IEC 61000-3-2amendmenta14explained[资料]

Harmonic Analysis perCENELEC Amendment -14to EN/IEC61000-3-2IntroductionWorking Group no.1 of Technical Committee 77 ( IEC TC77/SC77A/WG-1 ) agreed on the text of the so-called Millennium Amendment to EN/IEC61000-3-2 [1] in a special meeting, held in December 1999 in Rome,Italy. This special meeting was called after a basic compromise was reached during the preceding meeting in London. The IEC amendment, which was partly based on an earlier CENELEC proposal, was subsequently adopted by CENELEC [2], and processed through an accelerated acceptance procedure. After approval in September 2000, it was published in the Official Journal (OJ ) in December, just about one year after agreement was reached within the IEC committee. Per this amendment, harmonics tests per EN61000-3-2 can either be done per the existing standard, or the user may test per the amendment. The amendment changes the classi-fication of many products, i.e. allows the manufacturer to test them per the more relaxed Class-A limits, instead of using the potentially more strict Class-D limits. Also, the amendment changes the way limits are computed for Class-D and Class-C products, and requires that the overall measurement is performed in accordance with IEC61000-4-7. In fact, the complete Annex-B to IEC61000-3-2 (second edition) which describes the measure-ment methodology is to be replaced by IEC61000-4-7. These changes impact various areas, which are not all that well understood as yet. This paper explains all the changes from a practical viewpoint.Classification of products & limit calculationsThe existing standard EN/IEC61000-3-2 ( second edition ) has four product classes, each having their own set of limits for harmonic currents, as shown in Table 1. The limits for Class-A and Class-B are given in ampere (rms) and are fixed, irrespective of the power level of the product. Note that EN61000-3-2 applies to all products up to 16 Ampere per phase, although limits for professional equipment > 1000 Watt are still under consider-ation. Limits for Class-D equipment are proportional to the power level, and limits for Class-C (lighting products)depend on the fundamental current, and the Power Factor (for H 3). The limits for classes C & D are therefore proportional, while A & B are fixed.The existing standard classifies allmotor driven equipment as Class-A, and all portable electrical equipment ( for short time use ) is tested per Class-B. All lighting prod-ucts are tested per Class-C. For Class-D,the situation is a little more involved, and some historical perspective is in place.The IEC developed a so called Com-patibility Standard ( IEC61000-2-2 ) which permits specific voltage distortion percent-ages for the public supply network. These distortion levels are specified for individual voltage harmonics. Similar to IEEE-Std-519,the compatibility percentages concern power quality at the HV - MV and LV level.A portion of the overall distortion percent-ages are allocated to the LV level, and sub-sequently to products which are directly con-nected to the Low Voltage ( LV ) public dis-tribution system.Table -1 EN/IEC61000-3-2 Harmonic current limitsHarmonic Class-A Class-B Class-C Class-D Amp Amp % of Fund mA/Watt 2 1.08 1.6223 2.30 3.4530*PF 3.440.430.65 51.442.1610 1.960.300.45 70.77 1.127180.230.35 90.400.6050.5100.180.28 110.330.5030.35120.150.23130.210.3230.29614/40 (even) 1.84/n 2.76/n 15/39 (odd) 2.25/n3.338/n 3 3.85/nHarmonic Distortion Order %20.3030.8550.6570.6090.40110.40130.3015-390.254-400.20 These admissible contributions to voltage distortion levels by individual products formed the basis for the harmonics limits of Class-A type products per IEC/EN60555.2 and its successor standard IEC61000-3-2. Table -2 shows these compatibility levels, i.e. permitted voltage distortion contribution at the 230 V - 50 Hz LV level.The European Reference Impedance is given in IEC-60725, which was based on work in the 1970’s. The Reference Impedance values are set such that 95 % of the network will have an impedance at or below the selected values. The resulting impedance levels are given in Fig. 1. Recent tests [3] show that the IEC-60725resistive levels appear to be about correct, while the reactive component appears to have decreased somewhat ( probably due to a shift from overhead to underground cabling).The following example illustrates how the EN60555.2 Class-A harmonic limits were derived, using the compatibility levels, and system impedance. The LV distribution network impedance at the 3rd harmonic (150Hz) is (0.42 + (3 x j 0.25)2)0.5 = 0.85 Ohm. The 0.85 % (compatibility) permit-ted voltage distortion level for H 3 in Table-2, means a max. voltage distor-tion of 1.955 V ( 0.85 % of 230 V ). Hence, I H3 is allowed to be 1.955 / 0.85= 2.3 Amp before 0.85 % distortion for VH 3 is reached. The same calcula-tions apply to the other harmonics, and they in turn form the limit set for Class-A equipment. Class-B involves portable equipment, intended for short time use, and the limits for these electrical products are simply 1.5 times higher than those of Class-A. The limits for lighting products and other high volume consumer type products ( TV’s ) were set more stringent. Lighting products are easily defined as a class, and thus presented little problems.In EN60555.2 harmonic limits were defined for TV’s. These limits were based on the harmonics levels that a switch mode power supply with anacceptable current conduction angle would exhibit. In EN61000-3-2, this TVtest class was extrapolated into a new Class-D test class. Since the current waveform can be more or less defined by its conduction angle, the “specialwave shape envelope” was defined as the identifying method. The Class-Dlimits in mA/Watt were based on the samemethodology (for the 75 - 600 Wattrange). This special wave shape methodto define what constitutes a Class-D prod-uct caused some controversy. TV’s areso called “high impact” products. Not onlyare some 20 million sold every year inthe EC, but many TV’s are also operatedsimultaneously during the same peakhours. As the switch mode power supplycurrent pattern was used as the methodto identify Class-D products however, thiscauses many “low impact products” withswitch mode power supplies to be classi-fied as Class-D equipment as well, be-cause these power supplies exhibit the typical “Class-D special waveform”. Someproducts with fluctuating power levels have this “special waveform” for only partof the time, and exhibit a Class-A wave-form for the rest of the time.This is one of the key areas whereCENELEC-A14 changes the definition of Class-D products. Per the amendment, only TV’s, PC’s and PC monitors are to be tested per Class-D limits. This means that many products migrate to Class-A.Note that there is no change in classification for Class-A, B, and C products, but the harmonic analysismethod for products with fluctuating power is affected by A-14 for these products as well.= 0.40 + j0.25Fig. 1 - Impedance model for European LV distribution Table-2 Compatibility LevelsNew Class-D & Class-C Limits and Fluctuating HarmonicsThe limits for Class-C & D are proportional as earlier explained. Whereas this doesn’t cause any difficul-ties for products with a constant current/power level, the situation was less clear for products with fluctuating load levels. Most test systems implemented so called dynamic limits, with the limits constantly being adjusted per the measured power ( or the fundamental current for Class-C ) while others used some average power level to set the limits. The latter systems determine this average power/current using some arbitrary method, and pre-test period. Thus different test systems implement different limits for the same (fluctuating power) products,which can result in one system PASSING a product while the other REJECTS it.A second issue for fluctuating loads is the way the existing standard (second edition) defines criteria for passing and failing the harmonics test. The existing standard permits the unit under test to occassionally exceed the 100 % limit, provided the harmonics never exceed 150 % of the limit. In fact, the unit under test is allowed to exceed the 100 % level for 10 % of the test time. The test time for fluctuating loads is to be at least 2.5minutes, i.e. the harmonics can exceed the 100 % limit for 15 seconds in every 150 second ( 2.5 min ) period.For longer test times, one can perform this test in 2.5 minutes “time blocks” but another interpretation is to just take 10 % of the overall test time. Thus, the testing method for fluctuating loads was somewhat subject to interpretation by the test equipment manufacturer.CENELEC A-14 addresses these area in two ways. First of all, the manufacturer is required to declare the rated power ( the fundamental current and Power Factor for Class-C ) for the product. This “rated power”(current & PF ) will be used as the basis for the limit calculation of Class-D ( Class-C). Thus, a “rated power” of 150 Watt for example, will yield a third harmonic limit of 150 x 3.4 = 510 mA. Similarly, a rated fundamental current of 0.4 Amp, and a rated PF of 0.98 for an electronic ballast will result in a third harmonic limit of 0.4 x 0.3x 0.98 = 117.6 mA. Thus, the amendment requires that an automated compliance test system must allow the user to enter this rated power or rated current. Also, the test system must verify the power (fundamental current and PF) because these “rated vales” as declared by the manufacturer must be within +/- 10 % of the actual values. If not, the actually measured values are to be used for the limit calculation. The method to measure these actual power, fundamental current and PF differ from the “average method” used in existing test systems,and of course differ also from the “dynamic limit method”.The Pass/Fail criterion is given in Fig. 2 above. The “10 % of the time over 100 % of the limits” no longer exists. The average harmonic level during the whole test must simply be below the limit, and individual values in each acquisition window ( after 1.5 sec filtering) must be below 150 % of the limit. Also note the extra allowance for the higher harmonics from H 21 - H 39. The exact method to determine the “rated” power, fundamen-tal current and PF , and how to measure these, is discussed in more detail below.The average value for the individual harmonic currents, taken over the entire test observation period shallbe less than or equal to the applicable limits.For each harmonic order, all 1,5 s smoothed r.m.s. harmonic curent values, as defined in 6.2.2, shall be lessthan or equal to 150% of the applicable limits.Harmonic currents less than 0,6% of the input current measured under the test conditions, or less than 5mA, whichever is greater, are disregarded.For the 21st and higher odd order harmonics, the average values obtained for each individual odd harmonicover the full observation period, calculated from the 1,5 s smoothed values according to 6.2.2 may exceedthe applicable limits by 50% provided that the following conditions are met:-the measured partial odd harmonic current does not exceed the partial odd harmonic current which can be calculated from the applicable limits.- All 1,5 s smoothed individual harmonic current values shall be less than or equal to 150% of theapplicable limits.Fig. 2 Pass/Fail criterion per CENELEC Amendment-14 to EN/IEC61000-3-2Measuring “rated” power, fundamental current and PFThe process to measure this maximum averaged power therefore is as follows; Calculate the power for every acquisition window ( 200 ms is preferred, although 16 cycles of the fundamental are permitted) and pass this value through a 1.5 seconds smoothing filter. Take the maximum level of these 1.5 sec. smoothed values, and compare this with the “rated power” ( or fundamental current and PF ) and establish whether or not it is within +/- 10 % of the manufacturer’s rated power. If so, use the “rated power” for harmonic limit calculations and pass/fail comparison. If the measured value deviates by more than 10 % from the rated value, use the mea-sured value to compute the harmonics limits and perform the pass/fail test. It is extremely important to follow this procedure, as the peak level of the 1.5 sec. smoothed value will differ enormously from just the peak value for a given acquisition window ! Obviously, it behooves the manufacturer to determine the “rated” values using the same methodology.Assessment of Quasi Stationary and Fluctuating HarmonicsFor product testing per A-14, especially those with fluctuating power, one must complete the whole test before the peak 1.5 sec. smoothed power ( fund. current & PF ) can be determined, and before the limit calculation and Pass/Fail test can be finalized. Of course, intermediate results - based on either the “rated values” by the manufacturer or on peak levels observed thus far - can be given, but the official Pass/Fail decision cannot be made until the test is complete.The test duration has to be sufficiently long to obtain better than +/- 5 % repeatability between successive tests. For products with a steady state power level ( quasi stationary ) this test time can be short, even less than a minute. For products with short power cycles (cycles < 2.5 min.) the test should include at least 10 power cycles. For products with very long cycles, the manufacturer can provide information showing the worst case condition- highest Total Harmonic Current operating mode - in order to minimize test time. The test methods per the amendment therefore no longer makes the distinction between Steady State and Fluctuating harmonics test methods that the existing EN61000-3-2 (second edition) makes. In fact, the whole existing Annex-B (in 61000-3-2) is replaced by a reference to (the new) standard IEC61000-4-7. This new IEC61000-4-7 does have signifi-cant impact on the harmonic analysis methods, including the fact that interharmonic currents enter the equa-tion. This aspect is illustrated in Fig. 3 below, and explained in more detail in the following section.R.M.S. value of a harmonic groupThe square root of the sum of the squares of the amplitudes of a harmonic and the spectral components adjacent to it within the observation window, thus summing the energy contents of the neighbouring lines with that of the harmonic proper. See also equation 3 and Figure 4. The harmonic order is given by the harmonic considered.Fig. 3 Graphical representation and formulas to calculate harmonic group RMS value ( 200 ms cycles )The impact of the new IEC61000-4-7Before discussing the impact of this new standard, it is important to note that within TC77/SC77A/WG-1 there are several Task Forces. Whereas TF1a was busily working on the Millennium Amendment, TF2 has been working on a revision of IEC61000-4-7 for a number of years. Several of the participants in TF2 in fact also are members of TF1a. Hence, some of the work that was done in the context of revising IEC61000-4-7 was initially also included in the text of the Millennium Amendment. Since the new IEC61000-4-7 is already at the CDV (committee draft for voting ) level, the Millennium amendment at one time even included a reference to this CDV. This however, is procedurally not acceptable, and hence both the IEC and the CENELEC Amendment now simply state that the entire Annex B is replaced by “The requirements for measurement equipment are defined in IEC61000-4-7”.Whereas small portions of the IEC61000-4-7 CDV still make their appearance in the CENELEC amend-ment, it is relevant to examine what the new IEC61000-4-7 will say about the measurement methods, as it directly impacts how products are to be evaluated. The important factors of the revised IEC61000-4-7 will be the move to a 200 ms measurement window and the inclusion of interharmonics in the harmonic level computation. Until IEC61000-4-7 is revised again ( hopefully not until at least 5 years from now ) measurements using existing test equipment having 320 ms/16 cycles ( 266.7 ms for 60 Hz ) acquisition windows will be permitted. The new 200 ms acquisition window better accommodates both 50 and 60 Hz systems, as 200 ms includes exactly 10 cycles of 50 Hz and 12 cycles of 60 Hz. Also, there are exactly 15 such 200 ms periods in a 3 second averaging period which is commonly used in power quality analysis.Including interharmonicsIn the new IEC61000-4-7, interharmonics are “lumped in” with the (integer) harmonic subgroup level per a precisely specified algorithm. Although interharmonics may not necessarily have the cumulative effect as is the case with odd triplens ( 3 rd, 9 th , 15 th, etc ) they do generate heat, and adversely affect electrical gear. Hence the interharmonics are lumped into the adjacent integer harmonic level using the geometric averaging algo-rithm, similar to a RMS calculation, as given in Fig. 3. This geometric average of the harmonic subgroup is filtered (using the 1.5 sec filter) and then compared against the applicable limits. The new standard also defines the algorithm to calculate just the interharmonic subgroup ( see Fig. 3 ). Note that if the 16 cycle acquisition window were to be used, the algorithms will change somewhat, as the number of interharmonics from the FFT would differ. This is reflected in the different algorithms. Fig. 3 shows the algorithm for a 0.2 sec. acquisition period, and the graphical representation of this concept for a 50 Hz power system.ConclusionsCENELEC Amendment-14 to EN61000-3-2 moves many electronic products from test Class-D to the generally more relaxed Class-A, with limits initially being based on the manufacturer’s “rated” power. This rated power is to be determined in accordance with the 1.5 sec filtered measurement method as given in the amend-ment. The same method applies to determine the Fundamental Current and Power Factor for Class-C products.The amendment defines measurement methods and harmonic limit comparisons that apply to both, sta-tionary and fluctuating harmonic analysis. Per the amendment, the normative Annex-B of EN61000-3-2 is to be replaced by the measurement and analysis methods given in (the new) IEC61000-4-7. This new standard requires the use of a 200 ms acquisition window, and specifies algorithms to include interharmonics into the so called harmonic subgroups.The CENELEC Amendment therefore greatly affects compliance test systems. Especially manufacturers of Class-D products need to implement the new methods to determine what “rated” power levels they need to specify for their products. This applies also to manufacturers of lighting products ( Class-C ) for fundamental current and PF as well. Testing authorities and test laboratories should also take the changing requirements of IEC61000-4-7 into consideration.References: 1: EN61000-3-2 Second Edition 2000-08, IEC, Geneva - Switzerland2: Amendment A14:2000 to EN61000-3-2 :1995 OJ Publication date 14-12-2000.3: Low Voltage System Impedance Test, Mathieu van den Bergh - Oct. 2000。