Content uploaded by Jiri Drapela
Author content
All content in this area was uploaded by Jiri Drapela on Jul 01, 2014
Content may be subject to copyright.
ELECTRIC POWER ENGINEERING 2010
A Light-flickermeter – Part I: Design
Ji?í Drápela1), Jan ?lezingr2)
Department of Electrical Power Engineering, Brno University of Technology, Technická 8, 616 00 Brno, Czech Republic,
www.feec.vutbr.cz/UEEN
1) tel: +420 5 4114 9234, email: drapela@feec.vutbr.cz ,
2) tel: +420 5 41149292, email: xslezi01@stud.feec.vutbr.cz
ABSTRACT
Voltage variation is one of lamps flickering main cause;
thus evaluation of a disturbing light flicker is a part of volt-
age quality classification. The instrument for flicker level
measurement by means of voltage variation analysis is
named flickermeter and is specified in standard EN 61000-
4-15. The flickermeter is a specialized analyzer modeling
response of a chain consisting of reference 60W incandes-
cent lamp – eye – brain of an average observer. Hence,
other lamp types like discharge are mainly used in illumina-
tion systems at present, significant difference between
flicker measurement using the standard flickermeter and
observation can appear. In the paper, there is proposed a
light flicker measurement method based on analysis of an
irradiance (radiant flux) waveform, where the output flicker
index is independent on the lamp type. With regard to ex-
pected properties the proposed light-flickermeter can be
called objective flickermeter. Nevertheless the range of
application of the objective flickermeter is limited as it is
also shown.
Keywords: Voltage fluctuation, Interharmonics, Light
flicker, Light-flickermeter, Matlab Simulink
1 INTRODUCTION
As it is known, the visible light variation may leads to dis-
turbance of human’s visual perception. The origin of such
negative effect is in the eye anatomy and physiology [7]. The
major influence on the flickering light perception can be
found in eye viewing field, photoreceptors (cones of three
types and rods) and their distribution at retina as a part of eye
anatomy and in eye adaptation mechanisms like pupil, photo-
chemical and neural adaptation (response) to luminous varia-
tions (including the photoreceptors spectral luminous effi-
ciency [15]) as a part of eye physiology [7]. The human eye
can be seen as specialized luminance multi-detector. Never-
theless the stimulation contrast is as important as radiant
density. Thus visual perception as well as flicker perception
can be specified on the basis of following parameters of vis-
ual stimuli: object and background luminance and luminance
nonuniformity, contrast, object and background light spec-
trum, object size, object location compared with main axis of
view, and mainly in point of view of flicker perception: time
changes in luminance and color of stimuli, and movement or
spatial changes of stimuli in viewing field, etc. [7] [15]. On
the basis of the described perceived luminous variations,
there are differentiated three types of flicker: temporal, spatial
and chromatic [7][8]. All of these flickers are joyless and may
involve many psychological interactions.
In artificially illuminated areas, lamps light variations due
to variations in supply voltage may also lead to flicker per-
ception. Such lamp possibly will produce light with time-
varying radiant flux and its spectral distribution. As the prin-
cipal cause of lamps light flicker, there is viewed, for histori-
cal reasons, a voltage amplitude modulation (AM) by a
periodical and/or stochastic signal due to varying loads. Nev-
ertheless the lamps flickering is produced also by a phase
voltage modulation, mainly by phase jumps [12] and by the
interharmonic voltages superimposed on a voltage waveform
[11][3]. A lamp ability to transfer such changes in supply
voltage on generated luminous time-spectrum changes is
given by its transfer function which depends on the lamp
function principle and type as well as on the wiring, circuit
designing and dimensioning [2].
From the point of view of the perception of a lamp light
variation by a human eye, changes within the range of (0.1 ÷
35 (40)) Hz are significant. And the visual perception is dis-
turbed in a case when the magnitude of variation exceeds a
corresponding threshold value within this frequency range. It
stands to reason that the disturbing level of initiating voltage
variations depends on the lamp type, observer, and on given
observation conditions. Since the flicker is matter of voltage
variation, the flicker severity is currently evaluated by means
of voltage measurement using a specialized voltage analyzer
called flickermeter. The flickermeter is based on modeling of
the reference 60W incandescent bulb – human eye – brain of
average observer chain response obtained under specific
observation conditions. The EN standard for flicker meas-
urement is published in [5] and provides design specifications
for an analogue flickermeter, including validation methods.
Taking into account the standard flickermeter design specifi-
cation, the flicker severity measurement results and observa-
tion can differ significantly in a case of different lamp for
illumination will be used and/or observation conditions will
not be the same as reference and/or the observer response will
be different from the eye–brain response of the average ob-
server.
A possible solution in the first step is to exclude the light
source form the flicker severity evaluation process and to
measure the flicker severity directly by means of light meas-
urement and eye-brain of the average observer response
simulation [6][14], thus a light-flickermeter is going to be
realized. In the paper, the light-flickermeter topology is de-
signed including proposal of light measurement technique
DR
á
PELA, J.;
?
LEZINGR, J. A Light - flickermeter -
Part I: Design. In Proccedings of the 11th International Scientific
Conference Electric Power Engineering 2010. 1. Brno: Brno University of Technology, FEEC, DEPE, 2010. pp. 453-
458. ISBN: 978-80-214-4094- 4.
and corresponding signal processing on the basis of the stan-
dard UIE/IEC flickermeter, in the light of current and future
technologies towards the light flicker objective measurement.
Moreover a model of a simplified light-flickermeter for cer-
tain applications in Matlab Simulink is realized and prelimi-
nary verification results in terms of numerical simulations are
shown.
2 LIGHT-FLICKERMETER
The light-flickermeter is designed to have two main parts
which are light measurement and consequent luminous varia-
tion signal processing. Assuming that the human eye is a
luminance detector the spectral radiance time behaviour has
to be measured then the spectral luminous efficiency of ele-
mentary photoreceptor can be modeled. Since the eye percep-
tion system is composed of really large number of
photoreceptors of fourth types (long, middle and short cone
and rod) [7] each of them or sufficient number of them has to
be represented by correspondent spectral radiance detector
including the detectors surface distribution as it is across the
retina and including an optical system mediating visual stim-
uli from all viewing field to corresponding segment of the
detectors. Unfortunately realization of such multi-detector
modeling in fact particular anatomy of the human eye is a
technological problem at present. Technically realizable there
is a detector containing four elementary detectors each repre-
senting one of the basic photoreceptors, but typically a single
detector is available. Furthermore the detectors output signals
processing modeling all the described adaptation mecha-
nisms, the eye physiology in fact, taking into account also
neural interconnections in visual system [15] is necessary to
perform. Any deviation from introduce measurement system
makes the light flicker measurement non-objective.
In a case the objective light flicker measurement is not
required, the light flickermeter is suppose to be used for com-
parative measurements only then a significant simplification
in measuring process can be made. Considering that the light
flicker measurement result is subject to the same observation
condition whereat the reference eye-brain response of aver-
age observer was obtained, a single spectral radiance or ir-
radiance detector and signal processing based on the standard
flickermeter can be used. Another simplification consists in
presumption that the spectral luminous efficiency of the aver-
age observer is invariable and is equal to the spectral lumi-
nous efficiency of the normalized observer for photopic
vision. Then luminance or illuminance detector with built-in
special light filter can be introduced instead of spectral meas-
urements. In the same time it means that the eye adaptive
mechanisms are supposed to be dependent on only on the
light intensity, not at the light spectral distribution. Because
comparative measurement also implies that just relative varia-
tions in light are significant for light flicker measurement, the
luminance or illuminance can be substituted by luminous flux
in connection with lamps flickering. Resulting simplified
block diagram of the light-flickermeter flowing from the
standard flickermeter block diagram [5] is at Fig. 1.
The light-flickermeter architecture can be divided into
three parts:
– light measurement provided by block 1,
– eye-brain response modeling implemented in blocks
2, 3 and 4,
– statistical analysis of the instantaneous flicker read-
ings in real time realized by block 5.
light
measurement Signal conditioning
and normalization
Block 1 Block 2
Φ
e(
λ
,t) ~
Φ
(t)
Φ
(t)
Φ
DC
~
Filtration Squaring and
smoothing Statistical
evaluation
Block 3 Block 4 Block 5
P
f
(t)FS(t) P
st
Fig. 1. A simplified block diagram of the light-flickermeter
Block 1 – performs luminance or illuminance measurement
using a detector with appropriate characteristics.
Block 2 – carries out signal normalization without modifying
the resulting relative fluctuation. The output signal
from block 1 is normalized by its smoothed DC
value.
Block 3 – also consists of a cascade of three filters as well as
in case of the standard flickermeter [5]. The first it
is the high pass filter for DC component elimina-
tion. The 6th order Butterworth low pass filter re-
moving frequency components at frequencies over
35 Hz is the second. As the last one, there is the
weighting filter modeling the eye-brain response
of average observer.
Blocks 4 and 5 – are identical to the standard UIE/IEC flick-
ermeter.
2.1 Luminous measurement
For a case of light detection, utilization of a silicon photo-
diode (solid-state detector) having appropriate spectral re-
sponsivity, low noise, good stability and detectability is the
best solution suitable for extremely accurate measurements.
In consonance with previous, the basic detector has to be
equipped by additional light filter ensuring that the detector
overall spectral responsivity will be the same as it is shown
on the Fig. 2. Then the sensor becomes to be an illuminance
detector.
0
0.2
0.4
0.6
0.8
1
400 450 500 550 600 650 700 750
Wavelangth λ (nm)
Normalized spectral response v(
λ) (-)
Fig. 2. Normalized spectral luminous efficiency for photopic
vision
Another option is to use a spectroradiometer where the
output light spectrum will be weighted by spectral luminous
efficiency for photopic vision Fig. 2. However usage of a
spectroradiometer for such instance (to measure illuminance)
is inefficient. The spectroradiometer utilization is substanti-
ated in the case when the output spectrum will be decom-
posed to get the spectrum parts perceived by each type of the
eye photoreceptors, thus it is possible way to evaluate also
influence of variations in light spectrum on the flicker percep-
tion [15]. Nevertheless adaptation mechanisms for this de-
composed scheme are not sufficiently described at this time.
When a silicon photodiode is subject to light, a photocur-
rent IP is generated. The photodiode signal can be measured
as a voltage or a current. Current measurement demonstrates
far better linearity, offset, and bandwidth performance. The
photodiode is considered linear, as it is needed, if the gener-
ated photocurrent increases linearly with the incident light
power. Then the photocurrent at a given light wavelength is
[9]:
()
(
)()
λρλλ
λλλ
?Φ= eP
I (1)
where
()
λ
λ
e
Φ is the incident spectral radiant flux on equiva-
lent active surface of the detector SD and
()
λ
ρ
λ
is the detec-
tor spectral responsivity. If the detector is equipped by
photopic correction for illuminance measurement, the spec-
tral responsivity is in shape of V(
λ
) curve, thus the generated
photocurrent is proportional to detector illumination (if the
size of SD is suitable) and is as:
(
)
λλ
λ
λ
dII PP ∫
= (2)
The linearity is in principle ensured if the photodiode load
impedance is becoming to zero. Utilization of a linear meas-
uring operational amplifier in transimpedance configuration
seems to be the best measurement option from many reasons.
A scheme of a solid-state photo-detector with an operational
amplifier with single stage amplification which is reasonable
for intended application is at Fig. 3.
Op-amp
+
-
Photo-
Detector
+Ucc
-Ucc
U
A
PV
PC
R
F
C
F
C
A
C
J
R
Sh
R
S
I
P
Fig. 3. Scheme of a solid-state photo-detector with an opera-
tional amplifier
The scheme at Fig. 3 makes it possible to operate the pho-
todiode in Photo-Voltaic (PV) or Photo-Conductive (PC)
mode. The PC mode is set when a reverse bias is applied and
it can significantly improve chosen sensing parameters like
the speed of response and linearity. The circuitry also in-
cludes (parasitic) elements of equivalent diagram which are
essential for the Photodiode-Amplifier (PA) combination
performance formulation [13]. The shunt resistance RSh is
used to determine the noise current in the photodiode with no
bias (photovoltaic mode). For best photodiode performance
the highest shunt resistance is desired. The series resistance
RS of a photodiode arises from the resistance of the contacts
and the resistance of the undepleted silicon. Series resistance
is used to determine the linearity of the photodiode in PV
mode. Although an ideal photodiode should have no series
resistance. Connection parameters between the diode and
amplifier can also contribute by series resistance and induc-
tance causing impairment of linearity and introducing possi-
ble signal oscillations. For that reason the connection should
be as close as possible to prevent. The boundaries of the
depletion region act as the plates of a parallel plate capacitor.
The junction capacitance CJ is directly proportional to the
diffused area size SD and inversely proportional to the width
of the depletion region. In addition, higher resistivity sub-
strates have lower junction capacitance. The larger detection
area SD photodiode has the higher sensitivity and capacitance
CJ are. Furthermore, the capacitance is inversely dependent
on the reverse bias voltage magnitude. The junction capaci-
tance is used to determine the response of the photodiode.
The resistance RF in the amplifier feedback determines di-
rectly the amplifier gain and indirectly amplifier bandwidth.
Then the amplifier output voltage UA is as:
FPA RIU ?= (3)
Value of RF has also effect on the amplifier thermal noise
current whereas higher resistance is desired. The frequency
response of the photodiode-amplifier combination is deter-
mined by the characteristics of the photo-detector, pre-
amplifier as well as the feedback resistor RF and feedback
capacitor CF which limits the frequency response to avoid
gain peaking and to prevent instability. Cut-off frequency of
the detector-amplifier combination frequency response set by
the amplifier feedback is given by eq.:
FF
ACR
f??
=
π
2
1 (4)
However, the desired frequency response is limited by the
Gain Bandwidth Product (GBP) of the op-amp and by spuri-
ous components of the equivalent circuitry at Fig. 3. In order
to have a stable output, the values of the amplifier feedback
have to be chosen such that the fA frequency is less than:
()
JAFF
ACCCR
GBP
f++?
≤
π
2 (5)
where CA is amplifier input spurious capacitance.
In term of the comparative flicker measurement, there is
necessary to choice a Filter-Photodiode-Amplifier Combina-
tion (FPAC) having characteristics allowing accurate meas-
urements of light flickering up to 100Hz and/or of changes in
light with fall/rise time about 200 μs, in illuminance range
from 1 lx to 100 klx with that the maximal output voltage UA
of 10 V seems to be optimal regarding the amplifier power
supply level and output signal processing.
2.2 Signal processing
A block diagram of proposed signal processing of the
FPAC output signal, which is proportional to incident lumi-
nous flux, in blocks 2-5 of Fig. 1 is at Fig. 4.
An input analog signal of optimal magnitude is normal-
ized dividing the signal by its DC component in first two
blocks, since the normalization is also provided by the stan-
dard flickermeter [5]. The DC component is obtained using
the 1st order low pass filter transfer function with analogue
frequency response as follows:
()
1
1
+?
=s
sF
LPSC
LPSC
τ
(6)
where s is a complex Laplace variable and
τ
LPSC is the filter
time constant pre-set to 10 s. Corresponding filter cut-off
frequency is 0.016 Hz. In [5] is defined with a view to pre-
vent from affecting of very slow or isolated changes that the
rise/fall time of the normalization process response on step
change in input signal has to be equal to 1 minute. But it was
proved that such rise time can be shortened to 30 s without
any influence on result. The time constant of 10 s was chosen
with respect to reduction of the filter transition time after
start.
LP filter
F
LPSC
(s)
0.016Hz
~
Φ
(t)
Division
x
÷
F
HP
(s)
0.05Hz
F
LPBW
(s)
35Hz
LP filter
HP filt e r
F
WFO
(s)
Weighting
filter
F
LP
(s)
LP filter
u
2
0.53Hz
S
Scale
Squaring
A/D
Digitizer
{Cl}
Tst
Buffer
Cl(NC,M) P
x
({Cl},{X})
Percentiles
evaluation
Classifier
1600 S/s
P
st
({P
x
})
Flicker index
calculation
f
P
{P
st
}
Tlt
Buffer
P
lt
({P
st
})
Flicker index
calculation
st
Plt
P
Fig. 4. Block diagram of the luminous variation signal proc-
essing
Applying the signal normalization makes the flicker per-
ception measurement independent on illuminance (lumi-
nance) level which is in conflict with observations [7]. So it
should be better to remove this stage in future, but changes in
eye-brain response model have to be performed as well.
Transfer function with analogue frequency response of
the following 1st order high pass filter eliminating DC com-
ponent is as follows:
()
HP
HP
HP s
s
sF
τ
τ
?+
?
=1 (7)
where
τ
ΗΠ
is the filter time constant equal to:
(
)
HP
f
?
=
π
τ
21 ,
and fHP is the cut off frequency of the high pass filter having
attenuation of 3dB at fHP=0.05Hz.
The (pre-warped) 6th order Butterworth low pass filter
transfer function with analogue frequency response can be
given in cascade form as follows:
()
?
?
?
?
?
?
?
?
?
?
?
?
?
?
+
?
?
?
?
?
?
?
?
??
???
+
?
?
?
?
?
?
?
?
??
Π= =
2
2
2
3
1
2
2
2i
LPBW
ii
LPBW
i
i
LPBW
f
s
f
s
sF
ω
π
ω?
π
ω
(8)
where
ω
1=
ω
2=
ω
3=1,
ζ
1=0.26 ,
ζ
2=0.71 ,
ζ
3=0.97 and fLPBW is
the cut off frequency of the low pass filter with attenuation of
3dB at frequency of 35Hz.
The next weighting filter modeling the average observer
eye-brain frequency response is one of the most important
part of the designed light-flickermeter. Its transfer function
with analogue frequency response FWFO(s) can be obtained by
decomposition of known standard flickermeter weighting
filter transfer function FWF(s) using a procedure applied in
[16]:
() ()
()
sF
sF
sF
L
WF
WFO = (9)
where FL(s) is a transfer function with analog frequency re-
sponse of the reference 60W incandescent lamp. The transfer
function of the standard flickermeter weighting filter is de-
fined in [5] and consists of these two parts:
()
?
?
?
?
?
?
?
?+?
?
?
?
?
?
?
?
?+
+
?
++
??
=
43
2
2
1
2
1
11
1
2
ωω
ω
ωλ
ω
ss
s
ss
sk
sFWF (10)
where the transfer function parameters for 230V/50Hz system
are: k=1.74802,
λ
=2π4.05981,
ω
1=2π9.15494,
ω
2=2π2.27979,
ω
3=2π1.22535,
ω
4=2π21.9.
The reference lamp transfer function has been found as
approximation of measured gain factor curves of 10 bulbs
60W/230V from different producers using measurement
method published in [2]. Then the measured curves have been
fitted by 2nd order low-pass filter transfer function as:
()
1
2
2
1+?+?
=ss
K
sF
LL
L
ττ
(11)
where the time constants
τ
L1 and
τ
L2 and the gain K were
obtained using method of least squares and are K=3.57
τ
L1=0.02 ms,
τ
L1=21.2 ms.
After substitution of eq. (11) and (12) into (10) the light-
flickermeter weighting filter transfer function can be ex-
pressed as
()
35053805348201178132.196
298396.2715758.44041661.0
234
234
++++
+++
=ssss
ssss
sFWFO (12)
Comparison of the standard and light-flickermeter weight-
ing filters amplitude response is at Fig. 5.
10
-2
10
-1
10
0
10
1
10
2
10
3
10
4
10
-3
10
-2
10
-1
10
0
Frequency f (Hz)
Relative magnitude |F(f)| (-)
Reference inc.
bulb 230V 60W
Weighting filter of
the standard IEC
flickermeter
Weighting filter of
the Light-flickermeter
Fig. 5. Amplitude response of chosen filters
Than the flicker signal (FS) is squared within the frame of
simulation of the non-linear perception during flicker transfer
from eye to brain (block 4 of the flickermeter) and is filtered
to simulate memorization effect in the brain using the first
order low-pass filter with following transfer function:
()
LP
LP s
s
sF
τ
?+
=1 (13)
The time constant of the filter is
τ
LP=300ms.
The output signal is normalized using scaling factor S to
give a maximum value of instantaneous flicker level index
Pf=1, when a sine wave modulated input signal with funda-
mental frequency f1=50 Hz, the modulation frequency
fM=8.8 Hz and the modulation depth
Δ
U/U=0.25 % [5], is
applied.
In consonance with the standard [5], the analog signal has
to be converted to digital for evaluation in last block. The
required sampling rate is at least 50 Hz. In the proposed in-
strument there is used a 16-bit/ 1.6 kHz digitizer.
The classification routine in the block 5 is based on sort-
ing of instantaneous flicker level samples Pf(k) into logarith-
mically spaced classes using following equation:
() ()
()
()
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?+= M
kP
NC
floorkCl f
10log
10log
1
2 (14)
where NC is total number of used classes and M is a constant
defining the classifier range. The equation (14) provides that
half number of classes from NC is designed for Pf ≤1 and the
second half for Pf>1. The classifier parameters NC=100000
(which is much more than it is required according the stan-
dard [5]) and M=1000000 was chosen (Fig. 6).
Subsequently the classes Cl(k) are recorded for a short
time period Tst into a buffer. The short time period is set to 10
minutes, but can be changed to another time interval. At the
end of each successive Tst period a Cumulative Probability
Function (CPF) is calculated by means of the number of
samples to fall into each class. The percentiles defined in [5]
are searched consequently as class numbers of the flicker
levels exceeded for various percentages of the test time. Lin-
ear interpolation is used between classes where the percentile
value does not fall exactly into a particular class. The class
percentiles ClX calculation process using standard Matlab
functions can be expressed as:
[]
()
)100(, XClprctileClX?= (15)
where X is the percentile number defined in [5], X=0.1, 0.7, 1,
1.5, 2.2, 3, 4, 6, 8, 10, 13, 17, 30, 50, 80%.
0
20000
40000
60000
80000
100000
1.E-06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+0
6
Instantaneous flicker level P
f
(-)
Class number Cl (-)
Fig. 6. Classification function
The corresponding percentile values of the instantaneous
flicker level PX can be then calculated using the following
equation:
()
?
?
?
?
?
?
?
??
?
?
?
?
?
?
???
=1
2/
10log
10 NC
Cl
M
X
X
P (16)
And finally, the short-term flicker severity index Pst ac-
cording to the equation given in [5] is calculated from PX
percentiles as follows:
SS
SS
st PP
PPP
P
5010
311.0
08.028.0
0657.00525.00314.0
++
+++
= (17)
where
3
5
3
3
805030
50
17131086
10
432.2
3
5.117.0
1
PPP
P
PPPPP
P
PPP
P
PPP
P
S
S
S
S
++
=
++++
=
++
=
++
=
(18)
The Pst values can be also logged into another buffer for long
time period Tlt of 2 hours in case of Tst is equal to 10 minutes
and then a long-term flicker severity index Plt at the end of
each successive Tlt period is calculated using equation:
3
1
1∑
=
=N
istlt P
N
P (19)
where N is the number of Pst periods within the Tlt, i.e. 12.
3 PRELIMINARY VERIFICATION
For a verification of the proposed light-flickermeter, its
implementation in Matlab-Simulink has been realized. Since
there are not defined testing points for flickermeter based on
the evaluation of luminous variation, the verification proce-
dure for standard flickermeter has been adopted. In other
words, the test signals for the light-flickermeter model were
obtained by simulation of simplified reference 60W incan-
descent lamp model (Fig. 7) response on testing signals de-
fined in [5]. A block diagram of entire model for simulations
in Simulink is composed of diagrams at Fig. 7 and Fig. 4.
F
L
(s)
LP filter
u
2
7.51Hz
Wave
generator
N
U
u(t)
u(t)
Squaring
~
Φ
(t)
Fig. 7. Block diagram of the reference incandescent lamp
luminous flux fluctuation generation in Matlab Simulink
A sinusoidal/rectangular amplitude modulated normalized
signals with modulation parameters given in tables 1, 2 and 5
of [5] were generated using Matlab-Simulink signal sources
and have been subsequently coupled to a part simply model-
ing transfer of the input voltage on output luminous flux in
reference bulb, where the transfer function FL(s) is defined by
eq. (11). Then the implemented light-flickermeter model is
depicted on Fig. 4 and described in section 2.2. The photo-
detector is supposed to be an ideal in the model and is in fact
a part of the lamp model.
The results of performed preliminary compliance tests are
shown at Fig. 8 and Fig. 9. Whereas the required accuracy is
±5% of the reference value which is Pst=1 or Pf=1. As it can
be seen the deviations from expected values of Pst and Pf are
in allowed tolerance except testing points at modulation fre-
quency of 33? Hz. However it can be adjusted by setting the
scaling factor S.
10
0
10
1
10
2
10
3
0.9
0.95
1
1.05
1.1
Number of changes per minute r (1/min)
Flicker severity index P
st
(-)
Fig. 8. Rectangular voltage AM performance test results, test
No.1
0 5 10 15 20 25 30 35
0.9
0.95
1
1.05
1.1
Modulation frequency f
M
(Hz)
Instantaneous flicker level P
f
(-)
Rectangular AM
Sinusoidal AM
Fig. 9. Flickermeter compliance test results to normalized
sinusoidal and rectangular voltage AM, test No.2, 3
4 CONCLUSION
The paper introduced the light-flickermeter intended for
comparative measurements of disturbing light flicker pro-
duced by light sources of various types. The light-
flickermeter was designed on the basis of the standard IEC
flickermeter and its performance was verified by means of
simulations in Matlab-Simulink.
Even if such simple light-flickermeter has very short
range of applications and is far from an objective light-
flickermeter which could be used without any limitation, the
presented concept shows possibilities for future improve-
ments.
ACKNOWLEDGEMENT
This paper contains the results of research works funded
by the Czech Science Foundation under grant no.
GP102/08/P582.
REFERENCES
[1] Drapela, J. A Time Domain Based Flickermeter with
Response to High Frequency Interharmonics. In Pro-
ceedings of 13th International Conference on Harmon-
ics and Quality of Power. 1. University of Wollongong:
IEEE PES, 2008. pp. 1-7. ISBN: 978-1-4244-1770-4
[2] Drapela, J. Kratky, M., Weidinger, L., Zavodny, M.
Light Flicker of Fluorescent Lamps with Different
Types of Ballasts Caused by Interharmonics. 2005
IEEE St. Petersburg PowerTech Proceedings. St. Pe-
tersburg, Russia, IEEE PES, 2005, 7pp., ISBN 5-93208-
034-0
[3] Drapela, J., Toman, P. Interharmonic – Flicker Curves
of Lamps and Compatibility Level for Interharmonic
Voltages. 2007 IEEE Laussane Powertech, Switzerland.
IEEE PES, 2007, pp. 1552-1557, ISBN 978-1-4244-218
[4] EN 61000-2-2 ed. 2:2002. Electromagnetic Compatibil-
ity (EMC) – Part 2-2: Environment – Compatibility lev-
els for low-frequency conducted disturbances and
signalizing in public low-voltage power supply systems
[5] EN 61000-4-15:2003. Electromagnetic compatibility
(EMC). Part 4: Testing and measurement techniques.
Section 15: Flickermeter. Functional and design speci-
fications.
[6] Gallo, D., Landi, C., Pasquino, N. Design and Calibra-
tion of an Objective Flickermeter. IEEE Transactions
on Instrumentation and Measurement. Vol. 55, 2006,
pp. 2118-2125.
[7] Kaufman, J.E., et al. IES Lighting Handbook. Reference
Volume. IES, Waverly Press, USA, 1984, ISBN 0-
87995-015-3
[8] Lee, B.B., Martin, P.R., Valberg, A. Sensitivity of Ma-
caque Retinal Ganglion Cells to Chromatic and Lumi-
nance Flicker. Journal of Physiology. Vol. 414, pp.
223-243, 1989, ISSN
[9] McCluney, R. Introduction to Radiometry and Pho-
tometry. Artech House, MA, USA, 1994, 402 pp., ISBN
0-89006-678-7
[10] Mombauer, W. Calculating a New Reference Point for
the IEC Standard Flickermeter. European Trans. On
Electrical Power, Vol. 8, Part 6, December 1998, pp
429–436.
[11] Mombauer, W. Flicker caused by interharmonics. et-
zArchiv Vol. 12 (1990) p. 391-396
[12] Mombauer, W. Flicker caused by phase jumps. Euro-
pean Trans. On Electrical Power, Vol. 16, No. 6, 2006,
pp. 545–567
[13] OSI Optoelectronics. Photodiode Characteristics [on-
line]. 6pp., www.osioptoelectronics.com
[14] Peretto, L., et al. Theoretical Analysis of the Physiol-
ogic Mechanism of Luminous Variation in Eye-Brian
System. IEEE Transaction on Instrumentation and
Measurement, Vo. 56, No. 1, February 2007, pp. 164-
170, ISSN 0018-9456
[15] Rea, M.S. et al. The IESNA Lighting Handbook. Refer-
ence and Application. IESNA, USA, 2000, 9th ed.,
ISBN 0-87995-150-8
[16] Sakulin, M., Renner, H., Bergeron, R., Key, T., Nastasi,
D. International Recommendation for Universal Use of
the UIE/IEC Flickermeter. UIE XIII Congress on Elec-
tricity Applications 1996, 8pp.
