Electromagnetic Emission Measurement Determination of the Test Configuration of Telecommunications Power Rectifier Systems

Transcription

1 Electromagnetic Emission Measurement Determination Marcus Barthus Azevedo NMi Brasil /! " mpc.com.br Victor Vellano Neto CPqD Telecom. Research and Dev. Center /! " cpqd.com.br José Claudio de Oliveira e Silva Emerson Energy Systems /! " emersonenergy.com Mituo Okamoto Saft Power Systems /! " saftnife.com.br Abstract - This paper presents data of investigation of several experiments on the determination of the Minimum Representative System (MRS) for electromagnetic emission of large, which are by nature populated with identical rectifier modules. Two systems were used as study objects. The work is related to standardization activities in Brazil and trigged by the fact that most of the EMC test sites are not able to provide enough a.c. current for testing large power systems. The question of the validity of testing EUT s populated with small number of rectifiers within the recommendations of the emission standards has arisen in the standardization commission. The problem of determining a MRS is known - it can be very time consuming and therefore costly - so for the commission this work is in essence a feasibility study. The four parties, CPqD, NMi Brasil, EES and SAFT Power have conducted some tests and studies aiming to provide the commission with solid basis for adopting or not such practice. 1 Introduction In the last few years, Brazil significantly extended the development of technical specifications of energy systems and, in special, high frequency rectifying units and systems. Among the new requirements are those related to Electromagnetic Compatibility (EMC) and in there, the electromagnetic emission that was initially linked to the protection of the radiofrequency spectrum used for broadcast telecommunication services and that now assumes a more and more important role in complex information technology (ITE) installations. In general, an electromagnetic emission requirement must define the emission limits and testing procedures. The Brazilian Standard NBR [2], one of the basic references adopted in the standard for rectifying systems for telecommunications, comprises most of the items needed, e.g. it establishes limits, defines appropriate test sites and test equipment, test procedures and a general definition for the configuration of the Equipment Under Test (EUT). As any general definition, it lacks details, in our case, on how to define representative configurations of large power systems, what may give rise to a uncontrolled variety of interpretations and technical judgements, thus jeopardizing the objective of the standard, interfering with the ability of compare test results and on the settlement of a common set of rules. In addition to the inherent difficulty associated with the search for the Minimum Representative System (MRS), large power systems impose another test difficulty, that is the high a.c. current demanded from the test house facility. Trying to solve that, the Brazilian standardization commission analyzed several other international publications, among which are those generated by: CISPR 1

2 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto (Comité International Special des Perturbations Radioelèctrique), ETSI (European Telecommunication Standardization Institute), CENELEC (Comité Européen de Normalisation Electrotechnique), and ITU-T (International Telecommunication Union) - Group V, in charge of the recommendations related to electrical protection and EMC. The relevance of Group V is that it focuses on those requirement applicable to telecommunication equipments, taking into account their electromagnetic environment as well as functional aspects, etc.. Therefore, they actually complement the general and basic requirements published by IEC/CISPR. Among the documents analyzed on the subject, the most relevant are listed as references [1], [2], [3], [4], [5], [6] and [7]. CISPR publications are generally considered basic recommendations. CISPR 22 together with the ITU-T k38 and ETS - EN are the main references for the definition of applicable tests and procedures of large systems adopted by the telecommunications sector in Brazil. 1.1 Electromagnetic Emissions Requirements The NBR 12304, based on the first edition of CISPR22, establishes emission limits and test procedures for ITE, and is adopted for rectifying systems for telecommunications. The third edition of CISPR Publication 22 available today, incorporates details of measurements of telecommunications lines, validation of test site and test distances, and maintain the core and principles prescribed in NBR It is important to understand that, even though a general purpose power system does not fall into the scope of CISPR 22, it is adopted when it turns to be a system placed in the same electromagnetic environment as telecommunication equipments, where emissions levels allowed by CISPR 11 may be considered too high. On those standards, Class B is defined for residential environment equipment or equipment connected to the low voltage public network, where the emission levels must be lower due to the proximity to radiocommunications, receivers and other communications systems. Class A allows a higher level of emissions and therefore can cause interference in residential environment, being appropriate for those equipment installed in business or light industrial environment, where they are not connected to the low voltage public network and the distance to radiocommunications, receivers and other communications systems is at least 30 meters Radiated and Conducted Emissions Limits Table 1 and Table 2 present the limit lines for conducted and radiated electromagnetic disturbances respectively, valid for both NBR and CISPR 22. CONDUCTED (150 khz to 30 MHz) RADIATED (30 MHz to 1 GHz) Radiofrequency Disturbance Limits [db(µv)] Quasi-Peak Limit Frequency Frequency Range Class A Class B Range [db(µv/m)] [MHz] Quasi- Quasi- Average Average [MHz] Class A Class B Peak Peak 0,15 ~ 0, ~ ~ ~ ,50 a ~ a Table 1 CISPR 22 Conducted Emission Limits Table 2 CISPR 22 Radiated Emission Limits Note: It is important to understand that according to the class definitions in 1.1, equipment to be installed on telecommunications centers are generally considered Class A, and equipment as outdoor cabinets or those installed in subscriber premises must be considered Class B. 2

3 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto 1.2 General Disturbance Measurement Test Details CISPR 22 as well as most of the emissions standards require that the EUT configuration and exercising are such that the maximum emissions are found and recorded to be compared with the limits. For Power Rectifiers used in Telecommunications, that includes but is not limited to test under all possible output loads and all possible rated input voltages for all typical configurations, thus recommending an unaffordable number of measurements. CISPR 22 establishes limits, defines appropriate test sites and test equipment and test procedures. In special, it specifies that conducted emission are measured using an Artificial Mains Network (AMN) 1 because it provides the necessary decoupling of the measured noise from the a.c. voltage and provides a well defined impedance (refer to Figure 1) to the EUT as a noise source. In cases where the current is too high (say above 100 A, as it is usually the case when measuring conducted disturbances on d.c. terminals) so that a AMN connection is not possible, the Brazilian committee adopted the voltage probe method (refer to Figure 2) defined in [1]. Rede de Alimentação ESE 50 µ H 250µ H Rede de Alimentação C X C < 1500Ω 0,25 µ F 8µ F 1,2µ F ( R) Ω Terminação de 50 - Ω medidor de perturbação 1000 Ω 5 Ω 10 Ω X 1 >R R Instrumento de medição Figure 1 Typical AMN schematic Figure 2 Typical Voltage Probe All other specifications remains according to CISPR Configuration of the EUT The definition of the test configuration is as important as the limit settled by a standard. As mentioned before, [2] and [3] generally defines the configuration of the Equipment Under Test (EUT), allowing different interpretations and technical judgements. To solve that, other specific references related to large equipment emission measurement from ITU-T and ETSI was brought to light. It is important to mention that the recommendations of test configurations may not be feasible or viable since they may require too many rectifying units and unachievable load requirements. CISPR publications recommends: The EUT shall be configured, operated and its cabling arranged in a way that is consistent with typical applications and that the maximum emissions be generated; Where not specified, the EUT must be configured, installed, arranged and operated in a manner consistent with typical applications. Interface cables/loads/devices shall be connected to at least one of each type of interface port of the EUT, and where practical, each cable shall be terminated in a device typical of actual usage; Where there are multiple interface ports of the same type, additional interconnecting cables/loads/ devices may have to be added to the EUT depending upon the results of preliminary tests. The number of additional cables should be limited to the condition where the addition of another cable does not decrease the margin a significant amount (for example 2 db) with respect to the limit. The rational for the selection of the configuration and loading of ports shall be included in the test report; Where there are multiple interface ports all of the same type, connecting a cable to just one of that type of port is sufficient, provided it can be shown that the additional cables would not significantly affect the results. 1 AMN are also called Line Impedance Stabilization Network (LISN) 3

4 In addition to that, [7] defines a minimum representative system: 4 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto The minimum representative system is a system which contains the minimum number of units needed to perform all functions specified for the system.; The original text of [7] is given in [8], where a procedure for the determination of the minimum representative system is established: Physically large systems are modular in nature, i.e. they will generally be increased in size (and operational function) by the addition of like units. To ensure that the EUT is representative of installed systems, in terms of function and electromagnetic radiation, tests shall be repeated with the EUT configured with additional units. If, by adding additional units, which generates synchronized noise, the emission levels do not increase 4 db or more above the original maximum measured values, independent of frequency, then a minimum representative system in terms of radiation has been achieved. The additional units shall be composed to the largest possible extent of highest radiation sources, but they shall be typical of realistic installation. If, with the increase of additional identical units, the radiation increases by more than 4 db, then further additional units shall be added until the increase is less than 4 db. After the addition of identical units (as shown in figure 2 of [8]), the measured field levels from the representative system shall not increase beyond compliance limits. Another useful source of information is the EMC requirements for UPS [4]: The measurement shall be made in the operating mode producing the largest emission in the frequency band being investigated consistent with normal applications. UPS operating modes, normal mode and stored energy mode shall be investigated covered; An attempt should be made to maximize the emission by varying the test setup configuration of the test sample; UPS AC outputs shall be loaded with linear load capable of exercising the EUT for any load condition within its output rating. 2 Proposal The technical commission responsible for the Brazilian draft standard for switched rectifiers and power systems up to 2400 A d.c. (project number 03: ) has considered the introduction of a procedure to determine the MRS in emission tests. The main problem resides in testing large power systems. The flowchart in Figure 3 has been proposed in the draft standard. If the maximum output power that can be drawn from the EUT is lower than 50 % of its maximum capacity, no matter whether the MRS was reached or not, the CISPR 22 limits are reduced as indicated in the flowchart. The proposal implies then on a penalty to the manufacturer, driving then to force test sites to increase their current capacities. It may in fact happens as one realize that the emission limits can be tougher if tested in a test house with lower ampacity. In the other side, even risking to modify a standard, adding a penalty for the extrapolation of results of a partial test is important to: Have a fixed and reasonable criteria, based on the equipment itself; Use the test resources available to judge the performance of the EUT and still have a good correlation to the reality, without opening a escape channel the to norm; In line with the main objective of the standard, help protecting the electromagnetic environment in an efficient manner.

5 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto FLOWCHART - Determination of minimum representative (power) system for electromagnetic emission test. Nr = Nmax can be tested? y n Nr = 1 or Minimum Check & Record disturbances with margin to the limit < 10 db Nr = number of rectifiers composing the system under test (rectifiers at full power) Nmax = maximum number of rectifiers that can equip the system under test System under test = a power system comprising all functional parts that can equip the system, e.g. rectifiers, DC distribution and supervision units, in a configuration consistent with practical applications. If the system can grow by adding more cabinets, then at least two cabinets shall be tested. If more than two cabinets are required to permit that all functional parts be included, then a sufficient number of cabinets shall be tested so as to include all functional parts. According to CISPR22, all ports shall be connected and exercised. Nr = Nr+1 > + 20 db? Nmax n Nr Nmax/2? y y y Test lab capable of powering one more rectifier? n MRS Minimum Representative System n (corrected limit) L - 10 log (Nmax / Nr) L = specified emission limit as per CISPR 22 RECORD - Final Measurements Figure 3 Proposed flowchart for determination of minimum representative system (MRS) for EMI test. 2.3 Test Samples The two test samples used for the purposes of this study are defined in Table 3. EUT 1 (EES) EUT 2 (Saft) Manufacturer Emerson Sistemas de Energia Ltda. Saft Power Systems Ltda. Model Sophia Plus 24V / 25,5kW SR 1200 / +24V / Rectifiers 15 x 1700W ( V) 12 x 100 A / +24 V AC Input Nominal Voltage 220V, 3 phase 220 V, 3 phase 5

6 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto Frequency 50 60Hz 45 Hz a 65 Hz DC Output Nominal Voltage 24V 24 V Current 950 A (@ 27 V) 1200 A Photo for reference only Table 3 Test sample basic specifications 2.4 Test Conditions and EUT Configurations In general lines the system configuration started with a small number of rectifiers, three for EUT 1 and just one for EUT 2. The system grew by the addition of rectifiers, one by one, according to the specified or usual manner these systems grow, so that every step constituted a tested configuration. Preliminary emission measurements (fixed azimuth) were performed for each configuration. The maximum emission direction (azimuth and antenna height) was searched in some of the tested configurations. The system output current was set to a value corresponding to the maximum power that could be provided by the number of installed rectifiers. Rectifier positions in the cabinets as per Figure 4 and Figure Figure 4 - Rectifier positions of EES, Sophia+ (EUT1) Figure 5 - Rectifier positions of Saft, SR 1200 (EUT2) 6

7 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto 2.5 Emission Test Procedure Two types of radiated emissions measurement was performed, one that is basically a preliminary test, consisting on recording the measurement on an fixed antenna height and azimuth, used only to compare data, and the standard emissions measurement described on 2.5.1, used for final measurement according to the standard. The conducted emissions test procedure is much simpler, being the only remark to the standards that the 3 phases were always maximized (peak hold) and all data represents the maximum reading. Unless otherwise specified no reference to phase is noted Standard Radiated Emission Maximization Procedure h = antenna height; θ = azimuth; h max = h respective to maximum emission θ max = θ respective to maximum emission For θ = 0, a preview of the emission is done (just visual monitoring on the receiver screen) while h changes from 1 to 4 m, determining h max for each antenna With the antenna at h max, the radiated emission is acquired in peak hold mode so that the maximum emission values are accumulated and recorded while θ changes from 0 to 360, for both frequency ranges: MHz and MHz.. The emission values recorded are not at a given fixed θ, so θ max is not determined at this point. For the highest emission frequencies and/or for those close to the emission limit (margin < 6 db), a fine search of maximum emission is performed with respect to h and θ. The receiver is set to zero-span and tuned at the highest emission frequency(ies) for h as well as for θ scanning. The sweep times (in the order of seconds) are adjusted so as to match the time required for h to move from 1 to 4 m and then for θ to change from 0 to 360. Where θ max is achieved, a final antenna scan is performed, so that a very clear indication of h max and θ max is obtained on the receiver screen. The procedure is repeated for horizontal polarization. h max and θ max are then well determined for each particular selected high emission frequency. 2.6 Investigation Test Results The data that follows is also a result of the learning process taken during the test. It was planned that the data analysis would be performed with all peak measurement and it was noticed, during the process of testing EUT 1, that no coherence was achievable with that approach, requiring the test to be performed with Quasi-Peak measurement. Since Quasi-Peak measurement is time consuming, decision was taken to perform only on the highest emissions. Tests performed on EUT 2 already took the EUT 1 findings and all data was planned for Quasi-Peak Radiated Emissions Number of Measured Frequencies [MHz] Quasi-Peak Limit Rectifiers Quasi-Peak Peak 1) (N R ) CISPR22 - A Corrected ,8 31,2 33,6 22,1 23,8 23,8 19,5 19,4 19,3 23, ,0 4 35,5 30,6 34,0 18,3 23,4 22,7 18,2 18,3 21,4 22, ,3 5 35,6 31,4 34,2 23,1 22,2 23,8 19,7 18,5 20,3 23, ,2 6 35,0 31,6 32,6 21,2 22,1 23,7 18,9 21,3 20,8 25, ,0 7 34,2 31,6 32,1 22,8 22,3 23, ,6 19,2 24, ,7 7

8 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto 8 35,1 31,2 33, ,2 28,3 24,6 31,6 26,1 28, ,3 9 22,3 24,8 26,4 22,3 24,3 22,7 26, , ,7 32,1 34, ,2 28, ,3 28,5 26, ,0 1) Limit corrected as proposed to 10 log (N R / Total of Rectifiers of the System under Test) Table 4 EUT 1 Highest radiated emissions, vertical polarization Number of Measured Frequencies [MHz] (Quasi-Peak Only) Quasi-Peak Limit Rectifiers (N R ) CISPR22 - A Corrected 1) 1 33,7 34,3 28, ,2 2 33,8 34,2 28,8 26, , ,4 29, ,0 4 35,6 34,5 33, ,2 5 34,5 35,8 33,9 27, ,2 6 35,1 34, , ,0 7 35,6 34,6 31,9 28, ,7 8 35,2 34,8 32,8 28, ,2 9 35,7 35,6 34,1 27, , ,7 35,9 34,8 27, ,0 1) Limit corrected as proposed to 10 log (N R / Total of Rectifiers of the System under Test) Table 5 EUT 2 Highest radiated emissions, vertical polarization EUT 1 - Radiated Emissions (Vertical Polarization) - Peak and Quasi-Peak Measurements 42 QP LIM Cispr22 A 37 Corrected Lim 129 MHz emission (dbuv/m) MHz 149 MHz 152 MHz 156 MHz 163 MHz 176 MHz MHz Q-Pk 90 MHz Q-Pk 96 MHz Q-Pk number of Rectifying Units Graphic 1 EUT 1 Highest radiated emissions, vertical polarization 8

9 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto 41 EUT 2 - Radiated Emissions (Vertical Polarization) - Quasi-Peak Measurements emission (dbµv) number of Rectifying Units QP LIM CISPR 22 A Limite Corrigido 30 MHz 31 MHz 32 MHz 228 MHz Graphic 2 EUT 2 Highest radiated emissions, vertical polarization Conducted Emissions Number of Measured Frequencies [MHz] (Peak Only) Average Limit Rectifiers (N R ) 0,15 0,16 0,55 CISPR22 - A Corrected 1) 3 54,6 54,5 49, ,0 4 57,9 57,8 47, ,3 5 57,9 57,3 46, ,2 6 59,3 59,1 44, ,0 7 57,7 58,3 42, , , ,3 9 54, , , , , ,0 1) Limit corrected as proposed to 10 log (N R / Total of Rectifiers of the System under Test) Table 6 EUT 1 Highest conducted emissions Number of Measured Frequencies [MHz] (Peak Only) Average Limit Rectifiers (N R ) 0,83 0,86 0,92 5,57 7,42 CISPR22 - A Corrected 1) 1 49,3 44,6 48, ,2 2 54, , ,2 3 55,5 52,7 57, ,0 4 55,4 53,1 54,1 53,6 59, ,2 5 55,4 56,3 54,8 51,7 57, ,2 6 55,2 55, ,6 55, ,0 7 56, , , ,7 8 54,8 55,4 53,7 45, ,2 9 56,9 55, , ,8 9

10 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto 12 55,8 55,8 40,8 50, ,0 1) Limit corrected as proposed to 10 log (N R / Total of Rectifiers of the System under Test) Table 7 EUT 2 Highest conducted emissions EUT 1 - Conducted Emissions - Peak Measurements Avg LIM Cispr 22 A emission (dbuv) Corrected Lim 0,1573 MHz 0,1615 MHz 0,55 MHz number of Rectifying Units Graphic 3 EUT 1 Highest conducted emissions 62 EUT 2 - Conducted Emissions - Average Measurements 58 emission (dbµv) Avg LIM Cispr 22A Corrected LIM 0,83 MHz 0,86 MHz 0,92 MHz 42 5,57 MHz 7,42 MHz number of Rectifying Units Graphic 4 EUT 2 Highest conducted emissions 2.7 Final Test Results The final test graphic results corresponding to the investigation data (vertical polarization / maximum emission only) are added below to help identifying each EUT radiated and conducted emission pattern. 10

11 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto EUT 1 - Radiated and Conducted Test Results Graphic 5 EUT 1 Final Conducted EMI Graphic 6 EUT1 Final Radiated EMI, Vert. Pol EUT 2 - Radiated and Conducted Test Results Graphic 7 EUT 2 Final Conducted EMI Graphic 8 EUT2 Final Radiated EMI, Vert. Pol. 11

12 3 Data Analysis and Discussion J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto The most relevant findings from the investigation data and from the measuring experiences for those two study objects, were: The reduction of the limit has to be reconsidered carefully. The two systems passed all tested configurations but would fail if a small number of rectifiers were tested according to the flowchart in Figure 3, as Graphic 1, Graphic 2 and Graphic 4 shows; To really take advantage of the flowchart, it would be recommended that the emission from the rectifiers and other non-modular sources, like supervision unit, should be separated in the analysis; The MRS can not be easily found on the receiver screen during the execution of the tests. It was necessary to tabulate the emission of the frequencies within 10 db margin to the limit and plot them so as to get a better view of the emission increase as the number of rectifier increases. A trend analysis may be helpful. In other words, a rather elaborate assessment of data is required along the measuring process in order to give directions for the continuation of the tests, turning the test longer than an ordinary emissions test. Quasi-peak measurements provide better-behaved results and therefore may be necessary if the MRS is searched for. It also means longer tests. The behavior of the emissions does not follows a trend line like it was predicted by the authors of this paper. It is in fact highly unpredictable, sometimes even leading to the exit of the flowchart on a wrong point (points where much higher emissions not yet appeared); There was only one pattern repeated on the two sample tested: the results with 50 % and 100 % population were somewhat coherent, with differences within the expected values (< 3 db increase from 50 % to 100 %). No new or unexpected frequency appeared from one to the other. The change in the direction of maximum emission as the modules are introduced was not studied in this work. Such effect can make the data book impossible to analyze. The maximization of the emission as modules are introduced is very time consuming and can by no means be adopted at each step of the MRS search in ordinary tests. 4 Conclusion It must be realized that testing large power systems at their maximum configuration may be prohibitive sometimes and that testing poorly equipped systems is not reasonable and may be risky considering that, besides running the risk of spectrum pollution, the test may be rejected by one or another certification agency or customer. Searching for MRS as proposed is time consuming and may lead to wrong solution. The authors of this paper recommend that: a) Emissions test be performed on a System defined as follows: System under test = a power system comprising all functional parts that can equip the system, e.g. rectifiers, DC distribution and supervision units, in a configuration consistent with practical applications. If the system can grow by adding more cabinets, then at least two cabinets shall be tested in order to include the interconnection cable. If more than two cabinets are required to permit that all functional parts be included, then a sufficient number of cabinets shall be tested so as to include all functional parts. b) The above system must be tested fully and halfway populated, at full load and other conditions verified. That is the only way to guarantee full compliance to the standards. c) When the above is not possible, the test should proceed as close as possible to that balancing between the two extreme (population-wise) test conditions, not being recommended that less than half a system is tested. 12

13 J. C. Oliveira e Silva, M. Barthus Azevedo, M. Okamoto, V. Vellano Neto d) If after all, the test can only be performed on systems as defined above less populated then 50%, a penalty as the one suggested on the flowchart of Figure 3 be adopted. 5 References [1] CISPR 11 : 1997 Limits and methods of measurement of radio disturbance characteristics of ISM equipment [2] NBR : 1992 Limites e Métodos de medição de características radioperturbação de Equipamento de Tecnologia da Informação. [3] CISPR 22 : Limits and methods of measurement of radio disturbance characteristics of information technology equipment [4] EN : 1995 Uninterruptible power systems (UPS) Part 2: EMC requirements [5] ITU K48 Product Family EMC Requirements For Each Telecommunication Network Equipment [6] ITU-T k38 Radiated emission test procedure for physically large systems [7] Draft EN : 1998 Electromagnetic compatibility and Radio spectrum Matters (ERM) - Radiated emission testing of physically large telecommunication [8] ETS : 1994 Equipment Engineering (EE) - Radiated emission testing of physically large telecommunication systems 13