An Analysis of Harmonic Diversity Factors
Applied to LED Lamps
Sahel Uddin, Hussain Shareef, Member, IEEE, Azah Mohamed, Senior Member, IEEE, and M A
Hannan, Member, IEEE
Abstract--This paper presented an investigation of diversity
factors for Light Emitting Diode Lamps (LEDs) which are used
in residential and commercial applications as energy efficient
lighting bulbs. It is done by performing experimental tests on
various LED lamps available in the market. The current drawn
by several LED lamps are first measured to investigate the
electrical characteristics. These measurements provide current
harmonic characteristic of individual lamps. Then various
configurations of LEDs were connected to a laboratory scale
feeder to analyze the diversity factors. The results show that
LED bulbs produce considerable amount of harmonics. However,
various combinations of LED lamps can damp harmonics
penetration into the distribution system.
Index Terms--Light Emitting Diodes (LEDs), Harmonic
Distortion, Diversity factors, Power quality.
I. INTRODUCTION
ECENT research reports highlight that, Light Emitting
RDiode Lamps (LEDs) has almost 2 times better efficiency
compared to Compact Fluorescent Lamps (CFLs) and 8 t010
times more efficient then incandescent lamps. Also the life
time of LED is much longer than other types of lamp and they
are environmentally friendly due to absence of mercury inside
the LED lamps.
However, like CFLs, a compact AC/DC converter should
be used in the lighting fixture to supply DC current to LED
chips which introduce nonlinearity to the system. Due to the
non-linear characteristic, LED bulbs produce highly distorted
currents [1]. This distorted current is injecting into the power
system network. Although the input power of a single LED
bulb is quite low, a large number of customers using LED
bulbs and CFLs per premises could create significant power
quality problems [2].
This work was carried out with the financial support from the MOHE
under the research grant UKM-KK-02-FRGS0193-2010.
Sohel Uddin is with the Department of Electrical, Electronic & Systems
Engineering, Universiti Kebangsaan Malaysia (UKM), Malaysia, (e-mail:
sohel_091@yahoo.com).
Hussain Shareef is with the Department of Electrical, Electronic &
Systems Engineering, Universiti Kebangsaan Malaysia (UKM), Malaysia, (e­
mail: shareef@eng.ukm.my).
Azah Mohamed is with the Department of Electrical, Electronic &
Systems Engineering, Universiti Kebangsaan Malaysia (UKM), Malaysia, (e­
mail: azah@eng.ukm.my).
M A Hannan is with the Department of Electrical, Electronic & Systems
Engineering, Universiti Kebangsaan Malaysia (UKM), Malaysia, (e-mail:
hannan@eng.ukm.my).
A large number of studies were conducted on LEDs as an
energy efficient lamp but most of the researchers pay their
attention on the internal ballast circuit design and enhancing
their performance [3-5]. Several other researchers have
concentrated on light distribution and visual performance of
LED lamps [6-8]. A few contributions focus on harmonic
emissions of LEDs lamps [1], [9]. It is found that most of the
new design of LEDs with lower power rating «25 Watt) have
a power factor (PF) up to 0.6 and current total harmonic
distortion (THD,) between 100-140 % [1], [9]. Even if the
individual effect is very small on a distribution feeder, a large
number of LED lamps connected to a single feeder may
introduce considerable harmonic current distortion and power
quality problems. But it should be noted that arithmetic sum of
individuals may lead to wrong estimation of current harmonic
distortion levels. Therefore it necessary to measure the
diversity factors to know the impact of large number of LED
lamps on power quality. Diversity factor is defined as [10],
Vector sum of current harmonic (as measured)
DF (1)
n = Arithmetic sum of current harmonic (as calculated)
where DFn is the Diversity factors.
This paper investigates the diversity factors of LED lamps
without other non-linear loads. This is characterized by
measurement tests, using various available LED bulbs in the
market. In the investigations, a total of 35 pieces with different
ratings from 7 different manufactures (brands) are observed.
The electrical tests are carried out to observe their current and
voltage waveforms. The characteristics are also compared for
their performance with IEC 61000-3-2 harmonic standard.
II. HARMONIC STANDARD OF LED LAMPS
Similar to any other appliance, LED lamps also must
comply with several directives which are applicable to the
product. The IEC 61000-3-2 standard assesses and sets the
limit for equipment that draws input current :S16A per phase
[11]-[12]. Harmonic emission limits for lamps are subdivided
based on their active power up to 25W and above. Lamps
having an active input power less than or equal to 25W must
satisfy at least one out of the two following criterions. One of
the criteria is that the third harmonic current should not exceed
86% of the fundamental and the fifth harmonic current should
978-1-4673-2868-5/12/$31.00 ©2012 IEEE
 2

not exceed 61%. That gives the value of the current THD on for 10 minutes before the measurements are taken for
approximately 105%. The recommended voltage distortion stabilization. Each lamp is tested for four times to eliminate
limit for class C equipment is 3% and 5% for individual any error during different period of the day. The captured
harmonics and total harmonic distortion (THDv) respectively. current waveforms were analyzed by using Fluke 434 and
The other criterion is given as a Table I for each harmonic MATLAB software. After individual characterization of
order, where limits are expressed as percentage of harmonic spectrum, (1) is used to obtain the results.
fundamental.
TABLE II
TAB LE T TECHNICAL DATA FOR TESTED LAMPS
lEC 61000-3-2 LiMITS FOR CLASS C EQUIPMENT (P -
< 25W) Trade Nominal Power Life
Name Power Factor Span
Harmonics Class C
P (W) (PF) (Years)
(n) (% of fundamental) Philips Cool
4 0.69 25
Daylight
3 30 x PF
Philips Warm
5 0.71 25
5 10 Daylight
Philips Cool
7 7 7 0.72 25
Daylight
9 5 Osram Daylight 4 0.48 25
Osram Warm
II 3 4.5 0.47 25
White

13 3 Osram Daylight 6 0.5 25

Osram Warm
8 0.79 25
15:=;n:=;39 3 White
Toshiba Cool
5.5 0.58 40
White
Bright cool White 5 0.46 6
III. EXPERIMENTS Cash Warm White 7 0.48 20
To analyse the characteristics of the LED lamps, 12 Evenzo Cool White 3 0.47 18

samples with different power ratings from various

manufacture, as shown in Table II, were tested for harmonics.
The lamps have build in ballast which is powered using E-27,
E-14 or GU-I0 type sockets, commonly available in retail
stores. All the tested lamps are designed to operate at 220-240
V and have power consumptions rating of 3 W to 8 W.
Computer for Data
Download
IV. RESULTS
In this section, measurements of various test conducted on
LED lamps are analyzed and discussed. The lamp current
waveforms were analyzed using the Fourier Theorem. It
provide frequency spectrum of the lamp currents represented
by the fundamental sinusoidal component and a series of
higher order harmonic components at frequencies that are
integer mUltiples of the fundamental frequency. The square
roots of the sum of the amplitudes of the harmonic present the
total harmonic distortion (THD).

A. Harmon ic Characteristic a/In dividual Lamps

Lamps Under Test
PQ Analyzer Fluke
434
Fig. I. Experimental setup for harmonics
To obtain accurate data concerning the exact current
harmonic content of LED bulbs, an experimental setup as
shown in Fig. 1 is assembled. It consists of four components
namely, Fluke 434 power quality analyzer, Fluke i30s current
clamp, LED bulb(s) under test, and a personal computer to
analyze the signals. Before conduction with group of lamps,
measurement on individual lamps was performed to obtain
harmonic and other characteristic. Each lamp is kept switched
The purposes of these tests are to get an insight about the
harmonic generation levels from different LED lamps. The
typical current waveforms illustrated in Fig. 2 are obtained
from different lamps tested. From the figure it can be noted
that the current waveforms is not sinusoidal. It means that they
inject harmonic into the power system. Furthermore the figure
also indicates that dissimilar LED bulbs use different filtering
methods to reduce harmonic generation. For example, most of
the Phillips brand LED bulbs tested utilize the Valley-filled
circuit, the Toshiba 5.5 W lamp contains a passive filter,
Osram 8 W sample embed an active filter, and some other
tested lamps does not implement any filtering technique.
To demonstrate further, relative harmonic currents of all
tested lamps are presented in Table III. Note that some of the
tested LED bulbs generate quite a high level of harmonic
which produce unacceptable limits of harmonics when
referred to IEC 61000-3-2 standard described earlier. From
Table III, it can be noted that lamps having the trademark of
 3

Philips, have THDI values between 63 %-65 %, while Osram 0.2,.------=-�----� ---�_,___:-�

brand LED bulbs with power rating less than 6 W produce

high levels of harmonics which is in the range of 173 to 175 % :'.
.5
THDI. This is because these LED bulbs do not use any filter in -=
·200 '
their ballast circuit. However, in case of 8 W Osram LED
----::':: 1 --:---=--:"15,--
- --20400
lamps, they produce the lowest THDI which is in the rage of
-0.2;;---
0
------;:-
5 0
time (ms)
30 to 35 %. (a)
�IOO�I-,-����������-,-���-�
_Philips4W
Philips 5 W
_ I
o.4 ,------,---r-- .� 75 ,'i
o C-:-IPhili s7W
0.3
-Nofilte..ing
---Valley fill
E I "
"'tl �
•• 50 ,

0.2 - Passive filtering .:

= ,
$ 0.1
_.- Active filtering
�
a
25 I
,I 'I
'LI'
-J'U'L
,
J"I..... /.J IJ..., I".a ,lJ lLJ ill ilJ L.J iLJ 1 1..1"J
L
,I ,
tU Ii_H...I
� Ol ������--�����������
� 0
1 3 5 7 9 11 13 15 17 1921 23 25 27 29 31 33 35 37 39

; -0.1 harmonic order (n)

(b)
-0.2
"

-0.3 Fig. 3. Test results of Philips 4 W, 5 W, 7 W LED bulbs: (a) Lamp current
-0. and voltage waveforms (b) Individual harmonic spectrum
40�------:-5-----:"10,--
- ---IL5-------20-l
time (ms)
o.4 400
-Osram4W
---Osram6W
Fig. 2. Various types of current waveforms obtained from different tested � 200 �
A,\ ...... Osram8W
0.2
LED bulbs Voltage
�t t
=

o
� "-
TABLE III U -0.2 200 g
.r
-

HARMONIC CONTENTS OF SINGLE LED LAMPS

-0.4 5 10 15
0
-

Tested Harmonic (%) Time (ms)

d �
(a)
Lamp Fund 3' 5'h 7'h THO, --�o-sc-am-4-,w
�00 �-,-�--,-������ ��-,--,-r...

Philips 4 W 100 34.61 6.28 22.27 63.05

'E 75 : _Osram6W
c=]Osram8W
E I

Philips 5 W 100 36.92 7.16 19.81 63.83

� 50 :
,
-
I
,
� 25
Philips 7 W 100 32.34 11.6 23.4 64.23 �
U O �I������
Osram 4 W 100 89.24 70.65 51.19 173.92

Osram 4.5 W 100 91.72 77.47 59.43 157.76

Osram 6 W 100 91.96 77.05 58.51 174.37 Fig. 4. Test results of Osram 4 W, 6 W, 8 W LED lamps: (a) Lamp current
and voltage waveforms (b) Individual harmonic spectrum
Osram 8 W 100 22.25 15.04 2.34 30.94

Evenzo 3 W 100 90.7 77.34 58.89 164.45 B. Harmon ic Characteristic ofMultiple Lamps
Bright 5 W 100 86.65 75.92 61.75 167.25 The aim of these tests is to investigate the effect of
Cash 7 W 100 91.23 74.81 56.41 168.22 harmonic characteristic from group of LED lamps. For this
purpose, the diversity factors of various groups of lamps are
Toshiba 5.5 W 100 73.4 45.4 35.62 106.34
measured. A small value of diversity factor indicates that a
significant amount of cancellation occur because of
superposition of individual current wave shapes. The diversity
The current wave form and harmonic spectrum of tested
factors of different group of LED lamps are presented on
Phillips and Osram lamps are shown in Figs. 3 and 4
Table IV. From Table IV, note that Combination-A have a
respectively. From Fig. 3, it is clear that the current
diversity factor of 0.83 while combination-D shows much
waveforms of different wattage lamps have same
lower diversity factor which is equal to 0.72. That means
characteristics although the magnitude of current increases
combination-A with 3 lamps produces more harmonic than
with the increase of power rating. This also observed for the
Combination-D with 7 lamps. This shows that the diversity
case of Osram lamps in Fig. 4 if the lamp uses same ballast
factor decreases with increasing number of lamps.
circuit. Fig. 4(b) depicts the harmonic spectrum of these
To further investigate, diversity factors of individual
lamps. It is noticed that the magnitude of harmonic current is
harmonic orders are plotted for Combination-A and D as
decreased with increasing harmonic order and their
shown in Fig. 5 and 6. In the case, only odd order harmonic
magnitudes are comparable with each other for different
are presented because even order harmonic are almost zero.
wattage lamps. All other tested lamps have similar observation
From Fig. 5, it can be noted that diversity factor for all
and their THD, value greater than lOO as seen in Table III.
harmonic orders are greater than 0.5 for Combination-A while
for the Combination-D shown in Fig 6, diversity factors for
individual harmonic order is much lower and falls below 0.5
 4

starting from 19th order harmonic. Therefore, Combination- A

is more problematic than Combination-D. TABLE V
HARMONIC CONTENTS OF GROUP OF 5 W LED LAMPS FROM PHILIPS

TABLE IV No. of Power Harmonic (%)

DIVERSITY FACTORS OF SEVERAL GROUPS OF LED LAMPS d th th
Diversity Lamps P (W) Fund 3' 5 7 THO,
Combination Lamps
Factors
1 5 100 35.3 11.0 18.5 64.1
Philips 4 W, Osram 4 W,
A 0.83
Bright 5 W 2 10 100 35.7 11.2 17.0 64.5

Philips 7 W, Osram 4.5 3 15 100 35.0 12.1 18.7 62.3

B W, Osram 6 W, Bright 5 0.77
W, Evenzo 3 W

Philips 7 W, Osram 4.5 TABLE VI

C W, Bright 5 W, Cash 7 0.78 HARMONIC CONTENTS OF GROUP OF 4 W LED LAMPS FROM OSRAM
W, Evenz03 W No. of Power Harmonic (%)
Philips 7 W, Osram 4.5 d th th
Lamps peW) Fund 3' 5 7 THO,
W, Osram 6 W, Bright 5
0 0.72
W, Toshiba 5.5 W, Cash I 4 100 91.4 74.2 54.9 174.1
7 W, Evenzo 3 W
2 8 100 92.1 77.8 59.9 174.6

3 12 100 89.1 70.8 51 173.2

...
� 0.8
'" V. CONCLUSION
...
cO.6
.� This paper has presented several experimental results of
� 0.4
current harmonic distortion and analyse the effect of diversity
factors for energy efficient LED lamps. In the experiments
=
'"
� 0.2 various types of LED lamps from different manufactures were
tested. It is found that LED lamps diversity factors become
o 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 smaller if larger amount of LED lamps from various brands
Harmonic Order (n) are connected. Low diversity factor indicates low injection to
the system. However, it is found that group of lamps with
Fig. 5. Diversity factors of different harmonic orders for Combination-A
same model will not reduce diversity factor and harmonic
distortion. Therefore it is recommended to use various types of
... LED lamps together in lighting applications to reduce power
� 0.8
quality problems.
'"
...
CO.6
.� VI. REFERENCES
� 0.4 [1] M. Blanco and E. E. Parra, "Effects of High Penetration of CFLs and
= LEOs on the Distribution Networks," in Proc 2010 the 14th int. Con!
'"
� 0.2 on Harmonics and Quality of Power (iCHQP), pp. 1-5.
[2] R. Dwyer, A. K. Khan, M. McGranaghan, L. Tang, R. K. McCluskey,
0�-- � --� �- -- Ll� I- 3-- L 5 - -7�IL �L 3- -5�L 7- -9u R. Sung and T. Houy, "Evaluation of harmonic impacts from compact
3 5 7 9 1 1 1 9 -2-1 2 2 2 2
fluorescent lights on distribution systems," iEEE Trans. Power
Harmonic Order (n)
Systems, vol. 10, pp. 1772-1779, Nov 1995.
[3] X. Qu, S. C. Wong and C. K. Tse, "Resonance assisted buck converter
Fig. 6. Diversity factors of different harmonic orders for Combination-D for offline driving of power LED replacement lamps," iEEE Trans.
Power Electron, vol. 26, pp. 532-540, Feb 2011.
A second test is conducted to observe the effect of same [4] X. Qu, S. C. Wong and C. K. Tse, "Electronic ballast for mUltiple LED
lamps with independent brightness control. ," in Proc 2009 iEEE int.
wattage LED lamps with same trademark is connected in
Symposium on Circuits and Systems (lSCAS), pp. 2878-2881.
groups. Tables V and VI show the results of a group 5 W and [5] L. Gu, X. Ruan, M. Xu and K. Yao, "Means of eliminating electrolytic
4 W LED lamps from Phillips and Osram. From Tables V and capacitor in AC/DC power supplies for LED lightings," iEEE Trans.
VI it can be observed that, there is no effect of adding more Power Electron, vol. 24, pp. 1399-1408, May 2009.
[6] S. Muthu, F. J. Schuurmans and M.D. Pashley, "Red, green, and blue
LED lamps in the circuit on harmonic generation. This is LED based white light generation: issues and control," in Proc 2002 the
because the wave shapes generated by adding similar lamps is 37th Annual Meeting on industry Applications Con! (lAS), pp. 327-
unaffected except the magnitude of current drawn from the 333.
[7] I. Kiyak, V. Topuz and B. Oral, "Modeling of dimmable High Power
system. Similar observation is obtained when the lamps with LED illumination distribution using ANFIS on the isolated area,"
different power is combined from the same manufacturer Expert Systems with Applications, vol. 38, pp. 11843-11848, Sept 2011.
considering the type of ballast used. As a result, diversity [8] M. A. Reyes, S. Gallagher and J. Sammarco, "Evaluation of visual
performance when using incandescent, fluorescent, and LED machine
factor is not affected for those cases and the diversity factor lights in mesopic conditions," in Froc 2009 Annual Meeting of iEEE
will be close to 1. industry Applications Society (iAS), pp. 1-7.
 5

[9] V. Cuk, J. F. G. Cobben, W. L. Kling and R. B. Timens, "An analysis of

diversity factors applied to harmonic emission limits for energy saving
lamps," in Froc. 2010 the 14th 1nt. Con! Harmonics and Quality of
Fower, pp. 1-6.
[10 ] A. Mansoor, W. M. Grady, A. H. Chowdhury and M. 1. Samotyj, "An
investigation of harmonics attenuation and diversity among distributed
single-phase power electronic loads," iEEE Trans. Power Delivery, Vol.
10, pp. 467-473, Jan 1995.
[II] iEEE recommended practices and requirements for harmonic control in
electrical power systems, IEEE Standard 519-1992, April 1993.
[12 ] Limits for harmonic current emissions (equipment input current � 16A
per phase), IEC Standard 61000-3-2, Ed. 3.2, May 2009.