An Analysis of Harmonics from LED Lamps

Sohel Uddin#i, Hussain Shareef2, Azah Mohamed#3, M A Hannan#4

Department a/Electrical, Electronic and Systems Engineering, Universiti Kebangsaan Malaysia

43600 Bangi, Selangor, Malaysia
lsohel 091@yahoo.com; 2shareef@eng.ukm.my; 3azah@eng.ukm.my; 4hannan@eng.ukm.my

Abstract- Light Emitting Diodes (LEDs) are versatile, energy [6]. The deviation of waveform from perfect sinusoid is
efficient when compared to the conventional light sources. LEDs usually expressed in terms of harmonic distortion of the
represent a transformational change in how light is produced. current and voltage waveforms.
However, there are significant differences between conventional A large number of studies were conducted on LEDs as an
lighting sources and LEDs in terms of voltage and current energy efficient lamp but most of researchers pay their
operating requirements which may affect the power quality (PQ)
attention on the internal ballast circuit design and enhancing
of AC mains supplying the LED bulbs. Therefore, this paper
their performance [7-8]. Several other researchers have
investigates harmonic generations from LED bulbs available in
concentrated on light distribution and visual performance of
the market. It is done by conducting laboratory tests on various
LED lamps [9-10]. A few contributions focus on harmonic
LED lamps and tapping the load current behaviour under
different conditions. Then frequency domain analysis is emissions of LED lamps [4], [11]. According to [4], the
performed to investigate the generated harmonics. Harmonic widespread use of LEOs and CFLs could increase voltage
levels of different wattage, various branded LED bulbs are distortion in the distribution networks depending upon the
experimentally evaluated and compared. The results show that characteristics of the network. However, the work mainly uses
most of the LED lamps produce high level of harmonics with CFLs in the analysis.
reference to the IEC 61000-3-2 standard. This paper presents a detail harmonics analysis of LED
lamps. This is characterized by measurement tests, using
Keywords: Light Emitting Diodes (LEDs), Total Harmonic
various commercially-available LED bulbs. In the
Distortion (THD), Power Quality.
investigations, a total 35 pieces of different ratings from 7
I. INTRODUCTION different manufacturers (brands) are observed. All tests are
carried out to observe their current and voltage waveforms
Lighting accounts for roughly 20% of the electricity
and analyse them in terms of power rating, and brands. The
consumption all over the world. To promote energy saving,
test results are also compared with IEC 61000-3-2 harmonic
many governments in the world have introduced directives to
standard.
ban energy inefficient incandescent light bulbs [1]. Light
Emitting Diode (LED) lamps are becoming viable
II. BASIC OPERATION OF LEOS AND ITS HARMONIC
replacements in industrial and commercial lighting
STANDARDS
applications due to high energy efficiency and long life [2].
The principle operation behind LED bulbs and the
LEOs work on a completely different principle than the
harmonic emission limits for LEOs as defined by IEC 61000-
compact tluorescent lamps (CFLs). Individual electrons jump
3-2 are discussed in this section.
across a p-n junction of a semiconductor material. When each
electron recombines with an atom, it emits a particle of light A. Operation Principle ofLED Lamps
known as a photon. Because all of the light is being produced
LEDs require a constant current source from a low DC
at the junction, the resulting light source is a point source and
voltage source, but they must also operate from the AC mains.
requires many LEOs to light a large area [3].
Therefore, it is necessary to use a converter to regulate the
In general lighting applications, a compact AC/OC
voltage and control the current applied to the LEOs. The buck,
converter should be used in one lighting fixture to supply DC
boost, tlyback and resonant converters are well known in
current to LED chips which introduce nonlinearity to the
literature and widely used to supply LEDs, [12].
system. As a nonlinear load, LED bulbs produce highly
Fig. 1 depicts a block diagram of typical low-wattage LED
distorted currents [4]. Although the input power of a single
ballast. It includes the AC line input voltage, typically 220-
LED bulb is quite low, a large number of customers using
240 VAC 50/60 Hz, an EMI filter to block circuit-generated
LED bulbs and CFLs per premises could create significant
switching noise, a rectifier with smoothing capacitor, a PWM
power quality problems and disturb to other consumer
controlled constant current source converter for DC to DC
connected to the electrical power distribution systems [5].
conversion, and an array of LEOs. Since the rated load powers
Therefore a good understanding of the LED bulb's harmonic
are low, the directives governing the injection of harmonics
producing characteristics becomes necessary.
are not particularly strict [13], and therefore power factor
Harmonics are defined as the sinusoidal voltages or
control circuits may or may not be found in low-wattage LED
currents having frequencies that are integral multiples of the
lamp ballasts. However, to reduce the generated harmonics
frequency at which the supply system is designed to operate

978-1-4577-1559-4/12/$26.00 ©2012 IEEE 837

 and to improve the power factor it is possible to introduce TABLE II
TECHNICAL DATA FOR TESTED LED LAMPS
either an input passive filter, valley filled circuits, [C
Life
controlled active filtering configurations. Trade Name Nominal Current Luminous
Span,
Power,
l(mA) Flux (1m) (Year)
peW)
EM! Filter Bridge Rectifier Smooth DC Converter LED lamps
Philips Cool Daylight 4 27 250 25

N_c";:'7:�-
Philips Cool Daylight

~
5 32 350 25

,,�- �-
�
-+

-�... �
Philips Warm Daylight

Osram Daylight
7

4
39

40
350

250
25

25

Fig. Block diagram of LED electronic ballast Osram Daylight 6 53 365 25

Osram Cool White 8 47 450 25

B. Harmonic Injection Limits/or LED Lamps Osram Warm White 3 26 250 25

Similar to any other appliance, LED lamps also must Osram Warm White 10 48 950 25
comply with several directives which are applicable to the Toshiba Cool White 5.5 44 290 40
product. The IEC 61000-3-2 standard assesses and sets the Viri-bright cool White 5 55 230 6
limit for equipment that draws input current :S16A per phase
Cash Wann White 7 67 580 20
[14]. Harmonic emission limits for lamps are subdivided
Evenzo Cool White 3 33 150 18
based on their active power up to 25W and above. Lamps
having an active input power less than or equal to 25W must
Each lamp is tested for four times to eliminate any error
satisfy at least one out of the two following criterions. One of
during different period of the day, namely morning, evening
the criteria is that the third harmonic current should not
and afternoon. This is necessary to make sure that during the
exceed 86% of the fundamental and the fifth harmonic current
measurements period, the total voltage distortion (THOv) of
should not exceed 61%. That gives the value of the total
the supply system is below 3% and does not vary much to
harmonic distortion (THO) current approximately 105%.
have a consistent and data for analysis. Then all the
The other criterion is given as a TABLE I of limits for
measurements on each lamp were performed simultaneously.
each harmonic order, where limits are expressed as absolute
The captured current waveforms were analysed by using
currents per one watt of active power (mA/W).
Fluke 434 and MATLAB software.
TABLE I
lEC 61000-3-2 LIMITS FOR CLASS C EQUIPMENT (P <:: 25w)

Hannonic Order Maximum Permissible Hannonic

(n) Current (mA/W)
3 3.4

5 1.9

7 1

9 0.5

11 0.35

13 ::;n ::;39 3.85/n

PQ Analyzer Fluke
434
Fig. 2 Schematic diagram of experimental setup

III. METHODOLOGY [V. RESULTS AND DISCUSSION

To analyse the characteristics of the LED lamps, 12

In this section, measurements of various test conducted on
samples of with different power ratings from various
LED lamps are analysed and discussed. The lamp current
manufacturers, as shown in TABLE II, were tested. The lamps
waveforms were analysed using the Fourier Theorem. It
have build in electronic ballast which is powered using E-27,
provide frequency spectrum of the lamp currents represented
E-14 or GU-I0 type sockets, commonly available in retail
by the fundamental sinusoidal component and a series of
stores. All the tested lamps are designed to operate at 240 V
higher order harmonic components at frequencies that are
and have power consumptions rating of 3 W to 10 W.
integer multiples of the fundamental frequency. The square
To obtain accurate data concerning the exact current
roots of the sum of each the amplitudes of the harmonic
harmonic content of LED bulbs, an experimental setup as
present the THO and expressed in percentage of the
shown in Fig. 2 is assembled. [t consists of four components
fundamental frequency.
namely, Fluke 434 power quality analyser, Fluke i30s current
clamp, LED bulb(s) under test, and a personal computer to A. Findings from Single Lamps
analyse the signals. Each lamp is kept switched on for 10
minutes before the measurements are taken for stabilization.

838
 The purposes of these tests are to explore the type of lamps have same characteristics although the magnitude of
harmonic filters used in the LED ballast circuit and harmonic current increases with the increase of power rating.
emissions from all the tested lamps. The tests also provide an
TABLE III
insight about the harmonic generation levels of same branded
HARMON'C CONTENTS OF SINGLE LED LAMP S
lamps with different power rating, different branded lamps
Tested Lamp Harmonic (%)
with same active power rating.
The typical current waveform illustrated in Fig. 3 is Fund 3rd 5th 7th 9th THDr

obtained from different lamps tested. From the figure it can be Philips 4 W 100 34.61 6.28 22.27 18.94 63.05
noted that the current waveforms is totally distorted that
Philips 5 W 100 36.92 7.16 19.81 19.39 63.83
means they are inject harmonic in power system. Furthermore
from Figure 3, it can be seen that the ballast in dissimilar LED Philips 7 W 100 32.34 11.6 23.4 19.03 64.23

bulbs use different filtering methods to reduce harmonic Osram 4 W 100 89.24 70.65 51.19 38.57 173.9
generation. For example, most of the Phillips brand LED
Osram 6 W 100 91.96 77.05 58.51 41.01 174.3
bulbs tested utilize the Valley-filled circuit, the Toshiba 5.5 W
lamp contains a passive filter, Osram 8 sample embed an Osram 8 W 100 22.25 15.04 2.34 9.73 30.94

active filter, and some other tested lamps does not implement
Osram lOW 100 32.12 4.68 2.67 6.45 34.78
any filtering technique. Based on the type of filtering and peak
Evenzo 3 W 100 90.7 77.34 58.89 45.32 164.35
current drawn by the lamp, one can also classity these bulbs as
"good", "moderate" and "bad" LED lamps in terms of Bright 5 W 100 86.65 75.92 61.75 48.87 167.2
harmonic injection into the system.
Cash 7 W 100 91.23 74.81 56.41 41.06 168.2
0.3 ,------.---,--,--,
Toshiba 5.5 W 100 73.4 45.4 35.62 31.38 106.3

It further indicates that they use same type of ballast circuit.

Fig. 4(b) depicts the harmonic spectrum of these lamps. It is
noticed that the magnitude of harmonic current is decreased
with increased harmonic order and their magnitudes are
comparable with each other for different wattage lamps. Also
5 10 15 20
Time (ms) all the even harmonic is zero due to property of sine wave

Fig. 3 Various types of current waveforms obtained different LED bulbs

symmetry. Fig. 5 illustrated the current waveform of four
tested Osram lamps compared.

0.2 ;;- 400

To demonstrate further, relative harmonic currents of all
tested lamps are presented in Table III in terms of percentage � 0.1 --��:-- -PhilillS4W
-Ph,,,I"'W
200 �
'.:'5 0 �-;�'�.�.-.!...fc .l. - - - � VO!�:;t' 0 �
- - - - - - - - - - - - - - -
·····PhlllllS7W
to the fundamental frequency. Note that some of the tested
� -0.1 - V:.-� ;V: - - - 7- - - -200 �:;
.4,a" - - - �. .. u�&.�."••-.-"i- ....... ..

LED bulbs generate quite a high level of harmonic which

U
- - - - - - - - - - - - - - - - � - -u ."

produce unacceptable limits of harmonics when referred to -0.20'-----�5 ----'-----�----l0 400

2
lEe 61000-3-2 standard described earlier. When the THO
:::':
values presented in Table III is compared for lamps having � 100rn:��-,--���--';"'--,-�� - �:::o==�
trademark of Philips, it can be noted that the THO value °c 80 I �:�!!::�:
6 60
-'� ----------------------------
1_ - - - - - - - - - - - - - - - - - - - - - C -1Phili s 7W
generally varies between 63 to 65 % THO], while Osram ,
= 40
,-- - n l-In - - - - - - - - - - - - - - - - - - - - -
..

brand LED bulbs with power rating less than 6 W produce

E 20 , , . Ii
-

LtllLlliLru � �LlIiUO:� : i
10' 10' l[ �LIll
high levels of harmonics which is in the range of 173 to l:: 0
" I 3 5 7 9 II 13 15 17 19 21 23 25 27 29
175 % THO,. This is because these LED bulbs do not use any u Harmonic Order (n)
(b)
filter in their ballast circuit. However, in case of 8 W and 10
W Osram LED lamps produce the lowest THO, which is in Fig. 4 Test results of Philips 4 W, 5 W, 7 W LED bulbs: (a) Lamp current
the rage of 30 to 35%. All other tested lamps have THO, value and voltage waveforms (b) Individual harmonic spectrum

greater than 100% as seen in Table 3.

From Fig. 5, it can be seen that the current wave shapes
B. Findingsfrom Same Branded Lamps are totally different from Philips LED lamps and THO,
The purpose of this test is analysing the harmonic levels (TABLE III) values are much more for Osram 4 W and 6 W
from different wattage lamps from the same manufacturer. bulbs compared to 8 W and 10 W. This is because 8 W and 10

Here Philips 4 W, 5 W, 7 W lamps and Osram 4 W, 6 W, 8 W, W bulbs use different ballast configurations which produce
lOW LED bulbs are investigated. current waveforms similar to the pure sine wave. Hence, they
A typical voltage and current waveforms obtained from produce much less harmonic.
Philips 4 W, 5 W and 7 W is shown in Fig. 4(a). From the Fig.
C. Findings/rom Same Wattage Lamps
4 (a) it is clear that the current waveforms of different wattage

839
 These tests aim to identify the harmonic levels from same is also noted that different manufacturers of LED lamps use
wattage lamps introduced by different manufacturers. For this diverse ballast technologies to reduce harmonic generation
purpose 4 W, 5 W and 7 W LED bulbs are investigated. which includes ballast with passive filters, active filters, and
Figure 6 depicts the wave forms obtained from 4 W LED valley-filled circuit. However, some do not use any type of
lamps from Osram and Philips. It can be noted from the figure harmonic filter.
that the current waveform characteristics from these two
ACKNOWLEDGMENT
lamps are quite different. Since the Philips lamp used Valley­
filled configuration, the current peaks observed for the case of This work was carried out with the financial support from
Philips lamp is about 50% lower than that required by Osram the MOHE under the research grant UKM-KK-02 FRGSO 193-
lamp. Therefore it can be concluded that the 4 W Philips LED 2010.
bulbs is better than the Osram 4 W.
REFERENCES
0.4 400 [1] C. DiLouie, Advanced Lighting Controls: Energy Savings,

�:; : = =� = = ===== �:�:: ��

-Osram4W
Productivity, Technology and Applications, 1st ed., Lilburn, GA:
$ ��� 200 � The Fainnont Press, 2005.
B. 1. Huang, P. C. Hsu, M. S. Wu and C. W. Tang, "Study of
=
0
/��� Voltage
0
�
[2]
system dynamics model and control of a high-power LED

� -0.1 --------------- �� �� 1
· � -"�2 '=- � lighting
2007.
luminaire," Energy, vol. 32, pp. 2187-2198, Nov.

u :�:� = = = = = = = = = = = = = = = = = :'S- j = = =2< = - 200 >

[3] (2009) LED professional Review website. [Online]. Available:
http://www.led-professional.coml
V. �uk, l.F.G. Cobben, W.L. Kling and R.B. Timens, "An
-0.40 5 10 15
00 [4]
204 analysis of diversity factors applied to hannonic emission
Time (ms) limits for energy saving lamps," 14th International
Conference on Harmonics and Quality of Power (ICHQP'I 0).
Fig. 5 Current and voltage wavefonns of Osram 4 W, 6 W, 8 W, lOW LED 2010, pp. 1-6.
lamps
[5] R. Dwyer, A. K. Khan, M. McGranaghan, 1. Tang, R. K.

0.2 ,------c--,---�--_____,
McCluskey, R. Sung, and T. Houy, "Evaluation of harmonic
impacts from compact fluorescent lights on distribution systems,"

$ 0.1
IEEE Transactions on Power System, vol. 10, pp. 1772-1779, Nov.
1995.
[6] P.W. Tam , S.Y.R. Hui and S.H. Chung, "An Analysis and
Practical Implementation of a Dimmable Compact Fluorescent
Lamp Ballast Circuit Without Integrated Circuit Control,"
37th IEEE Power Electronics SpeCialists Conference
(PESC'06), 2006, pp. 1-8.
[7] Xiaohui Qu, Siu-Chung Wong and Chi K. Ts e, "Resonance
5 10 15 assisted buck converter for offline driving of power LED
Time (ms) replacement lamps," IEEE Transactions on Power Electronics,
vo1.26, pp. 532-540, Feb 2011.
[8] Linlin Gu, Xinbo Ruan, Ming Xu and Kai Yao, "Means of
Fig. 6 Comparison of harmonic levels from 4 W LED lamps from Osram
eliminating electrolytic capacitor in AC/DC power supplies for
and Philips: Current and voltage waveforms
LED lightings," IEEE Transactions on Power Electronics, vo1.24,
pp. 1399-1408, May 2009.
To further illustrate the performance of same wattage LED S. Muthu, F.J. Schuunnans and MD. Pashley, "Red, green, and
[9]
lamps from different manufacturers, 5 W and 7 W LED lamps blue LED based white light generation: issues and control," 37th
from Philips are compared Bright and Cash Brand bulbs of IEEE Annual Meeting on Industry Applications (IAS'02), 2002, pp.
327 - 333.
M.A. Reyes, S. Gallagher and 1. Sammarco, "Evaluation of visual
same wattages. From investigation, it is very clear that Philips
[10]
bulbs perform much better than the other compared branded performance when using incandescent, fluorescent, and LED
lamps. This is because Bright and Cash brands do not machine lights in mesopic conditions," Annual Meeting of IEEE
incorporate any type of harmonic elimination circuit in their Industry Applications Society (IAS'09), 2009, PP 1 - 7.
[11] M. Blanco and E. E. Parra, "Effects of High Penetration of
ballast design.
CFLs and LEDs on the Distribution Networks," 14th
International Conference on Harmonics and Quality of Power
V. CONCLUSIONS (ICHQP'IO), 2010, pp. 1-5.
This paper has presented several experimental results on [12] A. A. M. de Oliveira, T. B. Marchesan, A. Campos and R. N. do
Prado. "Distributed Emergency Lighting System LEDs Driven by
harmonic generation from LED lamps that are currently being
Two Integrated Flyback Converters," 42 nd Annual Industry
used for domestic, commercial and industrial lighting. In the Applications Conference (IAS'07), 2007, pp. 1141-1146.
experiments various types of LED lamps from different [13] H. Shared: A. Mohamed and N. Marzuki, " Analysis of Ride
manufacturers were tested to evaluate their harmonic Through Capability Of Low-Wattage Fluorescent Lamps During
Voltage Sags," International Review of Electrical Engineering,
performance in terms of power rating, brand, and type of
vol. 4, pp. 1093 - 1101, Sep-Oct 2009.
ballast used. From the experiments it was found that all LED [14] IEEE recommended practices and requirements for harmonic
bulbs generate harmonics due to the use of power electronic control in electrical power systems, IEEE std. 519-1992,
converter as a ballast to drive LED arrays in the bulbs. The 1993.

tested lamp THDJ values ranges between 30.94 and 174.38. It

840