Investigation of AMN Impedance Uncertainty Contribution in Conducted Emission Tests

Soydan Cakir1, Osman Sen1, Alexander Kriz2, Serdar Büyük1, Aykut Ayaydın1 1 TUBITAK National Metrology Institute (TUBITAK UME), Kocaeli, Turkiye 2 EMC and Optics, Seibersdorf Laboratories, 2444 Seibersdorf, Austria Abstract— The conducted emission test is one of the important EMC tests in the frequency range of 10 kHz – 30 MHz for commercial and military equipment. AMNs are required in order to perform measurements and AMN impedance contributes to measurement uncertainty to some extent. Although AMN impedance is defined in terms of magnitude & phase and its effects are studied in commercial standards, most standards such as military standards only define the magnitude and give no information about the uncertainty contribution. In this paper, firstly we developed a software solution to calculate the theoretical nominal AMN impedance values for a given AMN circuit and calculated the corresponding uncertainty contribution by using the magnitude & phase information in accordance with CISPR 16-4-2. Following this, we instructively studied a variety of widely used AMN types and performed their uncertainty analysis. Finally, we focused particularly on military tests and the uncertainty contribution of the defined military AMNs to military conducted emission tests where phase information is not defined. Keywords— AMN, Conducted, CE102, EMC, Emission, LISN, Military, Uncertainty. I. INTRODUCTION The conducted emission test is applied to commercial and military equipment by measuring conducted emissions from supply ports of the Equipment Under Test (EUT) often in the frequency range of 10 kHz - 30 MHz in order to determine emission levels of the EUT. The conducted emission test firstly requires a verification process performed with known calibrated signal levels or reference sources just before the test in some standards such as MIL-STD-461G [1] or as per quality systems of accredited laboratories. During the verification, generally a known voltage level is applied to the EUT ports of the Artificial Mains Networks (AMN) and it is checked that whether the system displays expected results. If the deviation between the known voltage level and the measured voltage is higher than ± 3 dB, the system is accounted problematic and it should be rectified before proceeding to the test stage. In the test stage, voltage emissions from the EUT are measured through AMNs and a frequency-selective receiver. Although the EUT is supplied through the AMNs during the test, the AMN impedance deviation from the nominal impedance affects the test results and contributes to measurement uncertainties to some extent. The specifications of the commercial AMNs are given in CISPR 16-1-2 [2]. There are two prominent types of AMNs defined in CISPR 16-1-2 to cover the frequency range from 9 kHz to 30 MHz; the 50 Ω / 50 μH + 5 Ω AMN from 9 kHz to 150 kHz and the 50 Ω / 50 μH AMN from 150 kHz to 30 MHz. In CISPR 16-1-2, the tolerance of the impedance is specified. Boundaries are given for the magnitude with ± 20 % and for the phase with ±11.5° across the whole frequency band. This means that the tolerance is accounted an annulus sector around the nominal impedance as seen in Fig. 1. It is assumed that both boundaries are not independent from each other as per CISPR 16-4-2 [3 - 6]. On the other hand, there seems to have been an ambiguity between the specification in CISPR 16-1-2 and the impact of the tolerance given in CISPR 16-4-2. A rectangular tolerance sector (independently ± 20 % for the magnitude and ± 11.5° for the phase) appears to be defined in CISPR 16-1-2 whereas the calculation of the impact is based on a tolerance circle (dependently ± 20 % for the magnitude and ± 11.5° for the phase as seen in Fig. 1) in CISPR 16-4-2. In fact, this is interpreted in a different manner by some people, also by many calibration laboratories and AMN manufacturers, as follows because of this ambiguity; the annular tolerance in CISPR 16-4-2 is only for defining nominal AMN uncertainty component δZAN which is used to calculate UCISPR in CISPR 16-4-2 for conducted emission tests. Having an AMN impedance uncertainty value (δZAMN) larger than δZAN will only mean a higher uncertainty contribution to the overall expanded test uncertainty but does not mean the AMN cannot be used for testing as long as the CISPR16-1-2 magnitude limits are met. If the CISPR16-1-2 magnitude limits are exceeded, that signifies the AMN cannot be used for testing and must be fixed before placing it in use. Nevertheless, exceeding phase limits is tolerable and it is evaluated in the uncertainty calculation. As a result of all, due to the aforementioned ambiguity, it might be deduced that CISPR 161-2 states independent magnitude and phase tolerances for checking the AMN adequacy while CISPR 16-4-2 states dependent and annular magnitude and phase tolerances for particularly defining UCISPR. In the scope of this paper, δZAN will denote the uncertainty component in CISPR 16-4-2 for UCISPR calculation whereas δZAMN will denote the uncertainty component that results from an actual AMN used in emission testing. Although CISPR16-1-2 defines a phase tolerance, MIL-STD461G defines a 50 μH AMN whose impedance is as shown in Fig. 3 along with its circuit diagram in Fig. 6 (a) without phase information and tolerance. In the appendix of MIL-STD-461G, 5 µH AMNs are also allowed to facilitate testing for EUTs not able to operate properly with 50 µH AMNs. In addition, there is no information or guidance about uncertainty calculation in Investigation of AMN Impedance Uncertainty Contribution in Conducted Emission Tests MIL-STD-461G. Therefore, in this paper, we firstly developed a piece of software which presents more options and features in comparison to the software solutions in literature. In fact, there is a useful software solution in literature in [6] to calculate the AMN uncertainty contribution but is limited to a number of AMNs and for one frequency at a time. Our developed software allows AMN circuit definition and subsequently calculates nominal impedance magnitude & phase information seen by the EUT along with the nominal AMN uncertainty component (δZAN) which is used for UCISPR calculation and based on the annular tolerance as per CISPR 16-4-2. Additionally, the same software calculates the actual AMN uncertainty component (δZAMN) to be directly placed in the uncertainty calculation budget of an actual emission test when impedance magnitude and phase values in an actual AMN calibration report taken from a calibration laboratory are entered into the software. Following this, we performed AMN impedance uncertainty analysis for most commonly used AMN types. Finally, we specially focused on the military conducted emission test called CE102 where a phase tolerance for AMNs is not available and which is very often performed in our laboratory. Fig. 1. Definition of impedance magnitude and phase tolerances [2-6]. (a) (b) Fig. 2. Disturbance voltage measurement model a) ideal case, b) real case [4]. II. ANALYSIS In CISPR 16-4-1 [7], a measurement model for the conducted disturbance measurement method is given. The disturbance source is modelled by the disturbance voltage VEUT and the source impedance ZEUT. In case of an ideal setup the voltage VNOM is measured at the nominal impedance of the ZNOM as seen in Fig. 2 (a). Due to tolerances the actual impedance will deviate from the nominal impedance. In this case the source is terminated with the impedance ZAMN, where the disturbance voltage VAMN is measured as depicted in Fig. 2 (b). The measurement error is defined in a log scale as in (1) [4, 6]. ∆𝑉 =20𝑙𝑜𝑔(|𝑉𝐴𝑀𝑁 𝑉𝑁𝑂𝑀|) (1) Fig. 3. MIL-STD-461G AMN impedance magnitude graph and tolerance [1]. In CISPR 16-4-2 a worst case analysis is performed to get the measurement error. The calculation is based on the work of [8], which uses reflection coefficients normalized to system impedance Zo = 50 as given in (2) to (4) [4, 6]; Γ𝑁𝑂𝑀 =𝑍𝑁𝑂𝑀 − 𝑍𝑜 𝑍𝑁𝑂𝑀 + 𝑍𝑜 (2) Γ𝐴𝑀𝑁 =𝑍𝐴𝑀𝑁 − 𝑍𝑜 𝑍𝐴𝑀𝑁 + 𝑍𝑜 (3) Γ𝐸𝑈𝑇 =𝑍𝐸𝑈𝑇 − 𝑍𝑜 𝑍𝐸𝑈𝑇 + 𝑍𝑜=𝜌exp (𝑗φ) (4) The voltages are derived by using (2) to (4) as given in (5) and (6); 𝑉𝑁𝑂𝑀 =𝑍𝑁𝑂𝑀 𝑍𝐸𝑈𝑇 + 𝑍𝑁𝑂𝑀 𝑉𝐸𝑈𝑇 (5) 𝑉𝐴𝑀𝑁 =𝑍𝐴𝑀𝑁 𝑍𝐸𝑈𝑇 + 𝑍𝐴𝑀𝑁 𝑉𝐸𝑈𝑇 (6) The measurement error is calculated using (7); ∆𝑈 = 20𝑙𝑜𝑔(| 1+Γ𝐴𝑀𝑁 1−Γ𝐴𝑀𝑁Γ𝐸𝑈𝑇 1−Γ𝑁𝑂𝑀Γ𝐸𝑈𝑇 1+Γ𝑁𝑂𝑀 |) (7) Due to the impedance specification, ZAMN with a tolerance circle is calculated by (8). It is assumed that the worst case will occur if ZAMN is exactly on this circle and not inside the circle. Consequently, in (4), 𝜌 = 1 and |Γ𝐸𝑈𝑇| is assumed to be 1 [4, 6]. 𝑍𝐴𝑀𝑁 = 𝑍𝑁𝑂𝑀 +0.2|𝑍𝑁𝑂𝑀|𝑒𝑗𝜃 (8) 0≤𝜃 <2𝜋 (9) All possible combinations of φ and 𝜃 were used with 0.5 degrees steps in (4) and (8) in the analysis. The software window developed in the scope of this study is presented in Fig. 4 (a) along with the most general AMN circuit diagram in Fig. 4 (b) analyzed by the software. As depicted in Fig. 4 (a), the software allows the definition of AMNs in a large scale by entering the required AMN circuit component values on the AMN circuit and subsequently produces nominal AMN impedance (ZNOM) seen from the EUT and based on the entered values. It also accepts the actual AMN impedance magnitude and phase values (ZAMN) taken from a calibration laboratory. At the end, while it produces the AMN impedance nominal values (ZNOM) seen from the EUT as depicted in Fig. 4 (b), it also produces uncertainty contribution δZAMN for the actual AMN impedance, obtained from a AMN calibration laboratory, along with the nominal AMN uncertainty component (δZAN) which is based on the annular tolerance and used for UCISPR calculation as per CISPR 16-4-2. The software additionally allows adjusting the annular impedance magnitude tolerance, if requested, from the 20 % default value to any desired value for special analyses, and changing the 0.5o default φ and 𝜃 step in order to increase and decrease the depth of the analysis. (a) (b) Fig. 4. (a) Software window for AMN impedance and uncertainty calculation, (b) the most general AMN diagram used in the software. In the scope of the project, we studied the commonly used AMN types whose circuits are given in Fig. 5 – Fig. 8. The power source symbols Fig. 5 – Fig. 8 should not be confused with the voltage source symbol used in Fig. 2. The voltage source symbol seen in Fig. 2 is the internal disturbance source (VEUT) inside the EUT, whereas the source symbol seen in Fig. 5 – Fig. 8 is the external power source (Vs) (e.g. 220/230 VAC, 50/60 Hz) that supplies the EUT. As seen in Fig. 5 – Fig. 8, depending on the related standard, the AMNs have slight or significant variations in design and components between them. When we examine the circuitry of the AMNs more carefully, we see that the CISPR 16-1-2 50 Ω / 50 µH + 5 Ω & CISPR 16-1-2 50 Ω / 50 µH combination, MIL-STD-461 50µH, ANSI C63.4 [9] AMNs are fundamentally similar apart from some minor components & values. For that reason, we selected the MIL-STD-461G 50µH AMN along with CISPR 16-1-2 50 Ω / 50 µH + 5 Ω & CISPR 16-1-2 50 Ω / 50 µH AMN combination for analysis to represent this AMN group. Likewise, the MIL-STD-461G 5 µH, CISPR 25 [10], and DO160G [11] AMNs are similar and we selected the MIL-STD461G 5 µH AMN for analysis to represent this second group. However, as ISO 7637-2 [12] and ESA ECSS-E-ST-20-07C [13] AMNs are markedly different in terms of circuitry and analysis results, we analyzed each of them separately. As a consequence, for each selected AMN type, we calculated nominal impedance magnitude & phase impedance curves and δZAN stated in CISPR16-4-2 for UCISPR calculation based on the tolerance circle shown in Fig. 1. In this study, regardless of the relevant standards of the studied AMNs, all the AMN types were analyzed in the same configuration with the following parameters for consistency; the short-circuit and open-circuit source impedance conditions, the frequency range of 9 kHz - 400 MHz that reflects the widest AMN frequency range considering most of the AMN-related standards, and annular 20 % magnitude & ± 11.5o phase tolerance. In fact, the treatment of the source side differs from one standard to another during impedance calibration. For instance, the opencircuit condition is demanded for ANSI C63.4, MIL-STD-461, DO-160G while the short circuit condition is for CISPR 25, ISO 11452-2. Both of the conditions have good arguments; the short-circuit condition may signify the theoretical correct treatment of a voltage source whereas the open-circuit condition may denote that the source side of the AMN is connected to the source with a cable. From an RF perspective the impedance of this cable is high due to the inductance of the cable. Additionally, it is practical in calibration with the opencircuit condition since no additional adapter is required. Our prominent aim was to show the usability of the software in any desired configuration even different from the configurations stated in the related AMN standards, if required. In addition, independently of the relevant standards of the analysed AMN types & their related frequency ranges, it was aimed to set out the results in the entire frequency range (9 kHz – 400 MHz) per AMN. Therefore, readers can also have the chance of seeing the full picture of the entire range which is not shown in the relevant AMN standards. Afterwards, we particularly focused on MIL-STD-461G 50 µH and 5 µH AMNs as they are often utilized in our laboratory in the MIL-STD-461G CE102 test and no phase tolerance is defined for them in the standard. In this context, we applied the uncertainty calculation methodology stated in CISPR 16-4-2 to the military CE102 test for both the open-circuit and shortcircuit power source impedance conditions. In this context, we firstly produced nominal impedance values and δZAN, which is stated in CISPR16-4-2 for UCISPR calculation based on the tolerance circle, for the military AMN by using the developed software. Thereafter, we used one of our actual AMNs and its calibration certificate taken from a calibration laboratory in an attempt to calculate the actual AMN uncertainty component (δZAMN) to be placed into the uncertainty budget of the CE102 test. Although phase information is not stipulated by the standard for military AMNs, we requested the phase information from the calibration laboratory along with the magnitude information particularly for this research. Consequently, the actual AMN uncertainty component (δZAMN) was calculated to be used in CE102 uncertainty budget. (a) (b) Fig. 5. CISPR 16-1-2 ideal AMN diagrams excerpted from the standard (a) 50 Ω / 50 µH + 5 Ω, (b) 50 Ω / 50 µH. (a) (b) Fig. 6. MIL-STD-461G AMN (a) 50 µH, (b) 5 µH. (a) (b) Fig. 7. Automotive test AMNs as per (a) ISO 7637-2, (b) CISPR 25. Finally, we formed a MIL-STD-461G 50 µH AMN simulation with an intentionally and slightly exaggerated phase deviation higher than ±11o in order to have an idea of the possible impact of a significant phase deviation on the AMN uncertainty component. In this simulation, while precisely keeping the AMN impedance magnitude at the nominal impedance magnitude, we intentionally increased the phase deviation by 50o from the nominal phase for the simulated AMN for all frequencies in the software and put it into analysis and evaluated the results. Conversely, we also reduced the CISPR 16-4-2 annular magnitude and phase impedance tolerance from 20% / 11.5o to 1 % / 0.57o for the MIL-STD 50 µH AMN (RS : short-circuit) selected just as an example to check the extent of improvement in voltage deviation through the significant tolerance reduction. III. ANALYSIS RESULTS AND DISCUSSIONS The nominal impedance magnitude and phase values of the most commonly used AMNs, which were calculated by the developed software, are presented in Fig. 9 (a) – Fig. 17 (a). Additionally, δZAN values based on the annular tolerance in accordance with CISPR 16-4-2 are given in Fig. 9 (b) – Fig. 17 (b) and Table I with the obtained peak values. The second column of Table I shows the maximum and minimum AMN voltage deviations in the range of 9 kHz – 400 MHz which is deemed as the widest AMN frequency range in this research regardless of the AMN’s relevant standards, whereas the third column shows the values particularly in the frequency range stated in the standard where the relevant AMN is defined. (a) (b) (c) Fig. 8. Other reputable AMN types (a) DO-160G, (b) ANSI C63.4, (c) ESA ECSS-E-ST-20-07C. As clearly seen in Fig. 9 (b) and Table I, the MIL-STD-461G 50 µH AMN with the open-circuit source impedance condition stipulated by the standard gives maximum voltage deviations of +3.32 dB / -2.94 dB when the annular tolerance is ± 20 % for the magnitude and ±11o for the phase in accordance with CISPR 16-4-2, which can be accounted reasonable performance in the frequency range of 9 kHz – 400 MHz. Also, the calculated magnitude curve in Fig. 9 (a) is the same as the military standard curve shown in Fig. 3, as expected. On the other hand, its analysis with the short-circuit source impedance condition yields a dramatic change in the impedance phase and voltage deviation as shown in Fig. 10 but it gives an impedance magnitude curve akin to the opencircuit source impedance condition, which means the MILSTD-461 50 µH AMN requires vigilant use in low-impedance source conditions below 50 kHz and the AMNs employed in CE102 tests should have an impedance tolerance as low as possible in the low frequency range. It shows very clearly how important the specification of the phase is. For the open condition (Fig. 9) the phase is limited to approximately 55° and the voltage deviation is acceptable. For short condition (Fig. 10) the phase comes close to 90° where a resonance occurs and the voltage deviation goes up. When we have a look at Fig. 11 and Fig. 12, it is noticed that the MIL-STD-461 5 µH AMN with both the open-circuit and short-circuit source impedance conditions appears to be not usable in the frequency range of 9 kHz – 1 MHz considering the CISPR 164-2 annular tolerance although the standard presents its usable impedance curve as from 150 kHz, on which careful consideration is needed. This signifies that there is great likelihood of potential high uncertainty in the use of 5 µH AMNs and also low annular impedance tolerance is required in the range of 9 kHz – 1 MHz. This information is crucial because EMC laboratories sometimes have to use 5 µH AMNs in the frequency range starting from 10 kHz despite the standard graph that starts at 150 kHz, on the grounds that some EUTs are not operating properly with 50 µH AMNs. Regarding the CISPR AMNs, the combination of CISPR 50 Ω / 50 μH + 5 Ω & CISPR 50 Ω / 50 μH AMNs is very usable from 9 kHz as given in Fig. 13 based on the aforementioned annular tolerance and its magnitude & phase results are completely compatible with the standard tables given in CISPR16-1-2. Additionally, the resemblance of the CISPR AMN curves in Fig. 13 to the curves of the MIL-STD-461 AMN (Rs: open-circuit) in Fig. 9 is striking. It should be noted that the CISPR AMNs analyzed in this paper are ideal AMNs in their simplest form as excerpted from the standard (see Fig. 5). Ultimately, ISO 7637-2 and ESA ECSS-E-ST-20-07C AMNs show their intrinsic behavior in magnitude, phase and δZAN curves as seen in Fig. 14 – Fig. 17 for both the shortcircuit and open-circuit source impedance conditions. The ESA ECSS-E-ST-20-07C AMN seems to show the worst performance and become usable beyond 2 MHz for both of the source impedance conditions as seen in Fig. 16 – Fig. 17 based on the same annular tolerance given in CISPR 16-4-2. Nevertheless, it should be stressed that the ECSS AMN is not intended to measure the disturbance voltage and the disturbance current is measured with a current probe in the standard. Here, an answer is sought to the question “what happens if an ECSS AMN is used in conducted emission tests?”. As a result, it looks that the ECSS AMN is not suitable at all for the use in conducted emission tests considering the CISPR 16-4-2 tolerances. When we examine the ISO 7637-2 AMN results in Fig. 14 – Fig. 15, we see at the first glance that there is significant difference between the short-circuit and open-circuit source impedance condition results. The impedance magnitude curve for the open-circuit source impedance is utterly incompatible with the curve given in the standard whereas there is good agreement with the standard for the short-circuit power source case as the ISO 7637-2 requires the source side is short-circuited in calibration. However, the voltage deviation is remarkably high below 1 MHz for the short-circuit source impedance condition while the severity of the deviation is alleviated for the open-circuit source impedance condition and seems to be usable beyond 50 kHz in terms of voltage deviation. In fact, for ISO 7637-2 the magnitude tolerance is only 10 %, but the 20 % / 11.5° annular tolerance from CISPR is implemented here in this paper just for information and for making it comparable to the other analysed AMNs. When the overall results are evaluated together, the best performance is shown by the ANSI C63.4 AMN independently of source impedance conditions in the entire frequency range in accordance with Table I based on the annular tolerance of CISPR 16-4-2. (a) (b) Fig. 9. MIL-STD-461 50 µH AMN (RS : Open) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN based on the annular tolerance for UCISPR calculation as per CISPR 16-4-2. (a) (b) Fig. 10. MIL-STD-461 50 µH AMN (RS : Short) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN based on the annular tolerance for UCISPR calculation as per CISPR 16-4-2. (a) (b) Fig. 11. MIL-STD-461 5 µH AMN (RS : open-circuit) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN based on the annular tolerance for UCISPR calculation as per CISPR 16-4-2. (a) (b) Fig. 12. MIL-STD-461 5 µH AMN (RS : short-circuit) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN based on the annular tolerance for UCISPR calculation as per CISPR 16-4-2. (a) (b) Fig. 13. Combination of CISPR 16-1-2 50 Ω / 50 µH + 5 Ω (9 kHz – 150 kHz) & CISPR 16-1-2 50 Ω / 50 µH (150 kHz – 400 MHz) AMNs (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN for UCISPR calculation based on the annular tolerance as per CISPR 16-4-2. (a) (b) Fig. 14. ISO 7637-2 AMN (RS : open-circuit) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN for UCISPR calculation based on the annular tolerance as per CISPR 16-4-2. (a) (b) Fig. 15. ISO 7637-2 AMN (RS : short-circuit) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN for UCISPR calculation based on the annular tolerance as per CISPR 16-4-2. (a) (b) Fig. 16. ESA ECSS-E-ST-20-07C AMN (RS : open-circuit) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN for UCISPR calculation based on the annular tolerance as per CISPR 16-4-2. (a) (b) Fig. 17. ESA ECSS-E-ST-20-07C AMN (RS : short-circuit) (a) nominal magnitude and phase information, (b) maximum and minimum deviation δZAN for UCISPR calculation based on the annular tolerance as per CISPR 16-4-2. Table I. Nominal AMN uncertainty component δZAN for UCISPR calculation by AMN types based on the annular tolerance as per CISPR 16-4-2 in the range of 9 kHz – 400 MHz and specifically in the relevant AMN standard ranges. AMN Type CISPR 16-4-2 δZAN (dB) (9 kHz – 400 MHz) CISPR 16-4-2 δZAN (dB) (AMN’s standard frequency range) CISPR 50 Ω / 50 μH + 5 Ω (9 kHz – 150 kHz) & CISPR 50 Ω / 50 μH (150 kHz – 400 MHz) 3.59 / -3.07 3.59 / -3.07 (9 kHz – 30 MHz) MIL-STD-461 50 µH (RS : open-circuit) 3.32 / -2.94 3.32 / -2.94 (10 kHz – 10 MHz) MIL-STD-461 50 µH (RS : short-circuit) 42.18 / -20.06 42.18 / -20.06 (10 kHz – 10 MHz) MIL-STD-461 5 µH (RS : open-circuit) 45.53 / -20.80 45.53 / -12.31 (150 kHz – 30 MHz) MIL-STD-461 5 µH (Rs : short-circuit) 44.29 / -18.78 42.82 / -10.62 (150 kHz – 30 MHz) ISO 7637-2 (RS : open-circuit) 11.31 / -5.24 1.70 / -2.01 (100 kHz – 100 MHz) ISO 7637-2 (Rs : short-circuit) 53.07 / -18.76 53.07 / -13.53 (100 kHz – 100 MHz) CISPR 25 (Rs : open-circuit) 45.42 / -21.06 45.42 / -16.02 (100 kHz – 100 MHz) CISPR 25 (Rs : short-circuit) 49.99 / -18.89 41.48 / -13.31 (100 kHz – 100 MHz) DO-160G (RS : open-circuit) 46.35 / -26.02 46.35 / -26.02 (10 kHz – 400 MHz) DO-160G 49.99 / -18.89 49.99 / -18.89 (Rs : short-circuit) (10 kHz – 400 MHz) ANSI C63.4 (RS : open-circuit) 3.33 / -2.94 3.33 / -2.94 (9 kHz – 30 MHz) ANSI C63.4 (Rs : short-circuit) 3.31 / -2.93 3.31 / -2.93 (9 kHz – 30 MHz) ESA ECSS-E-ST-20-07C (RS : open-circuit) 49.87 / -12.29 49.87 / -10.33 (100 kHz – 100 MHz) ESA ECSS-E-ST-20-07C (Rs : short-circuit) 48.15 / -10.88 48.15 / -10.88 (100 kHz – 100 MHz) Besides, although the voltage deviations seem to be very high and unacceptable in both the two and third columns of Table I for some AMNs, this should be reiterated here that all the analysis in Table I is based on the 20 % / 11.5° annular tolerance of CISPR16-4-2 irrelevant of the related standards of the analyzed AMNs, which means that the uncertainty contribution arising from these AMNs can be reduced by choosing AMNs with a lower impedance tolerance for conducted emission tests. When a calibration report for an AMN is received from a calibration laboratory, it can be checked through the developed software. However, as studied above, the way of the AMN source side termination could have a significant effect on the AMN impedance, especially in the low frequency range and users should be cautious about it. The actual calibration data of our selected military 50µH and 5µH AMNs obtained from a calibration laboratory in the range of 10 kHz – 30 MHz is given in Fig. 18 (a) – Fig. 19 (a). As we already know the nominal impedance magnitude & phase information from Fig. 9 (a) and 11 (a) for the open-circuit source impedance condition, the actual AMN uncertainty components are easily produced in the entire CE102 test frequency range 10 kHz – 10 MHz as given in Fig. 18 (b) – Fig. 19 (b) and Table II to be directly used in the CE102 uncertainty budget table. As our military AMNs were calibrated by the calibration laboratory up to 30 MHz beyond the CE102 standard frequency range, we also extended the frequency range in the graphs to 30 MHz for further information. As it is clearly seen in Fig. 19 (b), we should be very careful while using the 5 µH AMN in CE102 tests as uncertainty contribution is severe in the range of 10 kHz – 300 kHz whereas the 50 µH AMN contribution presented in Fig.18 (b) is fairly low almost in the full frequency range and it is even less than the δZAN of UCISPR given in Fig. 9 (b). (a) (b) Fig. 18. Results of our military 50 µH AMN (RS : open-circuit) (a) actual magnitude and phase information from a calibration laboratory, (b) actual uncertainty component (δZAMN) based on the actual AMN calibration data to be used in military CE102 uncertainty budget. (a) (b) Fig. 19. Results of our military 5 µH AMN (RS : open-circuit) (a) actual magnitude and phase information from a calibration laboratory, (b) actual uncertainty component (δZAMN) based on the actual AMN calibration data to be used in military CE102 uncertainty budget. Table II. Summary of military AMN Results in the range of 10 kHz – 30 MHz AMN Type Actual Uncertainty Component for Military CE102 Testing δZAMN (dB) (10 kHz – 30 MHz) MIL-STD-461G 50 µH 1.14 / -1.5 MIL-STD-461G 5 µH 24.63 / -30.48 (a) (b) Fig. 20. (a) Uncertainty component (δZAMN) of the AMN simulation with phase information increased by 50o, (b) re-analysis of MIL-STD-461 50 µH AMN (RS : short-circuit) for voltage deviation with 1 % / 0.57o annular magnitude and phase impedance tolerance. Ultimately, the results of the case where the phase deviation is intentionally increased by 50o in a military 50 µH AMN simulation are given in Fig. 20 (a). It is observed in Fig. 20 that significant deviation in the phase leads to unacceptable and intolerable voltage deviation, which is already explained in CISPR 16-1-2 Annex I. For that reason, the phase requirement should be also integrated into future MIL-STD461 versions to prevent the situation from going unchecked. Lastly, Fig. 20 (b) shows the improvement in the voltage deviation by means of the reduction in the annular impedance tolerance. When Fig.10 (b) and Fig. 20 (b) are compared, a marked improvement is noticed in the low frequency range, which underlines the importance of having an annular impedance tolerance as low as possible for a military AMN especially in the frequency range close to 10 kHz. The software developed in the scope of the research is downloadable through the link given in [14]. IV. CONCLUSION In this paper, we developed and introduced a user-friendly and useful software solution which will allow laboratories to calculate their own AMN uncertainty contribution (δZAMN) for any desired AMN type based on its actual LISN calibration certificate and we have made it downloadable. The software also can calculate the nominal AMN uncertainty component (δZAN) which is used to calculate UCISPR as per CISPR 16-4-2 for any AMN type for a standard annular tolerance or a customized one. The analysis made by using the developed software clearly demonstrated how essential the specification of the phase is and consequently all conducted emission related standards which use AMNs such as MIL-STD 461 should specify a phase tolerance to avoid going unchecked in terms of measurement uncertainty. Even this definition should be frequency-dependent as much as much as possible because much lower tolerance might be required in low frequencies for a suppressed uncertainty contribution. Ultimately, it was shown that both of the ways of the AMN source side termination, namely the open and short circuit conditions, should be particularly defined for the LISN impedance as significant difference might occur between them as reported in this paper and users should be aware of the risks even if the related standards define the impedance and/or phase requirements only for one of them. ACKNOWLEDGMENT This research is performed in the scope of the project “21NRM06 EMC-STD Metrology for emerging electromagnetic compatibility standards”. The project (21NRM06 EMC-STD) has received funding from the European Partnership on Metrology, co-financed by the European Union’s Horizon Europe Research and Innovation Programme and by the Participating States. REFERENCES [1] MIL-STD-461G, “Requirements for the Control of Electromagnetic Interference Characteristics of Subsystems And Equipment”. [2] IEC CISPR 16-1-2 : 2017, "Specification for radio disturbance and immunity measuring apparatus and methods – Part 1-2: Radio disturbance and immunity measuring apparatus – Coupling devices for conducted disturbance measurements". [3] IEC CISPR 16-4-2 : 2018, “Specification For Radio Disturbance And Immunity Measuring Apparatus And Methods - Part 4-2: Uncertainties, Statistics And Limit Modelling - Measurement Instrumentation Uncertainty”. [4] A. Kriz, "Uncertainty calculation for AMN impedance contribution using the Monte Carlo Method", 2019 IEEE Int. Symp. on EMC New Orleans (LO), pp. 565-569, Jul. 22-36, 2019. [5] Carlo F. M. Carobbi, “Quantification of the artificial mains network impedance contribution to the uncertainty of conducted emission measurements”, 2020 XXXIIIrd General Assembly and Scientific Symposium of the International Union of Radio Science, 29 August 2020 - 05 September 2020, Rome, Italy. [6] Application Note and File, Impedance Uncertainty Contribution of Artificial Networks (AN, AMN and ISN) & Maximum Deviation Calculator, Rohde & Schwarz, 2010. [7] IEC CISPR 16-4-1 : 2009, “Specification for radio disturbance and immunity measuring apparatus and methods - Part 4-1: Uncertainties, statistics and limit modelling - Uncertainties in standardized EMC tests”. [8] Manfred Stecher, “Uncertainty in RF Disturbance Measurements: Revision of CISPR 16-4-2”, EMC’09/Kyoto, 23R1-1. [9] IEEE/ANSI C63.4-2014, “American National Standard For Methods Of Measurement Of Radio-Noise Emissions From Low-Voltage Electrical And Electronic Equipment In The Range Of 9 KHz To 40 GHz”. [10] CISPR 25:2021, “Vehicles, boats and internal combustion engines - Radio disturbance characteristics - Limits and methods of measurement for the protection of on-board receivers”. [11] RTCA DO-160G, "Environmental Conditions and Test Procedures for Airborne Equipment". [12] ISO 7637-2:2011, “Road vehicles — Electrical disturbances from conduction and coupling - Part 2: Electrical transient conduction along supply lines only”. [13] ESA ECSS-E-ST-20-07C Rev. 1, “Space Engineering Electromagnetic Compatibility”. [14] https://tubitakgovtrmy.sharepoint.com/:f:/g/personal/serdar_buyuk_tubitak_gov_tr/EhYuB4 Ih2lFEq8ZMHpk8lbwBM0ZT87nkf62uGP7saD0uLg?e=uL0ENa