25th International Conference on Electricity Distribution Madrid, 3-6 June 2019

Paper n°1407

SUITABILITY OF TEST PROCEDURES IN IEC 61000-3-2 FOR ASSESSING

HARMONIC EMISSION OF MODERN MASS-MARKET EQUIPMENT

Jan MEYER, Ana Maria BLANCO Roberto LANGELLA Sasa DJOKIC

Technische Universität Dresden Università della Campania The University of Edinburgh
Germany “Luigi Vanvitelli” - Italy United Kingdom
[email protected] [email protected] [email protected]

ABSTRACT not to introduce intolerable disturbances into this

environment (emission aspect) [2]. There is a general
The standard IEC 61000-3-2 specifies harmonic emission expectation that the normative test conditions for
limits for equipment with input currents below or equal to emission limit compliance assessment should reflect
16 A. The limits, as well as the related test procedures, “worst case” conditions, regarding both setting of
have been developed several decades ago and are equipment operating points and test supply conditions
specifically dedicated to older types (“classical”) (voltage and impedance). In this way, it should be
equipment with line-commutated circuit topologies. ensured if the equipment complies with harmonic
Nowadays, the share of “modern” power electronic (PE) emission limits in tests, its harmonic emission under
equipment with self-commutating topologies is realistic supply voltage conditions will be in most cases
continuously increasing and the suitability of the limits, lower than under the test conditions.
methodologies and test procedures in IEC 61000-3-2 For classical line-commutated converter topologies, the
should be critically reviewed. This paper compares the pure/ideally sinusoidal supply voltage waveform will in
harmonic performance of around 100 “classical” and 50 the majority of cases result in a higher harmonic current
“modern” PE household appliances, which can be emission than under the flat-top distorted supply voltage
generally denoted as “mass-market equipment”. Based waveform. It should be noted that emission can be higher
on the identified differences, the paper discusses needs under pointed-top supply voltage waveform distortion,
and initial approaches for updating of existing limits and which is typical for industrial grids. This is not
test procedures. The paper aims to raise awareness to the considered in this paper.
qualitatively different behaviour of classical and modern Several studies for devices with modern circuit
topologies (e.g. PWM-based converters for photovoltaic
PE equipment and contribute to the development of and battery storage applications, or rectifiers with active
appropriate standards, which should ensure appropriate power factor correction for electric vehicle chargers and
EMC levels in existing LV networks and future grids. computer power supplies) have shown that the harmonic
currents emitted by these devices at flat-top supply
INTRODUCTION voltage distortion can be considerably higher compared
The standard IEC 61000-3-2 [1] defines harmonic to the emission measured under the reference sinusoidal
current emission limits for electrical and electronic conditions (e.g. [4-5]). This suggests a qualitatively
devices with a rated input AC current up to 16 A per different behaviour of modern PE equipment compared to
phase. According to the European directive on classical line-commutated topologies. Consequently, it
Electromagnetic Compatibility (EMC) [2], the has to be questioned if the test conditions in
compliance with this standard is required for each device IEC 61000-3-2 are still suitable to adequately reflect a
that enters the European market and generates harmonic typical or worst-case electromagnetic environment for the
currents when connected to public low voltage (LV) assessment of harmonic emission performance of modern
distribution systems. PE devices. A first indication that this is considered to be
To assess the compliance of a device, the harmonic a concern is found in IEC 61000-3-15 [6], which
current magnitudes have to be determined in tests with an proposes for dispersed generation units an additional
as sinusoidal as possible supply voltage with nominal assessment of emission limits under distorted supply
magnitude. However, field measurements in actual voltage, including one variant of a flat-top test voltage
residential public LV networks show that the supply waveform.
voltage is typically not sinusoidal, but has a more or less This paper presents a comprehensive survey of the
pronounced flat-top shape, with a total voltage harmonic sensitivity of harmonic emission of more than 150
distortion usually higher than 2% [3]. Moreover, the rms different “classical”, as well as “modern” PE devices (EV
value of the voltage is also not constant and equal to the chargers, PV inverters, LED lamps, computer power
nominal value, but it can vary significantly during the supplies, chargers, etc.) for two different supply voltage
day: allowed variations in European public LV networks waveforms (sinusoidal and flat-top voltage waveform)
are in the range up to ±10% of nominal voltage. and for three supply rms voltage magnitudes (230 V,
EMC is enforced in many countries by law and has two 230 V + 10 % and 230 V - 10 %). The presented results
aspects. Firstly the devices must be able to function shall contribute to the discussion on the differences
satisfactorily in their electromagnetic environment between “classical” and “modern” PE devices and the
(immunity aspect) and secondly the devices are assumed rationale to define revised and more realistic emission

CIRED 2019 1/5

 25th International Conference on Electricity Distribution Madrid, 3-6 June 2019

Paper n°1407

limits and test procedures for “modern” PE devices. This current waveforms of devices with different circuit
is important, as the numbers and sizes of “modern” PE topologies when a sinusoidal voltage supply waveform is
devices based on self-commutating converter topologies applied. The first two waveforms correspond to devices
are continuously increasing, while opposite is true for with no-PFC topologies; the first curve is a typical
“classical” line-commutated devices. waveform of CFLs and LED lamps, while the second
The paper starts with a brief overview of the influence curve is a typical waveform of small battery chargers of
of the supply voltage distortion on the harmonic emission smartphones and other similar devices. The passive and
of different types of mass-market equipment, generally active PFC topologies are common in devices with higher
termed as “household appliances”. Next, the devices input power, like computers, EV chargers, etc. Fig.1b
selected for analysis are classified according to their shows the current emitted by the same devices of Fig.1a
circuit topologies and indices used for the analysis and for a flat-top supply voltage waveform (230V rms; odd
comparison are introduced. For each device class, the harmonics up to order 15), with a similar shape as the one
impact of magnitude variations and flat-top supply suggested in IEC 61000-4-13 for immunity testing (cf.
voltage distortion is presented and discussed. The paper Fig. 2). A comparison of Fig. 1a and Fig. 1b reveals that
finishes with a generic proposal on how the test the harmonic current emission might depend significantly
procedure and emission limits could be extended, in order on the presence of supply voltage distortion; particularly
to take “modern” PE devices adequately into account. for devices 1, 2, and 4. More details about this behavior
can be found e.g. in [3-5, 8-9].
THEORETICAL BACKGROUND It should be noted that the power electronic interfaces
Power electronic based, or power electronic interfaced for generators (e.g. PV inverters) or converters for energy
household devices, have different circuit topologies, storage applications also show significant dependency of
which can be broadly classified into three categories, the harmonic current emission on the supply voltage
depending on the type of power factor correction (PFC): distortion [9], which is similar to the “modern”
consuming devices belonging to Category 3. Although
1. No-PFC topology: this paper mainly focuses on consuming devices with
They usually consist of a simple front-end with a input current below or equal to 16 A (scope of IEC
diode bridge rectifier and a smoothing capacitor. 61000-3-2), six PV inverters are also included in the
Capacitive divider-based power supplies also belong presented analysis and allocated to Category 3.
to this category. Devices with no-PFC topology have
a low true power factor (PF<0.6) and usually a high ANALYSIS FRAMEWORK
current harmonic emission.
2. Passive-PFC topology: Equipment selected for analysis
They usually feature a front-end with additional series All measurements of different devices in this paper are
capacitors or inductors before or after the diode obtained from equiPment hArmoNic Database [7]
bridge rectifier, in order to increase the true power (PANDA, https://www.panda.et.tu-dresden.de), which is
factor (usually 0.6≤PF≤0.9) and reduce current hosted at the Technische Universität Dresden. PANDA is
harmonic emission. Valley-fill circuit topologies also a global web-based platform for exchanging harmonic
belong to this category. emission measurements of household equipment
3. Active-PFC topology: (typically single-phase equipment with rated current
They have a complex front-end with advanced control below 16A) between different laboratories around the
circuits that shape the input AC current waveform. world. Currently, PANDA contains more than 3800
This topology has the best performance related to true measurements of more than 700 household appliances,
power factor (PF>0.9) and lowest current harmonic which were provided by 16 laboratories from Europe,
emission under sinusoidal supply voltage. Asia, Australia and North and South America. More
Devices in Categories 1 and 2 belong to “classical” information can be found on the website, or in [7].
topologies, while devices in Category 3 belong to A set of 145 appliances has been selected from
“modern” topologies. PANDA, which represents a broad range of household
As an example, Fig. 1a shows the normalized input AC electronic appliances sold in Europe, including CFLs,
1 No PFC 1 (0.4A)
No PFC 2 (4.4A)
in p.u.

0.5 Pasive PFC (2.2A)

Active PFC (1.6A)
0

0.5

1
0 2.5 5 7.5 10 12.5 15 17.5 20
Time in ms
a) b)
Fig. 1 Normalized input AC current waveforms of typical electronic appliances for sinusoidal (a) Fig. 2 Sinusoidal and flat-top
and flat-top (b) supply voltage waveforms (maximum i(t) value is indicated in Legend). supply voltage waveforms.

CIRED 2019 2/5

 25th International Conference on Electricity Distribution Madrid, 3-6 June 2019

Paper n°1407

Assessment indices and presentation of results

For this paper, the absolute deviation (ADI) and the
relative deviation (RDI) between the measured harmonic
current magnitude of a test point and a reference
point are presented exemplarily for odd harmonic
orders up to h = 19 (the most dominant in LV grids). The
values obtained for a supply voltage with sinusoidal
waveform and rms magnitude of 230 V serve as
reference, because they correspond to the test conditions
defined in IEC 61000-3-2.
Fig. 3 True power factor versus THDI of all selected appliances The absolute and relative deviations are calculated as:
measured for a sinusoidal 230 V supply voltage waveforms.
(1)
LEDs, PC power supplies, laptop chargers, battery
chargers for smart phones and tablets, flat-top panel TVs,
PC monitors, PV inverters and electric vehicle chargers. 100 % (2)
The selection was made in order to get approximately
same numbers of appliances for each topology category. The relative deviation allows a better comparison
Fig. 3 presents the true power factor and the THDI of the between the different appliances, but very small harmonic
selected appliances, grouped according to their circuit current magnitudes can result in high relative deviations.
topologies for sinusoidal supply voltage with nominal In order not to draw misleading conclusions, the absolute
(230 V rms) magnitude (50 appliances of Cat. 1: no-PFC, deviation is also presented. Finally, it is the absolute
45 appliances of Cat. 2: passive-PFC, and 50 appliances harmonic current, which has to be limited with regard to
of Cat. 3: active-PFC, including six PV inverters, have the impact on the harmonic voltages in the network.
been selected). The results are presented for each category using
modified boxplots (e.g. Fig. 4 and Fig. 5). These boxplots
Measurement procedure represent the 5th (lower line), 50th (middle line; median)
All the selected appliances were measured under test and 95th (upper line) percentiles of the obtained ADI and
similar conditions, using a programmable voltage source RDI values of all devices in the respective category. Data
and a measurement device (voltage transient recorder, or points outside the 5th – 95th percentile range are plotted
power quality analyzer) capable to measure the using cross symbol.
magnitude and phase angle of voltage and current
harmonics based on the specification in IEC 61000-4-7 IMPACT OF RMS VOLTAGE VARIATION
[10]. Three different supply voltage magnitudes (230 V, The test points correspond to the sinusoidal voltage
230 V + 10 %, 230 V - 10 %) under sinusoidal conditions supply waveform with rms values of 207 V (230 V -
and a flat-top supply with 230 V rms (cf. Fig. 2) were 10 %) and 253 V (230 V + 10 %). The results for current
used for the analyses. Both parameters of influence harmonic magnitudes are presented in Fig. 4. Due to the
(supply voltage magnitude and supply voltage distortion) very different sensitivities of “classical” and “modern”
are studied independently in this paper, which allows for PE devices, the value ranges of the Y-axes differ
a better understanding of their individual impact. significantly between plots for Cat. 1-3. The white and

Category 1 (No-PFC) Category 2 (Passive-PFC) Category 3 (Active-PFC)

Fig. 4 Impact of voltage magnitude on harmonic current emission for different topology categories (sinusoidal waveform).
White and grey bars correspond to the two test points of 207 V and 253 V.

CIRED 2019 3/5

 25th International Conference on Electricity Distribution Madrid, 3-6 June 2019

Paper n°1407

grey bars correspond to the two test points of 207 V and questionable if the methodology for setting and assessing
253 V. harmonic emission limits, which has been developed for
For no-PFC and passive-PFC devices (Cat. 1 and 2), a “classical” rectifier topologies, can simply be transferred
relatively small increase of harmonic currents, with RDI to the new generation of devices.
values typically below 50 % (no-PFC) and below 100 % With regard to the harmonic emission limits, a clear
(passive-PFC) is observed. In contrast, the active-PFC distinction between power-consuming devices and
devices including PV-inverters (Cat. 3) show very high power-generating, as well as energy storage devices
sensitivity to supply voltage magnitude variations, in should be introduced. The emission limits for these
particular for the 253 V test point). devices should be lower, because stricter requirements
should apply for devices, which generate energy and may
IMPACT OF FLAT-TOP VOLTAGE also provide network services. Another reason is that
WAVEFORM DISTORTION “modern” self-commutating topologies do not produce
harmonics like “classical” topologies and therefore
Only a single test point exists, which represents the should not be assessed against rather high harmonic
current harmonics magnitudes obtained for flat-top limits derived for “classical” devices. This applies also
supply voltage waveform. The sinusoidal supply voltage for “modern” consuming devices based on active-PFC
waveform with rms value of 230 V again serves as a topologies. The existing Classes A to D in IEC 61000-3-2
reference. The results are presented in Fig. 5. It should seem therefore not adequate and the introduction of a
again be noted that due to the very different sensitivities, new Class E should be considered. In this way, the
the value ranges of the Y-axes differ significantly existing concept of emission limits and their assessment
between the plots. for classical devices (Classes A to D) can be kept as it is,
Comparing Fig. 4 and Fig. 5 reveals that the presence of but a higher flexibility in dealing with modern devices
supply voltage distortion has a considerably higher will be achieved.
impact on the harmonic current magnitudes than the Only for the new class E the test procedure should be
variation of voltage magnitude for no-PFC and passive- extended in order to take into account the influence of
PFC categories. For no-PFC devices, the relative supply voltage distortion on the harmonic emission of
deviation increases with the harmonic order and it is “modern” devices. Accordingly, the proposed test
roughly comparable to active-PFC devices for higher procedure consists of three parts. Part A has two steps
harmonic orders, but significantly smaller for lower and tests the harmonic current magnitudes of equipment
orders. Passive-PFC devices do not show such tendency, under test (EUT) under two supply voltage waveforms:
but usually reduce the harmonic current emission at lower
harmonic orders (3rd and 5th harmonic). Again the active- A1. Apply sinusoidal supply voltage with nominal rms
PFC devices seem to be significantly more sensitive than voltage and compare current harmonic magnitudes
the devices with classical rectifier topologies. The very with a limit set A (existing limits). In case of non-
high sensitivity for the 5th harmonic is mainly determined compliance, the EUT fails (this equals the present
by PV inverters. test procedure), otherwise it goes to step A2.
A2. Apply a flat-top supply voltage and compare
POSSIBLE MODIFIED TEST PROCEDURE current harmonic magnitudes with limit set A. In
case of compliance, the EUT goes to step C,
The results of surveying 145 devices have shown that otherwise it goes to step B.
“modern” PE devices behave significantly different To consider an acceptable change of the emission due
compared to the “classical” ones. Therefore, it is to the presence of supply voltage distortion, step B is
Category 1 (No-PFC) Category 2 (Passive-PFC) Category 3 (Active-PFC)

Fig. 5 Impact of flat-top voltage waveform on harmonic current emission for different topology categories (nominal rms voltage).

CIRED 2019 4/5

 25th International Conference on Electricity Distribution Madrid, 3-6 June 2019

Paper n°1407

introduced, for which limit set B is applicable, which has PFC, behave in many cases significantly different than
higher, i.e. more relaxed limits than set A. “classical” devices based on line commutating front ends.
B. Check harmonic emission of EUT with respect to The paper also shows that sinusoidal test conditions
limit set B. In case of non-compliance the EUT specified in IEC 61000-3-2 do not represent a “worst-
fails, otherwise it goes to step C. case” condition anymore for “modern” devices and it has
to be questioned whether the main objective of
The change in harmonic emission between sinusoidal Electromagnetic Compatibility is still achieved with the
and flat-top test voltage waveform is mainly determined growth of the share of “modern” devices in the future.
by the design of the devices’ circuit. A certain A first idea for a revised test procedure for assessing
“robustness” of the EUT is of significant importance for a compliance with emission limits is proposed, which takes
reliable operation of the LV networks. Therefore, in the characteristic of “modern” devices into account. The
addition to limiting harmonic current emission, the proposed revision of current harmonic emission
sensitivity of the EUT should be assessed based on a limit assessment also checks “robustness” of tested device, as
set C, which tests the ratio of the change between flat-top too sensitive devices can seriously impact the reliable
and sinusoidal supply voltage waveforms. operation of public LV networks. This initial idea can be
C. Check the ratio of the current harmonic easily extended, e.g. by taking other supply voltage
magnitudes between flat-top and sinusoidal supply distortion into account. Additional flexibility could also
voltage for the EUT with limit set C. If the EUT be achieved by introducing criteria assessing the sum of
complies, it passes, otherwise it fails. the exceedances as total value (comparable to THC), or
The proposed extension of the existing test procedure in by introducing fixed minimum absolute deviations. The
[1] requires only one additional test point (A2, with flat- approach is not limited to consuming devices, but could
top voltage waveform), which, together with newly also be included in standards for generating and storage
introduced limits in Steps B and C, allows to correctly appliances.
assess harmonic emission of modern PE equipment. Further work is required to analyze the need to include
The above points are demonstrated using measurements magnitude variations of supply voltage into the test
of 14 on-board EV chargers (Ir ≤ 16 A). By intention, no procedure and to adapt existing emission limits to the
absolute values are provided, as the example shall serve characteristic of “modern” PE devices.
for illustrating the principle of the proposed method.
Fig. 6 presents the harmonic current magnitudes together REFERENCES
with the respective emission limits. EV3 fails already in [1] IEC 61000-3-2:2018 (RLV): EMC - Part 3-2: Limits -
Step A1. All EVs except EV5 go directly to Step C. EV5 Limits for harmonic current emissions (equipment input
passes Step B and goes to Step C as well. Fig. 7 presents current ≤16 A per phase).
the ratio of measured harmonic currents between flat-top [2] Directive 2014/30/EU of the European Parliament and of
and sinusoidal supply voltage for all EVs, along with the the Council of 26 February 2014 on the harmonisation of
respective limit C. Finally EVs 1, 2 and 6 fail due to too the laws of the Member States relating to electromagnetic
high sensitivity. compatibility, http://data.europa.eu/eli/dir/2014/30/oj
[3] S. Yanchenko and J. Meyer, "Harmonic emission of
household devices in presence of typical voltage
CONCLUSIONS distortions," 2015 IEEE PowerTech, Eindhoven, 2015
[4] A. M. Blanco, S. Yanchenko, J. Meyer, and P. Schegner,
Based on a comprehensive survey of harmonic emission “Impact of supply voltage distortion on the current
of around 150 household appliances, it is shown that harmonic emission of non-linear loads,” Dyna, vol. 82,
“modern” PE devices, generally equipped with active no. 192, pp. 150–159, 2015.
[5] X. Xu, A. J. Collin, S. Z. Djokic, S. Yanchenko, F.
Möller, J. Meyer, R. Langella, A. Testa, “Analysis and
Modelling of Power-Dependent Harmonic Characteristics
of Modern PE Devices in LV Networks”, IEEE Trans. on
Power Delivery, Vol. 32, No. 2, pp.1014-1023, 2017.
[6] IEC TR 61000-3-15:2011: EMC - Part 3-15: Limits -
Assessment of low frequency electromagnetic immunity
and emission requirements for dispersed generation
systems in LV network.
[7] A.M. Blanco, E. Gasch, J. Meyer, P.Schegner, "Web-
Fig. 6 5th harmonic current magnitude of 14 on-board EV chargers based platform for exchanging harmonic emission
for sinusoidal/flat-top waveform (white/gray). measurements of electronic equipment," IEEE ICHQP,
2012.
[8] S. Cobben, W. Kling, and J. Myrzik, “The making and
purpose of harmonic fingerprints,” Proc. CIRED, 2007
[9] S. Müller, J. Meyer and P. Schegner, “Characterization of
Small Photovoltaic Inverters for Harmonic Modeling”,
Proc. ICHQP, 2014
[10] IEC Standard 61000-4-7:2002+A1:2008 : EMC - Part 4-7:
Testing and measurement techniques - General guide on
harmonics and interharmonics measurements and
instrumentation, for power supply systems and equipment
Fig. 7 “Sensitivity” ratio of the EV chargers. connected thereto.

CIRED 2019 5/5