenario 1 Scenario 2 Scenario 3
102.5 MHz (FM) 96.3 MHz (FM) 180 MHz (VHF)
935 MHz (GSM) 440 MHz (UHF) 720 MHz (UHF)
the method under perturbations or conditions of uncertainty. Thus, the produced robust statistics have a reasonably small bias and are resistant to errors in the results due to deviations from assumptions (e.g. of normality). This is of high importance for the EMF measurements that are dominated by numerous uncertainty factors. The robust estimators are then used for the calculation of the performance statistic z-score from the relationship: z¼
All participants had 3 d at their disposal to deliver the 18 quantities (6 for each measurement scenario) along with their expanded uncertainties (95 % confidence level). PERFORMANCE STATISTICS Z-scores The basic concept of ILCs is the evaluation of the laboratories through a statistical indicator of their competence ( performance statistic). The initial stage of the statistical analysis is to define the best available estimations for the true value of the measurand ^ and the dispersion of the measure(assigned value m) ments (standard deviation for the proficiency assessment s ^ ). As in the first round of the scheme(8), these estimators are calculated in each test level by applying the iterative robust algorithm described in the standard ISO 13528 (Annex C, Algorithm A)(14) to the measurements of the laboratories. The main advantage of this algorithm is robustness, the persistence of
^ xx ^ xm ¼ ^s ^ s
ð2Þ
^ the where x is the measurement of the laboratory, x robust mean and ^s the robust standard deviation. The results are evaluated according to the following regions of the z-score values(14): † jzj 2: the performance of the laboratory is satisfactory. † 2 , jzj , 3: the accuracy and correctness of the measurement is questionable, and the performance statistic is a ‘warning signal’. † jzj 3: the performance of the laboratory is nonsatisfactory and the performance statistic is an ‘action signal’. An ‘action signal’ is a strong indication of problems in the measuring performance of the laboratory and further investigation is required for the determination of the error factors which affect the measurement quality. ‘Warning signals’ are milder indicators of inaccuracy sources and raise concerns mainly if they reappear in various test levels (or various test rounds).
Page 3 of 9
E. P. NICOLOPOULOU ET AL.
En numbers The calculation of z-scores in some of the test levels for the detection of the maximum emission frequencies was not possible. The majority of the participants reported exactly the same value of frequency. As a result, the robust standard deviation in the first iteration of the algorithm was zero. In these cases, instead of the z-scores, the En numbers were calculated as defined in ref. (14): xX En ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 þ U2 Ulab ref
ð3Þ
where Ulab is the expanded uncertainty of the measurement x of each participating laboratory and Uref the expanded uncertainty of the estimator X defined in a reference laboratory. The frequency of the produced waveform as applied to the settings of the generator was chosen to be the estimate value X. For the uncertainty uref of the estimator X, the uncertainty of the frequency generators of the NTUA High Voltage Laboratory was adopted: uref ¼ 0:0035 %:
ð4Þ
Thus, the expanded uncertainty of the estimator value X (Uref ) was calculated as follows: Uref ¼ X uref :
ð5Þ
With an jEnj 1, the performance of the laboratory is considered satisfactory, while an jEnj .1 indicates a non-satisfactory performance.
Acceptance criterion set by the organiser Based on the experience gained in the first round of the scheme, an acceptance criterion has been set by the organiser for the overall proficiency assessment of the participants. Under the assumption that the measurements in a test level follow the normal distri^ and s bution and the robust values m ^ are good estimators, the z-score is considered a normally distributed random variable, with a mean value of 0 and a standard deviation of 1, regardless of the measurand, the test level and the measurement method. This enables combining the z-scores of individual test levels to create an aggregated score for the whole scheme round. A variable described by the normal distribution N(m, s 2) lies with 95 % probability in the interval m + 2s and with 99.7 % probability in the interval m+3s. Thus, the z-scores of a laboratory—which follow the distribution N(0, 1)—are expected to lie outside the region +2 in 5 % of the measurements and outside the region +3 only for 0.3 % of the test levels. A well-behaving laboratory is
expected to comply with these limits. Therefore, by calculating for each participant the number of individual test levels where it has received a jzj .2 as the percentage of the total number of its evaluated measurements, the following acceptance criteria arise: † †
If this percentage exceeds 5, the overall performance of the laboratory is non-satisfactory. If this percentage does not exceed 5, the overall performance of the laboratory is satisfactory.
The suitability and efficiency of the 5 % threshold was demonstrated in the first round where error sources were detected in all laboratories that exceeded it (8). Non-satisfactory En numbers ðjEn j . 1Þare counted in the form of the aggregated score only at those test levels where the calculation of z-scores was not possible. RESULTS The performance statistics (z-scores and En numbers) for the results of the participants in all deliverable quantities and measurement scenarios are presented in Tables 3 and 4. Furthermore, the aggregated performance statistics are shown both for the present round as well as for the first round of the scheme, where possible. Non-satisfactory results are highlighted. For reasons of completeness, all the En numbers are presented even in the frequency detection test levels where z-scores were calculated. EVALUATION CONCLUSIONS The purpose of the presented work was to describe the organisation and execution of an ILC scheme on RF EMF measurements and to assess its overall function. The purpose of this ILC had three components, which were to: † † †
allow the participating laboratories for objective verification of their technical competence following international standards; evaluate the overall performance of the participants under totally controllable conditions and in comparison with a previous ILC procedure; determine possible error factors in RF EMF measurements.
Comments on the second high-frequency ILC scheme The analysis of the results provides useful conclusions about error sources within the laboratories. Instrumentation in general and inappropriate settings used in the measuring equipment are the major factors that affect the accuracy of the results. In the present ILC scheme, 15 out of the 25 teams that performed frequency-selective measurements with the Narda Selective Radiation Meters (SRM 3000 and 3006) had an overall satisfying performance. On the
Page 4 of 9
Table 3. Z-scores for the electric field strength (E) and the ER in the whole measurement range (TOT) and in the respective frequency bands with the maximum emission in each measurement scenario (FM, GSM, UHF and VHF). Laboratory
Measurement scenario 2
Measurement scenario 3
ETOT
EFM
EGSM
ERTOT
ERFM
ERGSM
ETOT
EFM
EUHF
ERTOT
ERFM
ERUHF
ETOT
EVHF
EUHF
ERTOT
ERVHF
ERUHF
1.451 20.725 20.350 0.580 0.657 0.277 20.317 0.586 0.021 0.328 0.634 0.703 20.801 20.656 1.602 12.98 14.67 0.293 20.086 20.066 20.874 20.814 0.528 21.144 20.176 0.396 21.741 21.089 21.487 20.869
0.761 20.890 20.617 0.259 0.301 0.085 20.345 0.261 3.302 0.068 0.610 1.050 20.717 20.583 2.305 21.36 3.771 0.141 20.062 20.168 20.735 20.835 20.103 21.227 0.396 20.984 21.723 21.167 — —
0.221 0.241 20.724 0.225 20.696 20.685 20.690 21.132 4.588 20.705 21.319 0.344 20.235 20.224 1.076 5.891 14.69 20.272 0.270 0.390 20.534 20.717 20.713 20.326 20.963 2.668 2.031 20.326 — —
1.006 20.339 20.578 0.254 0.299 20.033 20.522 0.022 3.795 0.030 6.897 7.051 20.672 20.552 0.283 15.59 5.128 20.001 20.296 20.306 20.924 20.878 0.207 21.245 20.388 20.318 21.754 21.121 — —
0.773 21.070 20.661 0.231 0.276 0.049 20.389 0.233 3.863 0.028 0.637 1.117 20.546 20.410 2.960 42.78 4.562 0.020 20.101 20.209 20.773 20.933 20.140 21.342 0.380 20.484 21.831 21.182 — —
0.441 20.936 20.936 0.441 20.936 20.606 20.606 20.998 5.723 20.771 20.936 0.441 20.135 20.125 1.013 8.016 30.74 1.130 0.441 0.579 20.337 20.792 20.936 20.420 20.936 0.186 8.016 20.248 — —
1.018 20.936 0.406 0.474 0.322 0.259 0.753 0.364 0.217 0.016 0.859 20.379 20.969 21.073 0.460 13.90 34.81 0.761 0.410 0.544 20.316 20.219 0.982 21.050 0.271 21.830 21.666 22.673 21.175 20.504
0.532 20.957 0.263 0.230 20.022 20.077 0.014 20.019 6.203 20.121 0.705 20.013 20.838 20.897 1.755 15.52 6.865 0.173 0.378 0.304 20.387 20.294 0.577 21.076 0.019 22.494 21.486 22.292 — —
2.494 0.226 0.814 22.389 0.578 20.363 20.327 20.178 20.363 20.379 0.988 0.192 20.561 20.447 28.810 3.099 88.63 0.797 0.226 21.841 0.105 20.001 21.910 0.595 20.778 0.629 0.058 20.312 — —
0.426 21.066 0.002 0.075 20.058 20.095 0.231 20.072 6.979 20.239 5.806 4.194 20.693 20.759 0.063 24.12 21.52 0.164 0.031 0.117 20.461 20.401 0.426 20.884 20.098 21.388 21.308 21.831 — —
0.576 21.231 20.304 0.250 20.009 20.063 20.063 20.004 8.241 20.114 0.671 20.042 20.625 20.682 1.860 27.79 8.996 0.433 0.407 0.329 20.375 20.351 0.624 21.003 0.029 21.624 21.445 22.064 — —
1.156 0.619 1.156 22.016 1.156 20.262 20.239 20.094 20.324 20.339 1.156 0.389 20.412 20.306 22.402 26.46 532.5 20.377 0.236 21.604 20.163 20.163 21.911 0.389 20.377 0.535 20.377 20.377 — —
1.120 4.628 0.820 20.417 20.072 20.404 20.606 20.178 21.368 0.675 21.179 20.744 0.384 0.349 1.184 17.53 6.019 20.216 20.188 20.181 20.575 20.759 0.372 21.356 0.239 21.070 0.113 0.833 21.611 0.226
0.964 5.298 0.887 20.395 20.052 20.268 20.787 20.311 20.544 0.838 21.048 20.596 0.784 0.513 27.341 4.836 20.437 20.129 0.059 0.094 20.376 20.732 0.349 21.665 1.396 21.419 0.214 0.995 — —
0.569 0.529 0.999 20.600 20.768 0.675 0.986 0.152 0.185 0.264 0.661 1.475 20.433 20.708 20.675 2.395 26.92 0.794 0.013 20.887 21.139 21.094 20.552 0.145 21.409 20.013 21.184 20.993 — —
0.907 5.161 0.557 20.676 20.357 20.739 20.935 20.577 20.820 0.460 5.193 6.069 0.467 0.429 0.376 6.137 0.330 20.561 20.441 20.434 20.843 20.843 0.123 21.861 0.019 21.939 20.059 0.622 — —
0.892 5.689 0.814 20.427 20.093 20.309 20.786 20.362 20.564 0.759 21.013 20.604 1.045 0.765 25.407 5.259 20.331 20.297 0.000 0.034 20.410 20.410 0.282 21.722 1.332 21.865 0.248 0.930 — —
1.243 20.383 1.391 20.041 20.235 1.257 1.582 0.741 20.634 20.531 20.235 0.209 0.238 20.013 21.023 4.792 41.75 20.826 0.549 20.220 20.557 20.557 20.678 20.664 20.531 20.309 21.418 21.270 — —
The coloured cells are warning and/or action signals jzj .2.
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
Page 5 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Measurement scenario 1
Table 4. Performance statistics (z-scores and En numbers) for the detection of the two maximum emission frequencies ( fMAX1 and fMAX2) for the three measurement scenarios and aggregated percentages of the first and second rounds of the ILC scheme. Laboratory
Measurement scenario 1 fMAX1
fMAX2
fMAX1
Measurement scenario 3
fMAX2
fMAX1
fMAX2
z-score
En number
En number
z-score
En number
En number
En number
En number
20.208 20.208 20.208 23.095 20.208 1.727 22.283 21.940 20.208 0.514 20.208 20.208 20.005 0.081 3.545 9.955 20.208 20.496 22.026 1.149 1.756 0.716 20.208 20.208 0.947 20.208 20.208 20.208 20.208 —
0.000 0.000 0.000 20.040 0.000 0.717 20.974 20.264 0.000 0.334 0.000 0.000 0.147 0.210 0.023 0.023 0.000 20.001 20.051 0.038 0.301 0.142 0.000 0.000 0.778 0.000 0.000 0.000 0.000 —
0.000 0.000 0.000 0.040 0.000 0.131 20.038 20.489 0.000 59.709 0.000 0.000 20.079 20.218 0.004 20.001 0.000 0.000 0.002 0.000 20.001 0.001 0.000 0.000 0.000 0.000 0.000 0.000 152.788 —
0.142 0.142 0.142 23.883 0.142 20.180 1.600 20.462 0.142 0.142 0.142 0.142 0.262 0.262 2.355 6.420 25.895 20.060 21.287 20.865 21.166 21.307 20.236 0.142 20.865 0.343 0.142 1.148 4.166 —
0.000 0.000 0.000 20.080 0.000 20.171 0.976 20.133 0.000 0.000 0.000 0.000 0.126 0.126 0.022 0.022 21.042 20.001 20.061 20.043 20.294 20.325 20.002 0.000 21.036 0.023 0.000 0.208 8.909 —
0.000 0.000 0.000 20.040 0.000 0.010 0.001 0.000 20.655 0.000 0.000 0.000 0.000 0.000 0.025 20.001 0.000 20.102 0.002 20.012 0.000 0.000 20.001 0.000 0.000 0.000 0.000 0.000 0.000 —
0.000 0.000 0.000 20.080 0.000 0.000 20.003 0.000 0.000 0.000 0.000 0.000 20.021 20.021 0.011 0.011 0.000 0.000 0.038 0.038 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 —
0.000 0.000 0.000 20.600 0.000 0.052 0.002 0.000 20.393 0.000 0.000 0.000 0.000 0.000 0.002 0.007 0.000 0.000 0.004 0.004 0.002 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 —
The coloured cells are warning and/or action signals jzj .2 and non satisfying En numbers jEnj .1. In the last two columns the coloured cells are percentages above 5 % (non satisfying performances).
Number of jzj . 2 or jE n j . 1 2nd round
1 4 0 4 0 0 1 0 8 1 3 3 0 0 8 20 16 0 1 0 0 0 0 0 0 2 2 3 2 0
% of jzj . 2 or jE n j . 1 2nd round
4.2 16.7 0.0 16.7 0.0 0.0 4.2 0.0 33.3 4.2 12.5 12.5 0.0 0.0 33.3 83.3 66.7 0.0 4.2 0.0 0.0 0.0 0.0 0.0 0.0 8.3 8.3 12.5 22.2 0.0
% of jzj . 2 1st round
(1a) 0.0
(1b) 10.9 4.8 5.6 — 1.7 2.9 2.0 18.4 — 0.0 11.5 — — — — — 4.0 0.0 — — 1.7 1.7 — — — 1.4 — — 11.8 —
E. P. NICOLOPOULOU ET AL.
Page 6 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Measurement scenario 2
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
contrary, three out of the four teams that used a classic spectrum analyser with antennas were negatively evaluated (Laboratories 15, 16 and 29). The adjustment of the instrument settings, the implementation of the measurement procedure and the processing of the results should be done with special precaution when using a spectrum analyser. Especially, the use of a non-isotropic antenna in conjunction with the spectrum analyser should be followed by a proper combination of the three spatial field components. In the cases of Laboratories 15 and 16, which belonged to the same ‘laboratory body’ and conducted the measurements using different spectrum analysers, the use of a non-isotropic antenna has yielded worse results for Laboratory 16 compared with the isotropic antenna of Laboratory 15. Also, the non-satisfactory fMAX1 results of Laboratory 29 in measurement scenarios 1 and 2 can be attributed to the spectrum analyser and the non-isotropic antenna which this team used specifically for the detection of the frequencies with the maximum emission level. Apart from the polarisation of the field, the frequency range of the antenna used is a significant factor. In measurement scenario 1, which included a maximum emission in the FM band (102.5 MHz), some of the action signals that Laboratory 16 has received can be justified by the operating frequency range (800–2600 MHz) of the used Log Periodic antenna. The majority of the participants conducted measurements that covered the frequency range of 75 MHz–3 GHz. Laboratories 2 and 17 that reported a wider total frequency range (27 MHz–3 GHz) recorded higher field values—possibly because they detected harmonics emitted by the generators—and received action signals for their electric field strength measurements. The emitted fMAX1 in measurement scenarios 1 and 2 were close to the lower operating frequency range of the instrument used by Laboratory 4 (87.5 MHz–960 MHz). Therefore, the reported fMAX1 values received a jzj .2. An additional very important technical characteristic related to the spectral sensitivity of the measuring equipment is the resolution bandwidth (RBW). Improper adjustment of this parameter explains the action signals of Laboratory 9. The RBW settings for measurement scenario 1 (2 kHz for the FM band and 30 kHz for the GSM band, whereas the other participants used 100 kHz and 200 kHz, respectively) resulted in an EFM value significantly higher than the total field ETOT. Similarly, in measurement scenario 2, the RBW in the FM band was retained at 2 kHz and in the UHF band it was set at 200 kHz, much smaller than the setting used by the other teams (1 MHz). Moreover, the measurement range of the instrument in measurement scenario 1 was set at 2.8 V m21, a value much smaller than the recorded field values, which are therefore considered to have questionable accuracy.
Laboratories 11 and 12 (which had the same operators but different measuring equipment—SRM 3000 and SRM 3006, respectively) received action signals, in all the three scenarios, only for the total ERs. This indicates an error during the processing of the measurements or the application of the exposure limits for the calculation of the ER. It should be pointed out that a measurement of the electric field evaluated with a jzj . 2 is very likely to cause a jzj . 2 in the corresponding ER. The performance of the participants at the test levels regarding the detection of the two frequencies with maximum emission was satisfying, in general. The comparison between the En numbers and the corresponding z-scores shows a disagreement as there are cases where a jzj . 2 has not led to an jEnj . 1 ( fMAX1 scenarios 1 and 2) and vice versa (Laboratory 25, fMAX1 in measurement scenario 2). These deviations are due to the fact that the measurement uncertainty is needed for the calculation of the En numbers. Therefore, the accuracy in calculating the uncertainty is also evaluated. Unlike z-scores, this performance statistic is unaffected by the results of the other participants. A non-repetitive value jzj . 2 without obvious or systemic error sources as in the cases of Laboratories 1, 7, 10, 19, 26, 27 and 28 could be treated as occasional errors. The detection of possible technical weaknesses in most cases proves the satisfying function of the scheme. The fact that over one-third of the participants (12 out of 30 measuring teams) had a non-satisfying performance shows that technical competence on RF EMF measurements needs to be constantly recorded and evaluated through participation in ILC schemes. Participants with non-satisfactory performance should undertake internal control actions to verify whether the error sources suggested by the organiser are true and to detect possible further problems. Some possible corrective measures could include the change in the measuring equipment, calibration of the equipment, change/improvement in the measurement method and/or processing of the results and training of the personnel. These actions are necessary especially if the laboratory is already accredited or preparing to apply for accreditation. Accreditation bodies demand not only participation in ILCs, but also implementation of all the necessary corrective actions in the case of non-satisfactory performance. Besides, some error sources such as delayed calibration of the equipment arise as a non-conformity during an accreditation audit. Comparison between rounds Improvement proposals arising from the first round of the scheme were adopted and implemented in this second ILC round. In contrast to the first ILC which
Page 7 of 9
E. P. NICOLOPOULOU ET AL.
took place outdoors at an open site, the execution of the second round in the controlled environment of an anechoic chamber with predefined emission conditions enabled assuring maximum comparability of the measurements by eliminating the temporal variation of the field sources and other environmental factors. Hence, measuring equipment and staff performance are indeed the evaluated inaccuracy sources. Apart from the measurement environment, the measurement procedure itself changed in this round. Due to space limitation in the anechoic chamber and in order to remain in the far field of the broadcasting antennas, the electric field value was recorded only in a predetermined position. Additionally, the measurements at this position were made only at a height to reduce not only the required measurement time but also the heterogeneity of the deliverable results, which complicates the evaluation process. Towards this direction of increasing the uniformity of the deliverables, the total spectrum was divided into spectral bands with limits prescribed by the organiser. The ability of laboratories to identify the main spectral components of the measured field was investigated through measurements of frequency. Furthermore, an attempt was made to exploit the uncertainties of the measurements and incorporate them in the evaluation
process through the use of En numbers for the measurements of the frequency. Fifteen out of the 30 participating laboratories in this second round have also participated in the first ILC scheme. A direct comparison of the performance of laboratories (combination of personnel and equipment) is not possible as participating groups, measurement environment and test levels were not the same. Thus, a comparison of the % non-satisfactory results provides conclusions only on the effect of the measuring equipment and its settings. To make this comparison on a common basis (given the contribution of En numbers in the second round), the percentages ,5 % are considered fully satisfactory, between 5 and 10 % partially satisfactory and .10 % non-satisfactory. The comparison of the performance in both rounds of the ILC is presented in Table 4 and Figure 2 (where the used Laboratory codes refer to the present round of the ILC). The two aggregated scores in the first round for Laboratory 1 are due to the fact that it participated in the first round with the same measuring equipment but with two measuring teams with a different operator, and it was therefore given two identification numbers. To avoid a false interpretation of the results, these two teams are named Laboratory 1a and 1b, respectively, in Table 4 and Figure 2.
Figure 2. Comparative presentation of the aggregated % performance statistics for the laboratories that participated in both rounds of the scheme.
Page 8 of 9
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
The comparison between the two rounds proves the necessity of the ILC scheme. Cases of satisfying performance in the first round that are now followed by a partially- or non-satisfactory performance in this round (Laboratories 2, 17 and 26) show that technical competence on RF EMF measurements needs to be constantly recorded and evaluated through participation in ILC schemes. Improved performances in the present round (Laboratories 3, 8 and 29) prove the efficiency of an ILC scheme in the detection and the elimination of factors that minimise the accuracy of a measurement.
CONCLUSIONS The main conclusion is the efficiency of ILCs as a way to record and improve technical competence in the field of RF EMF measurements. Especially, the comparison between the two rounds has highlighted the necessity to repeat the ILC scheme. Obvious technical reasons, mainly including instrumentation, were detected in most cases of non-satisfactory performance. The calculated performance statistics verified the effect of such factors that are in advance expected to degrade the measurement quality, thus demonstrating the proper function of the ILC scheme, both as a test procedure—improved through execution in a controllable environment—and as an evaluation procedure—including adoption of the aggregated performance statistics and the 5 % acceptance criterion. Given the fact that in both rounds of the ILC the use of a spectrum analyser with an antenna was identified as a measurement method that requires special attention, a need for the future rounds is the development of a test specification that will ensure the correct use of spectrum analysers from all the participants. A special ILC for participants with spectrum analysers only would also contribute to the training of the staff and the detection of the exact error sources. REFERENCES 1. Greek Government Gazette, Athens, Greece. Law No 4070, Act. No. 82/A/10.04.2012: Regulations concerning electronic communications, transport, public works and other provisions (2012).
2. Greek Standardization Organization, Athens, Greece. ELOT 1422-3. Radiocommunication antenna collocation— Part 3: Testing and measurement techniques—limits (2006). 3. European Conference of Postal and Telecommunications Administrations, Copenhagen, Denmark. CEPT Revised ECC/REC/(02)04. Measuring non-ionising electromagnetic radiation (9 kHz–300 GHz) (2004). 4. European Committee for Standardization, Brussels, Belgium. EN 50492: Basic standard for the in-situ measurement of electromagnetic field strength related to human exposure in the vicinity of base stations (2008). 5. International Committee for Non-Ionizing Radiation Protection (ICNIRP). Guidelines for limiting exposure to time varying electric, magnetic and electromagnetic fields (up to 300 GHz). Health Phys. 74, 494–522 (1998). 6. International Electrotechnical Commission Geneva, Switzerland. ISO/IEC 17025: General requirements for the competence of testing and calibration laboratories (2005). 7. International Electrotechnical Commission Geneva, Switzerland. ISO/IEC 17043: Conformity assessment— General requirements for proficiency testing (2010). 8. Nicolopoulou, E. P., Gonos, I. F., Stathopulos, I. A. and Karabetsos, E. Two interlaboratory comparison programs on EMF measurements performed in Greece. IEEE Electromagnetic Compatibility Magazine 1, Q.2, 50–59 (2012). 9. Nicolopoulou, E. P., Ztoupis, I. N., Karabetsos, E., Gonos, I. F. and Stathopulos, I. A. An interlaboratory comparison programme on high frequency electromagnetic field measurements in a controllable environment performed in Greece. In: 16th International Congress of Metrology, Paris, France, 7 –10 October, PB-28. doi:10.1051/metrology/201311010 (2013). 10. Hellenic Accreditation System, Athens, Greece. ESYD PDI/02/01/02-09-2011: ESYD policy relevant to PT schemes and interlaboratory comparisons. Available on http://www.esyd.gr/pweb/s/20/files/EN/kanonismoi/ PDI_27_1_11.pdf. 11. International Laboratory Accreditation Cooperation, Silverwater, Australia. ILAC G13: Guidelines for the requirements for the competence of providers of proficiency testing schemes (2000). 12. International Laboratory Accreditation Cooperation, Silverwater, Australia. ILAC-P9: ILAC policy for participation in national and international proficiency testing activities (2005). 13. International Electrotechnical Commission, Geneva, Switzerland. ISO/IEC Guide 43-1: Proficiency testing by interlaboratory comparisons—Part 1: Development and operation of proficiency testing schemes (1997). 14. International Organization for Standardization, Geneva, Switzerland. ISO 13528: Statistical methods for use in proficiency testing by interlaboratory comparisons (2005).
Page 9 of 9
doi:10.1093/rpd/ncu275
AN INTERLABORATORY COMPARISON PROGRAMME ON RADIO FREQUENCY ELECTROMAGNETIC FIELD MEASUREMENTS: THE SECOND ROUND OF THE SCHEME E. P. Nicolopoulou1, I. N. Ztoupis1, E. Karabetsos2, I. F. Gonos1,* and I. A. Stathopulos1 1 High Voltage Laboratory, School of Electrical and Computer Engineering, National Technical University of Athens, 9 Iroon Politechniou Street, Zografou Campus, Athens GR 15780, Greece 2 Non Ionizing Radiation Office, Greek Atomic Energy Commission, Patriarchou Grigoriou and Neapoleos Street, Agia Paraskevi, Athens GR 15310, Greece *Corresponding author: [email protected] Received 18 June 2014; revised 1 August 2014; accepted 3 August 2014 The second round of an interlaboratory comparison scheme on radio frequency electromagnetic field measurements has been conducted in order to evaluate the overall performance of laboratories that perform measurements in the vicinity of mobile phone base stations and broadcast antenna facilities. The participants recorded the electric field strength produced by two high frequency signal generators inside an anechoic chamber in three measurement scenarios with the antennas transmitting each time different signals at the FM, VHF, UHF and GSM frequency bands. In each measurement scenario, the participants also used their measurements in order to calculate the relative exposure ratios. The results were evaluated in each test level calculating performance statistics (z-scores and En numbers). Subsequently, possible sources of errors for each participating laboratory were discussed, and the overall evaluation of their performances was determined by using an aggregated performance statistic. A comparison between the two rounds proves the necessity of the scheme.
INTRODUCTION Serious public concerns are expressed about the possible adverse effects on human health as a consequence of exposure to electromagnetic fields (EMFs). In the case of radio frequency (RF) fields, attention is focused on mobile phone base stations due to their proximity to dwellings and on broadcast antenna facilities due to their large amount of radiated power. According to the Greek legislation(1, 2), measurements of the EMF levels carried out by accredited laboratories in the vicinity of such facilities are necessary in order to evaluate the exposure levels of the general public. The relevant metric used to assess whether the reference levels for human exposure to EMFs are exceeded is the total exposure ratio (ER)(3 – 5). This is defined as the sum of the individual ERs concerning the same quantity (electric or magnetic field) and the same effect (thermal or electrical stimulation) at a measurement location at a specific time slot. For frequencies .10 MHz and ,10 GHz and for measurements in the far field of the radiating antenna, the ER for a specific frequency (or frequency band) and measurement position is calculated as follows: ER ¼
E2 H2 ¼ LE 2 LH 2
ð1Þ
where E (or H ) is the average value of the measured electric (or magnetic) field strength at a certain frequency and measurement position; LE (or LH) the
corresponding reference level for the electric (or magnetic) field at this frequency. Within this growing need for reliable and comparable measurement results, proficiency testing schemes have proved to be an efficient method for assessing the adequacy and improving the performance of a laboratory. According to the standard ISO/IEC 17025(6), every accredited test laboratory should demonstrate traceability and proficiency by participating in interlaboratory comparison (ILC) programmes that cover all measurements lying within its scope of accreditation. The term ILC describes the organisation, performance and evaluation of calibration/tests on the same or similar calibration/test items by two or more laboratories under predetermined conditions(7). The type of ILC applicable to EMF measurements is a results comparison programme, where all participants measure the same EMF source, usually in normal working conditions(8, 9). The objective of a laboratory should not be just a single successful participation in an ILC scheme that would certify its technical competence for a limited time, but a constant monitoring of its performance. Following the relevant policy of the Hellenic Accreditation System S.A. (ESYD) for the participation of laboratories in proficiency testing schemes every 4 y, if possible(10), the Greek Atomic Energy Commission organised the second round of an ILC scheme on EMF measurements in May 2013(8, 9). The scheme, which involved RF EMF measurements
# The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
E. P. NICOLOPOULOU ET AL.
and calculations of ERs in a controlled environment in different measurement scenarios, was set up to comply with the international requirements of the ILAC G13: 2000 Guidelines(11) and the ILAC Policy(12, 13). PARTICIPANTS-MEASURING EQUIPMENT Nineteen laboratories with a total of 30 measuring groups participated in this ILC procedure. Each participating group (Laboratories 1–30) was assigned with a unique identification code number which is kept secret, thus guaranteeing data confidentiality. It must be noted that a measurement is defined as the unique combination of personnel handling of the instruments, measuring equipment and measurement procedure. The participating teams used different types of measuring equipment as presented in Table 1.
axis in order to remain in the far field of the antennas. Also, the first antenna (R&S HE200) was vertically oriented with respect to the floor and the second one (Teseq CBL-6111D BiLog Antenna) was placed at an angle of 458 in order to examine the capacity of the laboratories to detect the three spatial components of the fields. The location and orientation of the antennas are depicted in Figure 1b. The broadcast frequencies of each generator in each of the three scenarios are presented in Table 2. It should be noted that the laboratories did not know the characteristics of the incident field that they were asked to measure in any of the three scenarios.
MEASUREMENT TASK DELIVERABLES The participants were asked at each measurement scenario to deliver the following measurands: †
MEASUREMENT PROCEDURE
†
The high-frequency EMF measurements were conducted in the facilities of the High Voltage Laboratory of the National Technical University of Athens inside the shielded anechoic chamber as shown in Figure 1a(9). The measured field within the anechoic chamber was produced by two signal generators operating at the frequency ranges of 250 kHz –4 GHz and 9 kHz –6 GHz connected via two power amplifiers with two proper antennas (Teseq CBL-6111D BiLog Antenna and R&S HE200, respectively). Adjusting the characteristics of the generator waveforms (frequency, power level) at different combinations enabled controlling the field and forming three measurement scenarios. In all the three scenarios, the measurements were conducted at a single predefined position, i.e. at a height of 1.5 m and at a horizontal distance of 4 m from the antenna
†
A 6-min average value of the electric field strength (ETOT) across the total measured spectral range. The maximum values of the electric field strength (EFM, EVHF, EUHF and EGSM) in the two frequency bands at which they detected the maximum emission level. The two frequency values ( fMAX1 and fMAX2) (in MHz) at which the maximum emission level is detected.
For each laboratory, a time slot of 30 min was allocated to execute the measurements. After completing the measurements, the participating teams had to calculate for each measurement scenario: † †
The total ER (ERTOT) throughout the whole spectral range (applying reduction factor limits of 60 %, as imposed by the Greek legislation (1)). The ERs (ERFM, ERVHF, ERUHF and ERGSM) in both frequency bands with the highest measured value of electric field strength.
Table 1. The measuring equipment used by the participating teams in the present round of the ILC. Type of equipment Frequency-selective field strength measuring a device combined with isotropic probe Broadband survey field meter combined with isotropic probe Portable spectrum analyser in conjunction with antenna
Model
Laboratories
NARDA/SRM-3000 NARDA/SRM-3006 NARDA/EMR-300
18 teams 7 teams 1 team
NARDA/EMR-300 Rohde & Schwartz/ESMI Receiver, SCHAFFNER/6143 Bilog Antenna Rohde & Schwartz/FSH4 Spectrum Analyzer, Rohde & Schwartz Isotropic Antenna Rohde & Schwartz/FSH4 Spectrum Analyzer, Rohde & Schwartz HE200 Antenna GW Instek/GSP-830 Spectrum Analyzer (3 kHz–3 GHz), Aaronia/Bicolog 30100 Antenna Z Technology/Log Periodic AA8 Antenna
1 team
Page 2 of 9
1 team 1 team 1 team
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
Figure 1. (a) Schematic of the exposure set-up for the ILC. (b) Location and orientation of the antennas inside the anechoic chamber.
Table
2. Broadcast frequencies for the scenarios of the ILC scheme.
measurement
Emission level (dBm)
Hewlett Packard E444338 Signal Generator 217
Rohde & Schwarz SMB 100A Signal Generator 228
Scenario 1 Scenario 2 Scenario 3
102.5 MHz (FM) 96.3 MHz (FM) 180 MHz (VHF)
935 MHz (GSM) 440 MHz (UHF) 720 MHz (UHF)
the method under perturbations or conditions of uncertainty. Thus, the produced robust statistics have a reasonably small bias and are resistant to errors in the results due to deviations from assumptions (e.g. of normality). This is of high importance for the EMF measurements that are dominated by numerous uncertainty factors. The robust estimators are then used for the calculation of the performance statistic z-score from the relationship: z¼
All participants had 3 d at their disposal to deliver the 18 quantities (6 for each measurement scenario) along with their expanded uncertainties (95 % confidence level). PERFORMANCE STATISTICS Z-scores The basic concept of ILCs is the evaluation of the laboratories through a statistical indicator of their competence ( performance statistic). The initial stage of the statistical analysis is to define the best available estimations for the true value of the measurand ^ and the dispersion of the measure(assigned value m) ments (standard deviation for the proficiency assessment s ^ ). As in the first round of the scheme(8), these estimators are calculated in each test level by applying the iterative robust algorithm described in the standard ISO 13528 (Annex C, Algorithm A)(14) to the measurements of the laboratories. The main advantage of this algorithm is robustness, the persistence of
^ xx ^ xm ¼ ^s ^ s
ð2Þ
^ the where x is the measurement of the laboratory, x robust mean and ^s the robust standard deviation. The results are evaluated according to the following regions of the z-score values(14): † jzj 2: the performance of the laboratory is satisfactory. † 2 , jzj , 3: the accuracy and correctness of the measurement is questionable, and the performance statistic is a ‘warning signal’. † jzj 3: the performance of the laboratory is nonsatisfactory and the performance statistic is an ‘action signal’. An ‘action signal’ is a strong indication of problems in the measuring performance of the laboratory and further investigation is required for the determination of the error factors which affect the measurement quality. ‘Warning signals’ are milder indicators of inaccuracy sources and raise concerns mainly if they reappear in various test levels (or various test rounds).
Page 3 of 9
E. P. NICOLOPOULOU ET AL.
En numbers The calculation of z-scores in some of the test levels for the detection of the maximum emission frequencies was not possible. The majority of the participants reported exactly the same value of frequency. As a result, the robust standard deviation in the first iteration of the algorithm was zero. In these cases, instead of the z-scores, the En numbers were calculated as defined in ref. (14): xX En ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 þ U2 Ulab ref
ð3Þ
where Ulab is the expanded uncertainty of the measurement x of each participating laboratory and Uref the expanded uncertainty of the estimator X defined in a reference laboratory. The frequency of the produced waveform as applied to the settings of the generator was chosen to be the estimate value X. For the uncertainty uref of the estimator X, the uncertainty of the frequency generators of the NTUA High Voltage Laboratory was adopted: uref ¼ 0:0035 %:
ð4Þ
Thus, the expanded uncertainty of the estimator value X (Uref ) was calculated as follows: Uref ¼ X uref :
ð5Þ
With an jEnj 1, the performance of the laboratory is considered satisfactory, while an jEnj .1 indicates a non-satisfactory performance.
Acceptance criterion set by the organiser Based on the experience gained in the first round of the scheme, an acceptance criterion has been set by the organiser for the overall proficiency assessment of the participants. Under the assumption that the measurements in a test level follow the normal distri^ and s bution and the robust values m ^ are good estimators, the z-score is considered a normally distributed random variable, with a mean value of 0 and a standard deviation of 1, regardless of the measurand, the test level and the measurement method. This enables combining the z-scores of individual test levels to create an aggregated score for the whole scheme round. A variable described by the normal distribution N(m, s 2) lies with 95 % probability in the interval m + 2s and with 99.7 % probability in the interval m+3s. Thus, the z-scores of a laboratory—which follow the distribution N(0, 1)—are expected to lie outside the region +2 in 5 % of the measurements and outside the region +3 only for 0.3 % of the test levels. A well-behaving laboratory is
expected to comply with these limits. Therefore, by calculating for each participant the number of individual test levels where it has received a jzj .2 as the percentage of the total number of its evaluated measurements, the following acceptance criteria arise: † †
If this percentage exceeds 5, the overall performance of the laboratory is non-satisfactory. If this percentage does not exceed 5, the overall performance of the laboratory is satisfactory.
The suitability and efficiency of the 5 % threshold was demonstrated in the first round where error sources were detected in all laboratories that exceeded it (8). Non-satisfactory En numbers ðjEn j . 1Þare counted in the form of the aggregated score only at those test levels where the calculation of z-scores was not possible. RESULTS The performance statistics (z-scores and En numbers) for the results of the participants in all deliverable quantities and measurement scenarios are presented in Tables 3 and 4. Furthermore, the aggregated performance statistics are shown both for the present round as well as for the first round of the scheme, where possible. Non-satisfactory results are highlighted. For reasons of completeness, all the En numbers are presented even in the frequency detection test levels where z-scores were calculated. EVALUATION CONCLUSIONS The purpose of the presented work was to describe the organisation and execution of an ILC scheme on RF EMF measurements and to assess its overall function. The purpose of this ILC had three components, which were to: † † †
allow the participating laboratories for objective verification of their technical competence following international standards; evaluate the overall performance of the participants under totally controllable conditions and in comparison with a previous ILC procedure; determine possible error factors in RF EMF measurements.
Comments on the second high-frequency ILC scheme The analysis of the results provides useful conclusions about error sources within the laboratories. Instrumentation in general and inappropriate settings used in the measuring equipment are the major factors that affect the accuracy of the results. In the present ILC scheme, 15 out of the 25 teams that performed frequency-selective measurements with the Narda Selective Radiation Meters (SRM 3000 and 3006) had an overall satisfying performance. On the
Page 4 of 9
Table 3. Z-scores for the electric field strength (E) and the ER in the whole measurement range (TOT) and in the respective frequency bands with the maximum emission in each measurement scenario (FM, GSM, UHF and VHF). Laboratory
Measurement scenario 2
Measurement scenario 3
ETOT
EFM
EGSM
ERTOT
ERFM
ERGSM
ETOT
EFM
EUHF
ERTOT
ERFM
ERUHF
ETOT
EVHF
EUHF
ERTOT
ERVHF
ERUHF
1.451 20.725 20.350 0.580 0.657 0.277 20.317 0.586 0.021 0.328 0.634 0.703 20.801 20.656 1.602 12.98 14.67 0.293 20.086 20.066 20.874 20.814 0.528 21.144 20.176 0.396 21.741 21.089 21.487 20.869
0.761 20.890 20.617 0.259 0.301 0.085 20.345 0.261 3.302 0.068 0.610 1.050 20.717 20.583 2.305 21.36 3.771 0.141 20.062 20.168 20.735 20.835 20.103 21.227 0.396 20.984 21.723 21.167 — —
0.221 0.241 20.724 0.225 20.696 20.685 20.690 21.132 4.588 20.705 21.319 0.344 20.235 20.224 1.076 5.891 14.69 20.272 0.270 0.390 20.534 20.717 20.713 20.326 20.963 2.668 2.031 20.326 — —
1.006 20.339 20.578 0.254 0.299 20.033 20.522 0.022 3.795 0.030 6.897 7.051 20.672 20.552 0.283 15.59 5.128 20.001 20.296 20.306 20.924 20.878 0.207 21.245 20.388 20.318 21.754 21.121 — —
0.773 21.070 20.661 0.231 0.276 0.049 20.389 0.233 3.863 0.028 0.637 1.117 20.546 20.410 2.960 42.78 4.562 0.020 20.101 20.209 20.773 20.933 20.140 21.342 0.380 20.484 21.831 21.182 — —
0.441 20.936 20.936 0.441 20.936 20.606 20.606 20.998 5.723 20.771 20.936 0.441 20.135 20.125 1.013 8.016 30.74 1.130 0.441 0.579 20.337 20.792 20.936 20.420 20.936 0.186 8.016 20.248 — —
1.018 20.936 0.406 0.474 0.322 0.259 0.753 0.364 0.217 0.016 0.859 20.379 20.969 21.073 0.460 13.90 34.81 0.761 0.410 0.544 20.316 20.219 0.982 21.050 0.271 21.830 21.666 22.673 21.175 20.504
0.532 20.957 0.263 0.230 20.022 20.077 0.014 20.019 6.203 20.121 0.705 20.013 20.838 20.897 1.755 15.52 6.865 0.173 0.378 0.304 20.387 20.294 0.577 21.076 0.019 22.494 21.486 22.292 — —
2.494 0.226 0.814 22.389 0.578 20.363 20.327 20.178 20.363 20.379 0.988 0.192 20.561 20.447 28.810 3.099 88.63 0.797 0.226 21.841 0.105 20.001 21.910 0.595 20.778 0.629 0.058 20.312 — —
0.426 21.066 0.002 0.075 20.058 20.095 0.231 20.072 6.979 20.239 5.806 4.194 20.693 20.759 0.063 24.12 21.52 0.164 0.031 0.117 20.461 20.401 0.426 20.884 20.098 21.388 21.308 21.831 — —
0.576 21.231 20.304 0.250 20.009 20.063 20.063 20.004 8.241 20.114 0.671 20.042 20.625 20.682 1.860 27.79 8.996 0.433 0.407 0.329 20.375 20.351 0.624 21.003 0.029 21.624 21.445 22.064 — —
1.156 0.619 1.156 22.016 1.156 20.262 20.239 20.094 20.324 20.339 1.156 0.389 20.412 20.306 22.402 26.46 532.5 20.377 0.236 21.604 20.163 20.163 21.911 0.389 20.377 0.535 20.377 20.377 — —
1.120 4.628 0.820 20.417 20.072 20.404 20.606 20.178 21.368 0.675 21.179 20.744 0.384 0.349 1.184 17.53 6.019 20.216 20.188 20.181 20.575 20.759 0.372 21.356 0.239 21.070 0.113 0.833 21.611 0.226
0.964 5.298 0.887 20.395 20.052 20.268 20.787 20.311 20.544 0.838 21.048 20.596 0.784 0.513 27.341 4.836 20.437 20.129 0.059 0.094 20.376 20.732 0.349 21.665 1.396 21.419 0.214 0.995 — —
0.569 0.529 0.999 20.600 20.768 0.675 0.986 0.152 0.185 0.264 0.661 1.475 20.433 20.708 20.675 2.395 26.92 0.794 0.013 20.887 21.139 21.094 20.552 0.145 21.409 20.013 21.184 20.993 — —
0.907 5.161 0.557 20.676 20.357 20.739 20.935 20.577 20.820 0.460 5.193 6.069 0.467 0.429 0.376 6.137 0.330 20.561 20.441 20.434 20.843 20.843 0.123 21.861 0.019 21.939 20.059 0.622 — —
0.892 5.689 0.814 20.427 20.093 20.309 20.786 20.362 20.564 0.759 21.013 20.604 1.045 0.765 25.407 5.259 20.331 20.297 0.000 0.034 20.410 20.410 0.282 21.722 1.332 21.865 0.248 0.930 — —
1.243 20.383 1.391 20.041 20.235 1.257 1.582 0.741 20.634 20.531 20.235 0.209 0.238 20.013 21.023 4.792 41.75 20.826 0.549 20.220 20.557 20.557 20.678 20.664 20.531 20.309 21.418 21.270 — —
The coloured cells are warning and/or action signals jzj .2.
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
Page 5 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Measurement scenario 1
Table 4. Performance statistics (z-scores and En numbers) for the detection of the two maximum emission frequencies ( fMAX1 and fMAX2) for the three measurement scenarios and aggregated percentages of the first and second rounds of the ILC scheme. Laboratory
Measurement scenario 1 fMAX1
fMAX2
fMAX1
Measurement scenario 3
fMAX2
fMAX1
fMAX2
z-score
En number
En number
z-score
En number
En number
En number
En number
20.208 20.208 20.208 23.095 20.208 1.727 22.283 21.940 20.208 0.514 20.208 20.208 20.005 0.081 3.545 9.955 20.208 20.496 22.026 1.149 1.756 0.716 20.208 20.208 0.947 20.208 20.208 20.208 20.208 —
0.000 0.000 0.000 20.040 0.000 0.717 20.974 20.264 0.000 0.334 0.000 0.000 0.147 0.210 0.023 0.023 0.000 20.001 20.051 0.038 0.301 0.142 0.000 0.000 0.778 0.000 0.000 0.000 0.000 —
0.000 0.000 0.000 0.040 0.000 0.131 20.038 20.489 0.000 59.709 0.000 0.000 20.079 20.218 0.004 20.001 0.000 0.000 0.002 0.000 20.001 0.001 0.000 0.000 0.000 0.000 0.000 0.000 152.788 —
0.142 0.142 0.142 23.883 0.142 20.180 1.600 20.462 0.142 0.142 0.142 0.142 0.262 0.262 2.355 6.420 25.895 20.060 21.287 20.865 21.166 21.307 20.236 0.142 20.865 0.343 0.142 1.148 4.166 —
0.000 0.000 0.000 20.080 0.000 20.171 0.976 20.133 0.000 0.000 0.000 0.000 0.126 0.126 0.022 0.022 21.042 20.001 20.061 20.043 20.294 20.325 20.002 0.000 21.036 0.023 0.000 0.208 8.909 —
0.000 0.000 0.000 20.040 0.000 0.010 0.001 0.000 20.655 0.000 0.000 0.000 0.000 0.000 0.025 20.001 0.000 20.102 0.002 20.012 0.000 0.000 20.001 0.000 0.000 0.000 0.000 0.000 0.000 —
0.000 0.000 0.000 20.080 0.000 0.000 20.003 0.000 0.000 0.000 0.000 0.000 20.021 20.021 0.011 0.011 0.000 0.000 0.038 0.038 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 —
0.000 0.000 0.000 20.600 0.000 0.052 0.002 0.000 20.393 0.000 0.000 0.000 0.000 0.000 0.002 0.007 0.000 0.000 0.004 0.004 0.002 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 —
The coloured cells are warning and/or action signals jzj .2 and non satisfying En numbers jEnj .1. In the last two columns the coloured cells are percentages above 5 % (non satisfying performances).
Number of jzj . 2 or jE n j . 1 2nd round
1 4 0 4 0 0 1 0 8 1 3 3 0 0 8 20 16 0 1 0 0 0 0 0 0 2 2 3 2 0
% of jzj . 2 or jE n j . 1 2nd round
4.2 16.7 0.0 16.7 0.0 0.0 4.2 0.0 33.3 4.2 12.5 12.5 0.0 0.0 33.3 83.3 66.7 0.0 4.2 0.0 0.0 0.0 0.0 0.0 0.0 8.3 8.3 12.5 22.2 0.0
% of jzj . 2 1st round
(1a) 0.0
(1b) 10.9 4.8 5.6 — 1.7 2.9 2.0 18.4 — 0.0 11.5 — — — — — 4.0 0.0 — — 1.7 1.7 — — — 1.4 — — 11.8 —
E. P. NICOLOPOULOU ET AL.
Page 6 of 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Measurement scenario 2
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
contrary, three out of the four teams that used a classic spectrum analyser with antennas were negatively evaluated (Laboratories 15, 16 and 29). The adjustment of the instrument settings, the implementation of the measurement procedure and the processing of the results should be done with special precaution when using a spectrum analyser. Especially, the use of a non-isotropic antenna in conjunction with the spectrum analyser should be followed by a proper combination of the three spatial field components. In the cases of Laboratories 15 and 16, which belonged to the same ‘laboratory body’ and conducted the measurements using different spectrum analysers, the use of a non-isotropic antenna has yielded worse results for Laboratory 16 compared with the isotropic antenna of Laboratory 15. Also, the non-satisfactory fMAX1 results of Laboratory 29 in measurement scenarios 1 and 2 can be attributed to the spectrum analyser and the non-isotropic antenna which this team used specifically for the detection of the frequencies with the maximum emission level. Apart from the polarisation of the field, the frequency range of the antenna used is a significant factor. In measurement scenario 1, which included a maximum emission in the FM band (102.5 MHz), some of the action signals that Laboratory 16 has received can be justified by the operating frequency range (800–2600 MHz) of the used Log Periodic antenna. The majority of the participants conducted measurements that covered the frequency range of 75 MHz–3 GHz. Laboratories 2 and 17 that reported a wider total frequency range (27 MHz–3 GHz) recorded higher field values—possibly because they detected harmonics emitted by the generators—and received action signals for their electric field strength measurements. The emitted fMAX1 in measurement scenarios 1 and 2 were close to the lower operating frequency range of the instrument used by Laboratory 4 (87.5 MHz–960 MHz). Therefore, the reported fMAX1 values received a jzj .2. An additional very important technical characteristic related to the spectral sensitivity of the measuring equipment is the resolution bandwidth (RBW). Improper adjustment of this parameter explains the action signals of Laboratory 9. The RBW settings for measurement scenario 1 (2 kHz for the FM band and 30 kHz for the GSM band, whereas the other participants used 100 kHz and 200 kHz, respectively) resulted in an EFM value significantly higher than the total field ETOT. Similarly, in measurement scenario 2, the RBW in the FM band was retained at 2 kHz and in the UHF band it was set at 200 kHz, much smaller than the setting used by the other teams (1 MHz). Moreover, the measurement range of the instrument in measurement scenario 1 was set at 2.8 V m21, a value much smaller than the recorded field values, which are therefore considered to have questionable accuracy.
Laboratories 11 and 12 (which had the same operators but different measuring equipment—SRM 3000 and SRM 3006, respectively) received action signals, in all the three scenarios, only for the total ERs. This indicates an error during the processing of the measurements or the application of the exposure limits for the calculation of the ER. It should be pointed out that a measurement of the electric field evaluated with a jzj . 2 is very likely to cause a jzj . 2 in the corresponding ER. The performance of the participants at the test levels regarding the detection of the two frequencies with maximum emission was satisfying, in general. The comparison between the En numbers and the corresponding z-scores shows a disagreement as there are cases where a jzj . 2 has not led to an jEnj . 1 ( fMAX1 scenarios 1 and 2) and vice versa (Laboratory 25, fMAX1 in measurement scenario 2). These deviations are due to the fact that the measurement uncertainty is needed for the calculation of the En numbers. Therefore, the accuracy in calculating the uncertainty is also evaluated. Unlike z-scores, this performance statistic is unaffected by the results of the other participants. A non-repetitive value jzj . 2 without obvious or systemic error sources as in the cases of Laboratories 1, 7, 10, 19, 26, 27 and 28 could be treated as occasional errors. The detection of possible technical weaknesses in most cases proves the satisfying function of the scheme. The fact that over one-third of the participants (12 out of 30 measuring teams) had a non-satisfying performance shows that technical competence on RF EMF measurements needs to be constantly recorded and evaluated through participation in ILC schemes. Participants with non-satisfactory performance should undertake internal control actions to verify whether the error sources suggested by the organiser are true and to detect possible further problems. Some possible corrective measures could include the change in the measuring equipment, calibration of the equipment, change/improvement in the measurement method and/or processing of the results and training of the personnel. These actions are necessary especially if the laboratory is already accredited or preparing to apply for accreditation. Accreditation bodies demand not only participation in ILCs, but also implementation of all the necessary corrective actions in the case of non-satisfactory performance. Besides, some error sources such as delayed calibration of the equipment arise as a non-conformity during an accreditation audit. Comparison between rounds Improvement proposals arising from the first round of the scheme were adopted and implemented in this second ILC round. In contrast to the first ILC which
Page 7 of 9
E. P. NICOLOPOULOU ET AL.
took place outdoors at an open site, the execution of the second round in the controlled environment of an anechoic chamber with predefined emission conditions enabled assuring maximum comparability of the measurements by eliminating the temporal variation of the field sources and other environmental factors. Hence, measuring equipment and staff performance are indeed the evaluated inaccuracy sources. Apart from the measurement environment, the measurement procedure itself changed in this round. Due to space limitation in the anechoic chamber and in order to remain in the far field of the broadcasting antennas, the electric field value was recorded only in a predetermined position. Additionally, the measurements at this position were made only at a height to reduce not only the required measurement time but also the heterogeneity of the deliverable results, which complicates the evaluation process. Towards this direction of increasing the uniformity of the deliverables, the total spectrum was divided into spectral bands with limits prescribed by the organiser. The ability of laboratories to identify the main spectral components of the measured field was investigated through measurements of frequency. Furthermore, an attempt was made to exploit the uncertainties of the measurements and incorporate them in the evaluation
process through the use of En numbers for the measurements of the frequency. Fifteen out of the 30 participating laboratories in this second round have also participated in the first ILC scheme. A direct comparison of the performance of laboratories (combination of personnel and equipment) is not possible as participating groups, measurement environment and test levels were not the same. Thus, a comparison of the % non-satisfactory results provides conclusions only on the effect of the measuring equipment and its settings. To make this comparison on a common basis (given the contribution of En numbers in the second round), the percentages ,5 % are considered fully satisfactory, between 5 and 10 % partially satisfactory and .10 % non-satisfactory. The comparison of the performance in both rounds of the ILC is presented in Table 4 and Figure 2 (where the used Laboratory codes refer to the present round of the ILC). The two aggregated scores in the first round for Laboratory 1 are due to the fact that it participated in the first round with the same measuring equipment but with two measuring teams with a different operator, and it was therefore given two identification numbers. To avoid a false interpretation of the results, these two teams are named Laboratory 1a and 1b, respectively, in Table 4 and Figure 2.
Figure 2. Comparative presentation of the aggregated % performance statistics for the laboratories that participated in both rounds of the scheme.
Page 8 of 9
AN ILC PROGRAMME ON RF EMF MEASUREMENTS
The comparison between the two rounds proves the necessity of the ILC scheme. Cases of satisfying performance in the first round that are now followed by a partially- or non-satisfactory performance in this round (Laboratories 2, 17 and 26) show that technical competence on RF EMF measurements needs to be constantly recorded and evaluated through participation in ILC schemes. Improved performances in the present round (Laboratories 3, 8 and 29) prove the efficiency of an ILC scheme in the detection and the elimination of factors that minimise the accuracy of a measurement.
CONCLUSIONS The main conclusion is the efficiency of ILCs as a way to record and improve technical competence in the field of RF EMF measurements. Especially, the comparison between the two rounds has highlighted the necessity to repeat the ILC scheme. Obvious technical reasons, mainly including instrumentation, were detected in most cases of non-satisfactory performance. The calculated performance statistics verified the effect of such factors that are in advance expected to degrade the measurement quality, thus demonstrating the proper function of the ILC scheme, both as a test procedure—improved through execution in a controllable environment—and as an evaluation procedure—including adoption of the aggregated performance statistics and the 5 % acceptance criterion. Given the fact that in both rounds of the ILC the use of a spectrum analyser with an antenna was identified as a measurement method that requires special attention, a need for the future rounds is the development of a test specification that will ensure the correct use of spectrum analysers from all the participants. A special ILC for participants with spectrum analysers only would also contribute to the training of the staff and the detection of the exact error sources. REFERENCES 1. Greek Government Gazette, Athens, Greece. Law No 4070, Act. No. 82/A/10.04.2012: Regulations concerning electronic communications, transport, public works and other provisions (2012).
2. Greek Standardization Organization, Athens, Greece. ELOT 1422-3. Radiocommunication antenna collocation— Part 3: Testing and measurement techniques—limits (2006). 3. European Conference of Postal and Telecommunications Administrations, Copenhagen, Denmark. CEPT Revised ECC/REC/(02)04. Measuring non-ionising electromagnetic radiation (9 kHz–300 GHz) (2004). 4. European Committee for Standardization, Brussels, Belgium. EN 50492: Basic standard for the in-situ measurement of electromagnetic field strength related to human exposure in the vicinity of base stations (2008). 5. International Committee for Non-Ionizing Radiation Protection (ICNIRP). Guidelines for limiting exposure to time varying electric, magnetic and electromagnetic fields (up to 300 GHz). Health Phys. 74, 494–522 (1998). 6. International Electrotechnical Commission Geneva, Switzerland. ISO/IEC 17025: General requirements for the competence of testing and calibration laboratories (2005). 7. International Electrotechnical Commission Geneva, Switzerland. ISO/IEC 17043: Conformity assessment— General requirements for proficiency testing (2010). 8. Nicolopoulou, E. P., Gonos, I. F., Stathopulos, I. A. and Karabetsos, E. Two interlaboratory comparison programs on EMF measurements performed in Greece. IEEE Electromagnetic Compatibility Magazine 1, Q.2, 50–59 (2012). 9. Nicolopoulou, E. P., Ztoupis, I. N., Karabetsos, E., Gonos, I. F. and Stathopulos, I. A. An interlaboratory comparison programme on high frequency electromagnetic field measurements in a controllable environment performed in Greece. In: 16th International Congress of Metrology, Paris, France, 7 –10 October, PB-28. doi:10.1051/metrology/201311010 (2013). 10. Hellenic Accreditation System, Athens, Greece. ESYD PDI/02/01/02-09-2011: ESYD policy relevant to PT schemes and interlaboratory comparisons. Available on http://www.esyd.gr/pweb/s/20/files/EN/kanonismoi/ PDI_27_1_11.pdf. 11. International Laboratory Accreditation Cooperation, Silverwater, Australia. ILAC G13: Guidelines for the requirements for the competence of providers of proficiency testing schemes (2000). 12. International Laboratory Accreditation Cooperation, Silverwater, Australia. ILAC-P9: ILAC policy for participation in national and international proficiency testing activities (2005). 13. International Electrotechnical Commission, Geneva, Switzerland. ISO/IEC Guide 43-1: Proficiency testing by interlaboratory comparisons—Part 1: Development and operation of proficiency testing schemes (1997). 14. International Organization for Standardization, Geneva, Switzerland. ISO 13528: Statistical methods for use in proficiency testing by interlaboratory comparisons (2005).
Page 9 of 9
