Osram HQI-E 150 W/NDL CL Datasheet

Osram HQI-E 150 W/NDL CL Datasheet
www.osram.com
Metal
halide lamps
Instructions for the use and application
Contents
1
Introduction....................................................................................................................................................4
2
How a metal halide lamp works ......................................................................................................................5
2.1 Quartz discharge tube.............................................................................................................................6
2.2 Ceramic discharge tube (PCA = polycrystalline alumina) .........................................................................6
2.2.1 1st generation: cylindrical form .....................................................................................................6
2.2.2 2nd generation: freely moldable ceramic, POWERBALL ................................................................6
3
Ballasts for discharge lamps ..........................................................................................................................8
3.1 Inductive ballasts (chokes) ......................................................................................................................8
3.1.1 American circuits for ballasts ........................................................................................................9
3.1.2 Variation in supply voltage for adapted inductance .....................................................................10
3.1.3 Influence of deviations in supply voltage.....................................................................................11
3.1.4 Capacitor for power factor correction .........................................................................................11
3.2 Electronic control gear (ECG) ................................................................................................................12
3.2.1 Structure and functioning of an electronic ballast. ......................................................................12
3.2.2 Service life and temperature .......................................................................................................13
3.2.3 Advantages of operation with electronic ballast POWERTRONIC PTi ..........................................13
3.3 Influence of harmonic waves and corresponding filters .........................................................................15
3.4 Brief voltage interruptions .....................................................................................................................16
3.5 Stroboscopic effect and flicker .............................................................................................................17
4
Igniting and starting discharge lamps ...........................................................................................................19
4.1 External ignition units ...........................................................................................................................19
4.1.1 Parallel ignition unit ....................................................................................................................19
4.1.2 Semi-parallel ignition unit ...........................................................................................................19
4.1.3 Superimposed ignitor .................................................................................................................20
4.2 Warm re-ignition ...................................................................................................................................20
4.3 Hot re-ignition.......................................................................................................................................20
4.4 Ignition at low ignition voltage (Penning effect) .....................................................................................20
4.5 Ignition at low ambient temperatures ....................................................................................................20
4.6 Cable capacitance ................................................................................................................................21
4.7 Start-up behavior of metal halide lamps ................................................................................................21
5
Reducing the wattage of high intensity discharge lamps ..............................................................................23
5.1 Introduction ..........................................................................................................................................23
5.2 Wattage reduction techniques...............................................................................................................23
5.2.1 Reducing the supply voltage .......................................................................................................23
5.2.2 Phase control: leading edge, trailing edge ..................................................................................24
5.2.3 Increasing choke impedance or decreasing lamp current ............................................................24
5.2.4 Change in frequency for high-frequency mode............................................................................24
5.3 Recommendations for reducing the wattage in discharge lamps ...........................................................25
5.3.1 Metal halide lamps .....................................................................................................................25
5.3.2 Dimming for other discharge lamps ............................................................................................25
6 Lamp service life, aging and failure behavior ................................................................................................26
6.1 Lamp service life and aging behavior ....................................................................................................26
6.2 Storage of metal halide lamps ...............................................................................................................26
6.3 Failure mechanisms of metal halide lamps ............................................................................................26
2
6.3.1 Leaking arc tube .........................................................................................................................27
6.3.2 Increase in re-ignition peak.........................................................................................................27
6.3.3 Broken lead or broken weld ........................................................................................................28
6.3.4 Leaking outer bulb......................................................................................................................28
6.3.5 Lamps that do not ignite.............................................................................................................28
6.3.6 Breakage or differing wear of the electrodes ...............................................................................29
6.3.7 Scaling of the base / socket .......................................................................................................29
6.3.8 Bursting of the lamp ...................................................................................................................29
6.3.9 Rectifying effect .........................................................................................................................29
6.3.10 Conclusions................................................................................................................................31
7
Luminaire design and planning of lighting systems .......................................................................................32
7.1 Measuring temperatures, ambient temperature .....................................................................................32
7.1.1 General physical conditions for temperature limits for outer bulbs
and pinches in metal halide lamps ..............................................................................................32
7.1.2 2 Measurement with thermocouple .............................................................................................32
7.1.3 Measuring points for thermocouples in different lamp types .......................................................33
7.2 Influence of ambient temperature on ballasts and luminaires ................................................................36
7.3 Lamp holder .........................................................................................................................................36
7.4 Leads to luminaires ...............................................................................................................................37
7.5 Maintenance of lighting systems with metal halide lamps......................................................................37
7.6 Standards and directives for discharge lamps .......................................................................................39
7.6.1 Standards ...................................................................................................................................39
7.6.2 Directives ...................................................................................................................................41
7.6.3 Certificates .................................................................................................................................41
7.7 Radio interference ................................................................................................................................42
7.8 RoHS conformity ..................................................................................................................................42
7.9 Optical design of reflectors ...................................................................................................................42
7.9.1 Condensation on the lamp ..........................................................................................................42
7.9.2 Projection of the condensate ......................................................................................................42
7.9.3 Back reflection on the lamp ........................................................................................................43
8
Light and colour ...........................................................................................................................................43
8.1 Night vision...........................................................................................................................................44
8.2 Colour rendering ...................................................................................................................................46
8.2.1 Test colours from standard DIN 6169 ..........................................................................................47
8.3 Light and quality of life .........................................................................................................................48
8.4 UV radiation..........................................................................................................................................49
8.4.1 Fading effect ..............................................................................................................................50
8.4.2 Protective measures to reduce fading .........................................................................................50
9
Disposal of discharge lamps.........................................................................................................................51
9.1 Statutory requirements .........................................................................................................................51
9.2 Collection, transport and disposal of discharge lamps at end-of-life .....................................................51
9.3 Ordinance on Hazardous Substances ...................................................................................................51
10 List of abbreviations .....................................................................................................................................52
11 Literature .....................................................................................................................................................53
3
1 Introduction
Metal halide lamps offer a number of advantages that
favor their use in ever broader areas of application.
These include high luminous efficacy, a long service
life and good colour rendering. Because the light is
generated in a small space, the discharge lamps
almost correspond to a spot light source, with
advantages in terms of light control and brilliance
of the illumination.
Table 1: The properties of metal halide lamps and resulting application areas
X
X
X
Changing lamps is difficult and
expensive
Long service life with longer
change intervals
X
X
X
Long operating hours
High efficacy and long
service life
X
Realistic reproduction of high
value surfaces, pictures,
products and TV images
Good colour rendering
High illuminance values
High luminous flux permits
few fixtures light points
Luminaires should be small or
discreet
Small dimensions permit
compact luminaires
High mounting heights and
wide spacing demand precise
light control
Small arc tubes permit very
good light control
The means of generating light is technically complicated. The key principles regarding the operation of these lamps and instructions for their use
are listed below.
4
Sports
facilities
High luminous flux permits
few fixtures light points
Building
illumination
The height requires a lot of light
X
X
X
X
X
X
X
Street
Foyer
metal halide lamps
Trade
show
Property of
application
Industry
Requirements in the
Shop
High room, building
X
X
X
X
X
X
X
X
X
X
X
X
X
These application instructions address a large number
of users, such as luminaire designers, lighting planning
engineers, operating device developers and retailers.
Naturally, not all users will find all sections relevant,
but the aim has been to cover the interests of as many
users as possible.
In an arc tube, gas discharge works through excitation
of the luminous additives (metal halide salts) and the
mercury is excited by the current flow. Visible radiation characteristic for the respective elements is emitted. The mixture of the visible radiation of the different
elements results in the designed colour temperature
El
ec
tro
de
M
ol
yb
de
nu
Ar
m
c
fo
(Q tub
il
ua e
rtz
)
Di
sc
ha
rg
H
e
ea
ar
c
tr
efl
ec
to
r
Le
ad
-in
w
ire
M
ol
yb
de
nu
m
Ba
fo
se
il
y
er
cu
r
M
ha
lid
es
U
VQ Filt
ut er
(Q er b
ua u
rtz lb
)
M
et
al
G
et
te
r
In discharge lamps, light is generated by a gas discharge of particles created between two hermetically
sealed electrodes in an arc tube. After ignition, the
particles in the arc are partially ionized, making them
electrically conductive, and a “plasma” is created. In
high intensity discharge lamps, the arc tube is usually
enclosed in an evacuated outer bulb which isolates the
hot arc tube thermally from the surroundings, similar
to the principle of a thermos flask. But there are also
some discharge lamps without outer bulbs, as well as
lamps with gas-filled outer bulbs. In contrast to lowpressure discharge, there is high pressure and a high
temperature in a discharge tube.
C
on
ta
ct
Similar to high-pressure mercury lamps or high-pressure sodium lamps, metal halide lamps also belong to
the group of discharge lamps. Low pressure discharge
lamps include fluorescent lamps and compact fluorescent lamps.
pl
at
e
2 How a metal halide lamp works
Fig. 1: An example of how a metal halide lamp works
based on a double-ended lamp with a quartz arc tube.
and colour rendering for a particular lamp. In the operating state, the mercury evaporates completely. The
other elements involved are present in saturated form
at the given temperatures, i.e. they only evaporate in
part; the rest is in liquid form at the coolest point in
the arc tube. The fraction of the filling that has evaporated depends on the temperature of the coolest point
on the arc tube wall and also varies for the different
filling components. Changes to the temperature of the
arc tube wall can change the composition of the metal
halides in the discharge, thus also changing the colour
properties of the lamp.
Hg = Mercury
Me = Metals
Hal = Halids
Outer bulb
Electrodes
Arc tube
Mo-Foils
ignitor
Alternative: ECG
ballast
supply voltage
Fig. 2: Tasks of the metals [sodium (Na), thallium (TI), indium (In), tin (Sn), lithium (Li), rare earths; dysprosium (Dy),
holmium (Ho), thulium (Tm)]
5
Ceramic arc tubes can be produced with smaller
dimensional tolerances, reducing the variation in lighttechnical and electrical parameters.
Multiple lines
GREEN
Ceramic is less susceptible to attacks from the aggressive metal halide filling and is less permeable for
filling particles, resulting in a considerably longer
service life compared to quartz tube lamps.
YELLOW
RED
Ceramic arc tubes are now available in various different
forms: the original cylindrical version and the improved
round version.
BLUE
Variation possibilities of the Colour Temperature
Tn
2.2.1 1st generation: cylindrical form
CRI
Daylight
Neutral White
Neutral White de luxe
Warm White de luxe
Fig. 2a: Generation of the desired Spectral Distribution
Components in order to achieve high luminous Efficacies and good Colour Rendering
In the first version, the ceramic arc tube was designed
in a cylindrical form, based on the production technology for the high-pressure sodium lamp. The arc tube
was made up of cylindrical sub-sections sintered together. The arc tube consisted of a relatively thick plug
at either end of the tube: this was necessary for the
durability and functioning of the tube.
2.2.2 2nd generation: freely moldable ceramic,
POWERBALL ®
2.1 Quartz discharge tube
The discharge tubes in 1st-generation metal halide
lamps are made of high purity quartz glass. This quartz
material allows for stable operation at high temperatures, is resistant to sudden changes in temperature
and is transparent. The well proven HQI lamps are produced in various different forms using this technology.
H
Q
I
...
...
...
Hydrargyrum (Greek-Latin for mercury)
Quartz
Iodide
• Well proven lamp technology
• Wide wattage range 70 W – 2000 W
• Colour temperatures up to 7250 K
• Good optical properties thanks to transparent
discharge tube
2.2 Ceramic discharge tube
(PCA = polycrystalline alumina)
The use of arc tubes made of ceramic material further
enhanced some of the metal halide lamp’s properties.
Ceramic can withstand higher temperatures than
quartz glass. This permits higher wall temperatures
thereby evaporating more of the metal halide salts into
the gas arc and allowing for more efficient use of the
chemicals. Ceramic lamps offer improved luminous
efficacy and colour rendering as a result.
6
A second step with changed production technology
permitted production of freely moldable tube geometries. This made it possible to produce round ceramic arc tubes with a constant wall thickness – the
POWERBALL® arc tubes. The round form and constant
wall thickness brought considerable advantages. The
possibility of further increasing the wall temperature
improves luminous efficacy and colour rendering. The
absence of the thick plug at the end of the tube reduces light absorption in this area, resulting in a higher
luminous flux with more uniform irradiation characteristics. There are fewer differences in the wall temperature between the various burning positions and therefore also smaller differences in colour between the
burning positions. The reduced ceramic mass of the
round tube enables the tube to heat up faster, reaching
the photometric values more quickly. Similarly, when
the lamp goes off, warm re-ignition is possible more
quickly because the required cooler starting temperature for normal ignition devices is achieved faster.
The uniform wall thickness and round shape produce a
more even temperature curve along the inner tube wall
as shown in Fig. 3, based on the temperature shown in
colours diagram. The steeper temperature gradient in
the cylindrical ceramic favors chemical transport processes. During this process, aluminum oxide ceramic
dissolves in the liquid metal halide melt and settles
at cooler points of the arc tube. If erosion from the
wall goes too far, this can lead to leakage in the tube,
causing the lamp to fail. Failures due to so-called
“ceramic corrosion” are thus less likely to occur in
lamps with round ceramic tubes.
9000 h
12000 h
Convection
Convection
Solution of
Alumina
in MH-melt
= Corrosion
Condensation
of Metall Halides
Transport of
soluble
alumina in
MH-melt
Evaporation
of Metal Halides
Evaporation
of Metal Halides
1100
Deposition of alumina
by saturation of HM-melt
due to cooling
1060
1020
Condensation of
Metal Halides
980
940
Fig. 3: Comparison of ceramic corrosion between the different tube forms
The advantages of POWERBALL® technology compared to cylindrical solutions
• Better maintained luminous flux throughout the service life
• Improved color rendering, particularly in the red
• Improved color stability during the service life
• More uniform operation independent of burning position
• More constant luminous intensity distribution
• Faster start-up behavior
7
3 Ballasts for discharge lamps
Since the discharge reacts to increasing lamp current
with falling voltage (which would cause the current to
rise indefinitely until the fuse blows or another part of
the circuit fails), the lamp current must be limited by
a ballast during operation. This usually consists of an
inductive circuit (choke), although in rare cases up to
400 W capacitive circuits are also possible (although
this usually results in a shorter service life). In the
blended lamp (HWL), the resistance of the filament
serves as a series resistor for the high-pressure mercury discharge lamp. In most cases, additionally to the
current-limiting element, an ignition device is needed
to start discharge (see chapter 4 “Ignition and starting
discharge lamps”).
In modern luminaires, an electronic ballast fulfils the
function of igniting the lamp, limiting the lamp current
and controlling the lamp wattage.
3.1 Inductive ballasts (chokes)
The voltage across the electromagnetic ballast increases as the current increases, therefore a stable
working point can be achieved in the series connection of the discharge lamp and the choke.
Charting the equations results in the curves shown in
Fig. 5. The difference between lamp wattage and the
product of lamp voltage and lamp current is called
lamp power factor. It reaches values between 0.7 and
0.95 depending on the operating mode. The yellow
curve was generated by using a higher lamp power
factor λL [(to be more exact: 1.05*(1-n/3)].
Typical voltage and current waveforms as shown in
Fig. 6 show that while the current is (approximately)
sinusoidal, voltage is not. After the current zero crossing, the voltage initially increases (so-called re-ignition
peak) to then fall to a relatively constant value (saddle)
(see also chapter 6.2.2 and Fig. 30). Voltage remains
approximately the same beyond the maximum of the
current, and has the same zero crossing as the current.
So there are areas with high voltage which count
towards the effective value of the voltage but don’t
contribute to the wattage as the current at that point in
time is nearly zero. This results in lamp power factors
deviating from the value 1.
If the lamp voltage is equal to zero, the voltage drop
across the choke is the entire supply voltage, and the
choke short-circuit current is reached. This is the maximum current that can flow through the choke inasmuch the current has no DC component (see chapter
6.2.9 for effects of direct current components).
The following curves are typical for 150 W and apply in
the same way to other wattages.
CPFC ... PFC capacitor
La
La ... lamp
180
US ... supply voltage
160
Ch ... choke
UNUS
Describing the relationships of current and voltage
requires a system of differential equations which cannot generally be solved. The following approximation
formulas describe how the lamp current and lamp
wattage depend on the relationship of lamp voltage to
supply voltage [3]:
IL ≈
US
Z
[(1 − n )
[(1 − n )
2 1/ 2
2 1/ 2
]
− 0, 25n
]
− 0, 25n
Lamp power PL in W
Fig. 4: Discharge lamp with inductive ballast (ignition
unit has been left out, the various possibilities are featured in chapter 4 “Ignition and start-up of discharge
lamps”)
2
⎛ n⎞
US
⋅ n ⎜1 − ⎟
Z
⎝ 3⎠
2
140
K PFC
C
PL =
2,5
5 % higher lamp
power factor
120
PL
1,5
100
IL
80
1
60
40
0,5
20
0
0
0
0,5
UL/US
1
Fig. 5: Lamp current IL, lamp wattage PL over the ratio
of lamp voltage to supply voltage UL/US; Z=99 Ω for a
150 W lamp
Voltage in V
Current in A
(Gl. 4.1)
(Gl. 4.2)
whereby: (1-n/3) ....approximation for the lamp
power factor λL
PL ..... lamp wattage in W
US ..... supply voltage in V
n ..... ratio of lamp voltage UL to supply voltage US
Z ..... choke impedance
8
Lamp current IL in A
Ch
Dr
Time in ms
Fig. 6: Graph showing lamp voltage and current of a
150 W lamp when operated at a choke (applies in the
same way to other wattages)
This lamp behavior results from the relatively flat zero
crossing for sinusoidal current. When the current approaches zero, the plasma temperature decreases and
the electrodes also cool down. The recombination of
electrons with ions reduces conductivity. After the zero
crossing, the conductivity is too low to take up the
current that the choke wants to drive. As a result, the
voltage through the lamp increases again significantly
until the lamp “reignites”. The higher voltage results
in a higher ionization rate that increases conductivity
again so that voltage falls.
By contrast, current and voltage for the rectangular
waveforms of an electronic ballast change significantly faster from positive to negative half-wave, or
have a faster commutation times (see chapter 3.2
“Electrical ballasts (ECG)”), so that the plasma has
little chance to cool down. The instantaneous voltage
required from the electronic ballast is therefore significantly lower than for the choke. This is one of the
advantages of electronic ballast, as one of the failure
mechanisms of metal halide lamps is to extinguish
due to high re-ignition voltage. The re-ignition peak
of a lamp normally increases over the service life, and
when it exceeds what the supply voltage momentarily
can provide, there is no “re-ignition” and the lamp
goes out (see also chapter 6.2.2 “Increase of the
re-ignition peak”).
When operating on a conventional choke, the lamp
wattage runs through a maximum depending on the
lamp voltage (see Fig. 5). The maximum occurs for a
lamp voltage of slightly more than half the supply voltage. Near the maximum the lamp wattage changes
only slightly with the lamp voltage. During the lamp
service life, the lamp voltage increases, as also shown
in chapter 6 “Lamp lifecycle, aging and failure behavior”. In order for the lamp wattage to change as little
as possible, the nominal value for lamp voltage is generally chosen near the maximum, therefore at about
half the supply voltage.
The impedance of the choke is rated at a certain supply frequency and certain supply voltage. Deviations
from the nominal supply voltage will result in a different ballast curve and a related different working point
for the lamp and therefore different lamp wattage. To
limit the associated greater spread in the lamp parameters, a maximum deviation of 5% from the nominal
values is permitted in the short term for the supply
voltage, or maximum 3% in the long term. For deviations over a longer period of time, suitable choke tap
must be selected. Similarly, the choke impedance
must not deviate from the nominal values by more than
2% (see also chapter 3.1.3 “Influence of deviations in
supply voltage”).
As per the IEC 61167 standard, ballast units for MH
lamps must by protected from overheating through
rectification. This can be done e.g. with a thermal fuse
(tested according to IEC 598-1, Annex C).
3.1.1 American circuits for ballasts
In this context it is important to note that the supply
voltage in America has a different frequency (60 Hz).
As the inductive resistance of the choke depends
on the frequency, in this case it is important to use a
designated ballast for the corresponding frequency. In
addition, both lamps and ballasts in the USA are standardized by ANSI, the American National Standards
Institute. To operate the systems correctly, lamps must
be operated with corresponding ballasts. Ratings designations are required by ANSI to be marked clearly on
all products, allowing users to clearly identify system
agreement.
3.1.1.1 Autoleak transformer or high reactance autotransformer
If the supply voltage is smaller than around twice the
lamp voltage, as is the case, for example, in the USA
or Japan, then the supply voltage must first be stepped
up. A good way of doing this is use an autoleak transformer. Part of the secondary windings act as lamp
choke. On the one hand, this saves on material, and on
the other hand, a higher voltage (open-circuit voltage)
is available to start the lamp. These types of ballasts
are typically more economical than a constant wattage
style ballasts at the expense of wattage regulation.
CPFC ... PFC capacitor
La ... lamp
Tr
US
La
Tr ... autoleak transformer
CPFC
Fig. 7: Autoleak transformer
3.1.1.2 Constant wattage ballast
A constant wattage ballast such as those widely available in the USA consists of an autoleak transformer in
series with a capacitor. The advantage of this circuit is
the reduced impact of fluctuations in the supply voltage and the possibility of operating the lamp at supply
voltages (110/120 V in the USA, 100 V in Japan) that
lie within the range of the lamp voltage.
C
• Maximum permitted supply voltage deviations:
5% in the short-term, 3% in the long-term, use
other tap on the choke if necessary.
US ... supply voltage
C ... capacitor
La ... lamp
US
Tr
La
US ... supply voltage
Tr ... autoleak transformer
• Maximum deviation of choke impedance 2%.
• The choke must be protected against overheating
according to the standard (thermal fuse).
Fig. 8: Constant wattage ballast
9
3.1.2 Variation in supply voltage for adapted
inductance
3
200
180
3.1.2.1 Operation at supply voltage higher than 230 V
with adapted choke impedance
An increase in supply voltage shifts the maximum of
the choke characteristic curve (PL over UL/UN). In the
lamp voltage range of OSRAM lamps (approx. 100 V),
the change in lamp wattage with changing lamp voltage is steeper. In addition, the maximum wattage that
can be achieved with increasing lamp voltage is larger,
as shown in Fig. 9. Normally, the lamp voltage increases with increasing service life (see also chapter 6
“Lamp service life, aging and failure behavior”).
According to equation Eq. 4.1, wattage of about 150 W
is achieved for a 150 W choke with a lamp voltage of
100 V. The maximum the lamp wattage can increase
to for a lamp voltage of 150 V is 175 W. The higher
achievable wattage can reduce the service life and
possibly cause an increase of undesirable effects at
the end of the service life (e.g. lamp explosion).
140
Lamp power PL in W
Some countries have supply voltages that permanently
deviate from 230V. When using correspondingly adapted inductances, the following points must be taken
into account.
2
120
1,5
100
80
1
60
40
Lamp current IL in A
2,5
160
0,5
20
0
0
0,5
UL/US
1
0
Fig. 9: Lamp current IL, lamp wattage PL over the ratio
of lamp voltage to supply voltage UL/US for US=277 V
3.1.2.2 Operation at supply voltage less than 230 V
with adapted choke impedance
Supply voltages of less than 230 V shift the maximum
of the choke curve (PL over UL/US). Operation at 200 V
supply voltage for example is more favorable than at
230 V with regard to the change in lamp wattage with
lamp voltage. The PL(UL) curve runs flatter in the normal
range of lamp voltage. For lamp voltages exceeding
130 V Wattage falls again.
3
200
180
a) An increase in negative effects must be expected
at the end of the service life, because wattage
rises clearly above the nominal wattage when
lamp voltage increases on account of the shifted
choke characteristic curve. The increased wattage input for lamps with already aged arc tube
wall can cause increased lamp explosion rates,
for example. Operation in overload conditions
will probably cause accelerated aging.
b) The steeper characteristic PL(UL) in the range of
the normal lamp voltage causes a higher spread
of the wattage and therefore of the photometric
data, e.g. perceived colour variation.
We therefore discourage operating the lamps at 277 V
supply voltage. Our lamps have been developed and
undergone service life testing at 230 V supply voltage,
so that we cannot assume any warranty for the service
life behavior and photometric data for any deviating
operation.
10
140
2
120
100
1,5
80
1
60
40
Lamp current IL in A
2,5
160
Lamp power PL in W
OSRAM lamps are generally designed to operate at
230 V supply voltage and undergo corresponding service life testing. There are, however, also systems at
400 V, e.g. for some discharge lamps > 1000 W. For
these lamps, the following explanations apply in the
same way. The use of high intensity discharge lamps is
theoretically also possible at 277 V operating voltage
with adapted impedance and ignition devices, although
such operation is associated with considerable disadvantages.
0,5
20
0
0
0,5
UL/US
1
0
Fig. 10: Lamp current IL, lamp wattage P L over the ratio
of lamp voltage to supply voltage UL/US at UN=200 V
There is a major drawback that with lower supply voltage, there is also less voltage available for re-ignition
after the current has passed zero crossing. If the momentary supply voltage is lower than the re-ignition
voltage, the lamp goes off. Normally, the lamp voltage
and also the re-ignition peak increase with increasing
service life (see also chapter 6, “Lamp service life, aging and failure behavior”). That means that a reduction
in supply voltage causes a shorter service life in many
lamps.
When operating metal halide lamps on a choke, the
lamp parameters change depending on the supply
voltage. To limit the associated variation in lamp photometrics, a maximum deviation in supply voltage of
5% from the nominal values for the supply voltage is
permitted in the short term, or maximum 3% in the
long term. For deviations over a longer period of time,
suitable ballast tap must be selected. As choke impedance also influences the lamp parameters via the correspondingly adjusted lamp current, this is allowed to
deviate from the nominal values by maximum 2%.
Percentage of the nominal value
3.1.3 Influence of deviations in supply voltage
Percentage of the nominal supply voltage
Remote mounting can also cause noticeable decreases in voltage (see also chapter 7.4 “Leads to
luminaires”).
Long term reductions in lamp wattage cause the luminous flux to decrease, with a shorter service life and
a deviation in colour from the nominal values, as also
explained in chapter 5: “Wattage reduction in high
intensity discharge lamps”.
Percentage of the nominal value
If the supply voltage is too high, the arc tube is operated at too hot a temperature, causing increased
blackening and a shorter service life.
Percentage of the nominal supply voltage
UL in %
IL in %
PL in %
FL in %
UL ... Lamp voltage
IL ... Lamp current
PL ... Lamp power
FL ... Luminous flux
Fig. 11: Lamp parameters of a typical OSRAM HQI®
lamp over supply voltage
UL in %
IL in %
PL in %
FL in %
UL ... Lamp voltage
IL ... Lamp current
PL ... Lamp power
FL ... Luminous flux
Fig. 12: Lamp parameters of a typical OSRAM HCI ®
lamp over supply voltage
3.1.4 Capacitor for power factor correction
The capacitor for power factor correction is necessary to
correct the power factor of the system when operating
discharge lamps at electromagnetic ballasts. Inductively
stabilized discharge lamps achieve power factors of only
about 0.5 because of the dephased current flow. The
power factor of a load is defined as the ratio of effective
power to the apparent power actually withdrawn from
the grid (kW to kvar) and is referred to as cos ϕ. The
apparent power comprises the effective power used by
the consumers to create e.g. heat or mechanical energy
and the idle power that is used to develop magnetic or
electric fields of inductivity and capacity. However the
latter flows back into the grid after a half cycle length,
i.e. it is not actually “used”. The closer cos ϕ is to one,
the smaller is the share of wattless power withdrawn
from the grid. A higher share of wattless power results in
a higher flow in current for which the supply lines have to
be rated. Similarly, power dissipation in the supply lines
increases in a square progression with the current. In
order to achieve the values demanded by the utility companies of more than 0.85, a grid parallel capacitor must
be selected according to the lamp or choke current to
approximately correct the shift in phase. By including an
exactly calculated capacitor, the inductive wattless load
required by an electric consumer can be offset with a
capacitive wattless load. It is thus possible to reduce the
wattless power withdrawn from the grid; this is called the
power factor correction or wattless power compensation.
The capacitors are differentiated as follows,
depending on the arrangement and form of use:
INDIVIDUAL OR FIXED COMPENSATION, where the inductive wattless power is corrected directly where it occurs,
relieving the strain on the leads (typical for individual consumers usually working in continuous mode with constant
or relatively large wattage – discharge lamps, asynchronous motors, transformers, welding equipment, etc.)
11
The parallel capacitor has no influence on lamp
behavior.
GROUP COMPENSATION, where one joint fixed capacitor is allocated to simultaneously working inductive consumers, similar to individual correction (motors
located close together, discharge lamps). Here again
the strain on the leads is relieved, but only up to the
point of distribution to the individual consumers. Under unfavorable conditions, resonance can be caused
in two-phase grids.
CENTRAL COMPENSATION, where a number of capacitors are connected to a main or sub-distribution
station. This is the normal procedure in large electrical
systems with changing load. Here the capacitors are
controlled by an electronic controller which constantly
analyzes the demand for wattless power in the grid.
This controller switches the capacitors on or off to correct the current wattless power of the total load and
thus reduce overall demand in the grid.
Capacitor for power factor correction values are
stated for every lamp type in the Technical Information
and can also be calculated using the following equation.
CPFC =
1
2 × π × fS × U S2
× ⎛⎜ U S2 × IL2 − PW2 − (PW × tan ϕ K )⎞⎟ Eq. 4.3
⎝
⎠
CPFC in F Capacitance of the capacitor for power
factor correction
US in V Rated supply voltage
fS in Hz Supply frequency
IL in A
Lamp rated current
PW in W Total active power (lamp rated wattage plus
choke loss wattage)
ϕK
Tolerable or desirable phase difference between the fundamental waves of the supply
voltage and the supply current.
But this only corrects the power factor for the fundamental wave. Phase difference remains for distortion,
i.e. the harmonic waves, between current and voltage.
For this reason, the overall power factor can also only
reach values between 0.95 and 0.98 in practice.
A higher level of harmonic waves can cause resonance
effects and destroy the lamp. A power factor close to 1
is to be avoided as this can cause resonance between
the choke and correction capacitor.
While a discharge lamp is starting up, the power factor
undergoes significant changes in value. After ignition,
the lamp voltage is still very low with current higher
than in steady state. This is why the power factor in
this state is still low (inductive). While lamp voltage
increases and the lamp current falls, the power factor
increases to its nominal value of 0.85 – 0.9.
As discharge lamps age, it is normal for lamp voltage to
increase, causing the lamp current to fall according to
12
the ballast curve. Because the capacitor for power factor
correction is rated for a specific lamp and choke current,
the power factor varies according to lamp current. For an
extremely high lamp voltage, the choke current is so low
that capacitive current exceeds the inductive current and
the complete circuit becomes capacitive.
Under certain conditions, audio frequency central control systems have to be considered during installation.
In these cases, suitable audio frequency attenuation
chokes are to be provided. This kind of system is still
sometimes used for day/night circuits in street lighting,
although directional radio systems are finding increasing use here.
3.2 Electronic control gear (ECG)
Together with conventional ballasts, the use of electronic ballasts has meanwhile become widely accepted
practice, particularly for interior lighting.
Electronic ballasts offer clear advantages compared
to conventional ballasts. The main advantages include
in particular simplified handling (e.g. lighter ballasts),
lower energy consumption, a positive impact on lamp
service life and light quality, and, last but not least,
controlled and reliable shutdown of lamps at the end
of the service life.
Basically, most of the technical information provided in
this manual applies to both conventional ballasts and
electronic ballasts. This refers for example to wiring
requirements, wattage-reduced operation of MH lamps
or instructions for luminaire design.
But in addition, there are also considerable differences
between operation on an electronic or magnetic ballast.
The following section briefly explains the main differences and their effects.
3.2.1 Structure and functioning of an electronic
ballast
Electronic ballasts mainly consist of units with rectangular current and voltage. In principle, it is also
possible to operate the lamps with high-frequency
sinusoidal current similar to the fluorescent lamp. In
any case, it is important to ensure that no acoustic
resonances occur as these may result in arc instability
(lamp flicker) and in severe cases, lamp rupture.
3.2.2 Service life and temperature
There is a significant difference between conventional
and electronic ballasts particularly with regard to the
service life and thermal behavior of the units.
ECG
Luminaire
US
Lamp
Fig. 13: Simplified circuit diagram showing the electronic operation of high intensity discharge lamps
Voltage in V
Current in A
Time in ms
Fig. 14: Current and voltage of a metal halide lamp
operated on a rectangular electronic ballast
For a conventional ballast, it can be presumed that the
service life is defined by the choke temperature tw. A
10 °C increase in the tw temperature means that the
service life is halved.
In electronic ballasts, these circumstances are far
more complicated. The mortality rate of individual components, the circuit design and above all the electronic
load and the temperatures at which the units are operated have a considerable influence on the service life
behavior.
This is why the nominal service life of electronic
ballasts is stated in combination with a failure probability. For example, all units in the product family
POWERTRONIC® PTi have a nominal service life of
40,000 hours with failure probability of maximum
10% when operated at the maximum permissible
temperatures.
The service life of electronic ballasts is influenced
directly by the temperature at which the units are operated. This is why 2 temperature values are defined
to describe the thermal behavior. The ambient temperature ta describes the temperature immediately
surrounding the unit and thus prevailing around the
electronic components. To be clear, this is not the
room temperature or the ambient temperature of the
luminaire.
When an electronic ballast is fitted in a luminaire, the
real ambient temperature ta of the ballast can only be
measured with great difficulty and at great effort. This
is why a second temperature has been stipulated: the
tc temperature. Basically this is the casing tempera-
ture which can be measured by a thermocouple at a
set point – the tc point – and is defined as maximum
permissible temperature at which safe operation of the
electronic ballast is still guaranteed. In addition, the tc
temperature is set in relation to the ballast service life.
That means that the measured tc temperature permits
very precise conclusions as to the anticipated service
life of the electronic ballast.
OSRAM’s HID electronic ballast for example principally
reaches its full nominal service life at the maximum
permitted tc temperature. In practice, this means
that any temperature levels below the t c temperature
always prolong the effective service life. As a rule of
thumb, it can be presumed that a temperature 10 °C
below the printed maximum tc temperature will double
the service life of the electronic ballast.
However, it is not advisable to use only the absolute
maximum tolerable tc value for conclusions regarding
the quality and service life of an electronic ballast. This
is because on the one hand, the position and therefore
indirectly also the value of the t c point can be freely
defined by every electronic ballast manufacturer. On
the other hand, the rule of stating the nominal service
life at the maximum permitted t c temperature has not
yet become established throughout the electronic
ballast industry. In practice this means that many
electronic ballasts only achieve approx. 50% of their
nominal service life at maximum tc temperature.
Nominal service life (B10): max. 10% of the
electronic ballasts have failed
A serious evaluation of the electronic ballast
service life is only possible by comparing the
electronic ballast ambient temperature t a with
the corresponding service life.
Comparison of the service life using only the t c
temperature is not appropriate.
3.2.3 Advantages of operation with electronic
ballast POWERTRONIC ® PTi
The following table provides an overview of the advantages of operating lamps with the electronic ballast.
The corresponding values and statements are based
on tests and experience with POWERTRONIC® PTi
ballasts, so that they cannot necessarily be transferred
1:1 to ballasts of other makes.
In comparing the conventional and the electronic
ballast, the performance of the conventional ballast
constitutes the reference parameter and is given a
value of 100. This is also based on the fact that the
lamp parameters are defined with the reference
conventional ballast.
For more details, please refer to the POWERTRONIC®
Technical Guide – Electronic control gears for metal
halide lamps.
13
Magnetic ballast
Electronic ballast POWERTRONIC®
Energy consumption
100
10 to 15% savings over the service
life
Lamp service life
100
Up to 30% longer depending on
lamp type and kind of use
Lamp start-up
Depends on type: usually approx.
60 to 90 sec. to reach 90% of the
luminous flux level
Up to 50% faster
Colour stability
Colour variation possible
Clearly reduced scattering; initial and
over service life
Cut-out at end of lamp service
life
Not available or only simple
cut-out mechanisms
Permanent parameter control,
intelligent cut-out mechanisms
Ignition cut-out
Only with timer ignition units
Ignition time limited to 18 min
Light flicker
Visible flicker
Flicker-free thanks to 165 Hz operation
Consistent wattage
Increase in wattage over service
life, also dependent on fluctuations in temperature and supply
voltage, and on lead length
+ 3% over the entire service life,
regardless of fluctuations in
temperature and supply voltage
or lead length
Handling
3 components, complicated
wiring
1 unit, simple wiring
Size and weight
Heavy, several components, large
in some cases
Light and compact
Power factor correction (PFC)
0.5 – 0.95, considerable aging
fluctuations
> 0.95
Noise development
Clearly audible humming possible
Almost noiseless
Bidirectional data exchange
Not possible
Generally possible
The main advantages of electronic ballasts are
described in greater detail in the following section.
3.2.3.1 Reducing energy consumption
Compared to conventional ballasts, electronic ballasts
can considerably reduce energy consumption over the
service life. The energy savings result from two factors:
1) Unit power dissipation:
In conventional ballasts a large amount of energy
is lost in dissipated heat on account of the
design. By contrast, electronic ballasts have a
low-loss design with top quality components,
reducing dissipation to less than 10% of the
nominal power.
2) Increase in wattage over service life:
The system wattage with conventional ballasts
fluctuates significantly over the service life of the
lamp. This results from the change in lamp voltage, which can increase by up to 30% throughout
the service life (see also chapter 3.1.2), resulting
in considerable fluctuations in lamp wattage.
14
By contrast, electronic ballasts operate the
lamps always with constant wattage throughout
the entire service life. The maximal tolerated
fluctuation is 3%. This means for example that
for a 70 W ceramic arc tube lamp, the electronic
ballast constantly provides the lamp with the
rated 73 W.
3.2.3.2 Lamp service life and cut-out at the end of the
service life
A detailed description of the lamp service life and failure behavior using conventional ballasts can be found
in chapter 6.
Electronic ballast operation also offers considerable
advantages in terms of lamp service life and cut-out
behavior at the end of the service life.
Comprehensive laboratory tests and extensive practical experience show that operation on electronic ballasts has a significantly positive influence on the lamp
service life. Precise but gentle lamp ignition, a more
stable thermal balance thanks to constant wattage
supply and, above all, a clearly reduced tendency to
go out by avoiding re-ignition peaks all lengthen the
lamp economic life for ceramic arc tube lamps by up
to 30% on average.
current of the electronic ballast, producing a higher
wattage input into the lamp which therefore heats up
more quickly.
3.2.3.4 Size, weight and handling
The electronic ballast also shows its strengths at the
end of the lamp service life. The ignition time limit
ensures that old lamps, where stable operation is no
longer possible, are not subject to endless ignition
attempts. After max. 18 minutes and precisely defined
ignition intervals, the POWERTRONIC® ECG cuts out
automatically. If a lamp goes out 3 times, the electronic
ballast also cuts out. This avoids interfering, flickering
light, prevents EMC load on the cables and also an
excessive load on the electronic ballast itself.
Permanent monitoring of parameters such as lamp
voltage or lamp current by the integrated micro-controller and alignment with pre-defined nominal values
also makes it possible to turn lamps off well before
they reach critical or undefined conditions which often
can hardly be managed.
Electronic ballasts combine ignition component, compensation component and choke in one unit. This
3-in-1 combination clearly reduces the installation
workload, the risk of installation errors and the need
to replace individual faulty units. Multi-lamp electronic
ballasts (e.g. 2x35 W or 2x70 W) duplicate these advantages because to connect to 2 luminares, only one
power lead is required.
Electronic ballasts are also lightweight. They weigh
50% to 60% less than magnetic ballasts, which of
course offers direct advantages in terms of luminaire
design: they can be sleeker in structure; a wider range
of materials can be used, and a lighter load can be
placed on the fastening components.
3.2.3.5 Bidirectional data transfer
3.2.3.3 Light quality, light colour, drop in light output,
start-up
HID lamps with electronic ballasts offer considerably
improved colour quality, both when initially installed
and throughout the service life.
The constant wattage supplied to the lamp by the
electronic ballast can compensate for differences in
light quality resulting for example from production tolerances or differing aging states. The result is visibly
more even light colour and a more uniform chromaticity coordinate.
Similarly, supply voltage fluctuations or the length of
the power leads are no longer relevant when using an
electronic ballast, as the constant wattage supply to
the lamp means these have no effect.
Lamp electrodes cool down to a lesser extent with
the rectangular electronic ballast thanks to the steep
electrical transitions through the zero crossing. Less
cooling down results in reduced sputter effects of the
electrodes, which in turn means less bulb blackening.
The more constant, and on average slightly higher,
lamp plasma temperature also produces 3% to 5%
more luminous efficacy, which also has a positive influence in luminous flux behavior in addition to reduced
blackening effects.
Electronic ballasts also have a much faster start-up
behavior than magnetic ballasts. Fig. 24 in chapter
4.7 for example clearly shows that a double-ended
quartz lamp operating with an electronic ballast already produces more than 90% of its max. luminous
flux after approx. 40 seconds. The same luminous flux
level with a conventional ballast takes at least 25 to 30
seconds longer. This at least 50% faster start-up at
the electronic ballast is due to the higher start-up
Intelligent electronic units will in the future offer completely new possibilities of controlling and monitoring
lighting systems, thanks to bidirectional data transfer.
Features such as querying the lamp or ballast status,
integration in Building Management Systems (BMS)
and central or local actuation and management of
lighting solutions will not only bring a clearly expanded
range of functions but also optimize maintenance and
repair work. In the medium term, it is quite conceivable
to see developments here similar to those in low-pressure discharge technology.
• Electronic ballasts are state of the art.
• Electronic ballasts can be used to achieve
significant increases in quality, reliability and
safety of lighting systems with metal halide
lamps.
• Most new MH lighting installations today are
already equipped with electronic ballasts.
3.3 Influence of harmonic waves and corresponding
filters
The development of modern semiconductor technology
with a significant increase in the number of consumers with solid state switches and converter controllers
unfortunately results in undesirable side-effects on the
AC voltage supply by causing considerable inductive
wattless power and non-sinusoidal current.
A typical converter current consists of various superimposed sinusoidal partial currents, i.e. a first
harmonic with the supply frequency, and a number of
15
so-called harmonic waves whose frequencies are a
multiple of the supply frequency (in three-phase supplies these are mainly the fifth, seventh and eleventh
harmonic waves).
These harmonic waves increase the current of the capacitor for power factor correction, as the reactance
of a capacitor decreases with increasing frequency.
The increasing capacitor current can be accommodated by improving the design of the capacitor, but this
does not eliminate the risk of resonance phenomena
between the power capacitors on the one hand and
the inductance of the feeding transformer and the grid
on the other.
If the resonance frequency of a resonance circuit
consisting of power capacitors and inductance of the
feeding transformer is near enough to the frequency
of a harmonic wave in the grid, this resonance circuit
can amplify the oscillation of the harmonic waves and
cause immense overcurrent and overvoltage.
The harmonic wave contamination of an AC voltage
supply can have some or all of the following effects:
• early failure of capacitors
• premature triggering of protective switches and
other safety devices
• failure or malfunction of computers, drivers,
lighting installations and other sensitive
consumers
• thermal overload of transformers caused by
increased iron losses
• overload of the neutral conductor (particularly
by the 3rd harmonic wave)
• shattering or bursting of discharge lamps
• thermal overload of the lamp choke due to resonance between choke and capacitor for power
factor correction. The effects can be similar to
asymmetrical mode (see chapter 6.2.9), which is
why the use of a choke with thermal protection
can also protect the luminaire from burning.
The installation of so-called choked capacitors (capacitor in series with a filter choke) aims at forcing the
resonance frequency of the grid below the frequency
of the lowest prevailing harmonic wave. This prevents
a resonance between the capacitors and the grid and
would thus also prevent the amplification of the harmonic currents. This kind of installation also has a filtering effect by reducing the level of voltage distortion
in the grid. It is therefore recommended for all cases
where the wattage share of the loads that generate
harmonic waves is more than 20% of the total wattage. The resonance frequency of a choked capacitor
always lies below the frequency of the 5 th harmonic
wave.
16
In the electronic ballast OSRAM POWERTRONIC® PTi,
the influence of harmonic waves on the lamp is kept
extensively at bay by the ballast design which comprises an intermediate circuit. The immunity of the PTi
input stage to line-related interference is safeguarded
by tests according to the IEC 61000 standard.
Such line-related interference includes e.g.:
• burst as per IEC61000-4-4, 1000V peak, repetition
frequency 5kHz, low-energy pulse
• current feed as per IEC61000-4-6, frequency
range 0.15-80MHz, 3Vrms
• surge as per IEC61000-4-5, 1000V symmetrical,
2000V asymmetrical, high-energy pulse
• voltage interruptions as per IEC61000-4-11
• voltage fluctuations
3.4 Brief voltage interruptions
When the lamp current falls, the recombination rate
starts to exceed the ionization rate, causing a reduction in plasma conductivity. This occurs with magnetic
ballasts in every half-wave on passing the zero crossing and results in the so-called re-ignition peak. When
recombination of the charged particles has progressed
far enough, the remaining quantity of charge carriers is
not sufficient enough to generate an adequate quantity of new charge carriers when the voltage increases
again – the lamp goes out. The high pressures in the
arc tube mean that the ignition voltage is now no longer sufficient to re-ignite the lamp. It has to first cool
down for a few minutes before it can ignite again (see
also chapter 4.2 “Warm re-ignition”).
When the supply voltage is interrupted, both the length
and depth of the interruption (100% for complete
interruption) and the phasing of the interruption are
important. Older lamps with their increase in lamp voltage and higher re-ignition voltage are more sensitive
than unaged lamps. The capacitor for power factor
correction can act as voltage source during voltage
interruptions, at least in the short term, and extend the
time in which a voltage interruption is tolerated before
the lamp goes off. Voltage interruptions just before the
zero crossing are more serious, because the plasma
has already cooled down significantly.
3.5 Stroboscopic effect and flicker
Operation of a metal halide lamp on a magnetic ballast
under supply voltage with 50 Hz frequency results in
periodic fluctuation of the luminous flux with double
the supply frequency. When the current flow drops
near the zero crossing, the plasma also has far less
radiation. But even on passing the zero crossing, the
luminous flux does not reach zero so that the plasma
still has on-going radiation.
The human eye reacts with differing sensitivity to
varying flicker frequencies, and can, for example, no
longer perceive fluctuations in luminous flux above
100 Hz. Literature provides differing ways of depicting
the sensitivity of the human eye for periodic luminous
flux fluctuations at various frequencies. Fig. 17 shows
an example according to Kelly and Henger [1].
According to Afshar [2], adapting the evaluation also to
short-term changes and implementation in a filter for
a light signal, such as in Fig. 15, results in values for
the flicker factor as shown in Fig. 16. The perceptibility
threshold is assumed to be 1. The values in this example remain below 1, i.e. no visible changes can be
perceived in the light.
Fig. 15: Luminous intensity of a metal halide lamp at
50 Hz choke operation, shown in arbitrary units
When operating at 50 Hz, the luminous flux or intensity
fluctuates with wattage, i.e. with 100 Hz as shown in
Fig. 15. Literature uses various equations to evaluate
changes in luminous intensity that can be perceived
by the human eye. Flicker is evaluated according to
EN 50006 standard, for example, with a flicker factor
F10 as
∑ m²(f )G ² f )
i
i
)
F10 =
i
whereby m(fi) = time-dependent modulation depth
of the luminous intensity
G
= filter curve for flicker sensitivity
depending on flicker frequency
Fig. 16: Flicker factor calculated from the luminous
intensity signal for a metal halide lamp at 50 Hz choke
operation, shown in arbitrary units
Fig. 17: Eye sensitivity curve for flicker as per Kelly
1960 and Henger 1985
17
There is a delay of just a millisecond between the
current maximum and the luminous flux maximum as
shown in the following drawing.
+200V
+8A
+7.97V
+400VA
V
MPower Power D C RMS: 150.52 VA
MImpedance Impedance D C RMS: Under range
3
1
4
2
TWindow 1 M--
-400VA
V
-0.033V
-8A
-200V
MM
Ch 1: Lamp voltage
Ch 3: Light signal (zero-line at the bottom)
Ch 2: Lamp current
Ch 4: Lamp power
Fig. 18: Time curve for light signal and the electric parameters of a metal halide lamp
In fast-moving or rotating objects, the stroboscope effect can cause an optical illusion that the object is
moving more slowly or in the opposite direction or even at a standstill.
Stroboscope effects can be reduced or ruled out by operating luminaire groups on three different phases
or by using electronic ballasts.
18
4 Igniting and starting discharge lamps
Some discharge lamps do not require an external
ignition unit, as the supply voltage is sufficient to
ignite the lamp or because the lamp has an integrated
ignition unit. These lamps must not be used in installations with an external ignition unit or they will fail prematurely due to internal arcing.
4.1 External ignition units
4.1.1 Parallel ignition unit
Pulser
ignition unit
choke
All other discharge lamps must be ignited by an additional unit. Ignition units or circuits of varying types are
used for this purpose.
At room temperature, the filling particles are still present in solid form (metal halides or amalgam) or in liquid
form (mercury). The arc tube contains the start gas,
usually an inert gas such as argon or xenon, between
the electrodes. The insulating gas filling in the arc
tube must be made conductive in order to generate
hot plasma. This is carried out by high-voltage pulses
generated by a separate ignition unit or by the ignition unit in an electronic ballast. Constantly available
free charge carriers (electrons) are accelerated by high
voltage, providing them with sufficient energy to ionize
atoms on impact and generate more free charge carriers. This process, similar to an avalanche, finally produces conductive hot plasma within which the current
flow excites the partly evaporated metal halide filling
such that light is radiated.
Luminaire
US
Lamp
Capacitor for
PFC
Fig. 19: Simplified circuit diagram for conventional
operation of high intensity discharge lamps with pulse
ignition unit
With a pulse ignition unit, the choke must be insulated
for its surges. The pulse ignition units can normally
take a load of 1000 pF, permitting lead lengths of
about 15 m between lamp and choke. During ignition,
the lead carries high voltage from the choke to the
lamp so that care must be taken to ensure that the
supply lead, socket and luminaire are adequately insulated for the corresponding high ignition voltage.
This type of ignition unit is used in single phase grids.
4.1.2 Semi-parallel ignition unit
The ignition voltage required to generate a breakdown
between the electrodes depends on the spacing between the electrodes, the filling pressure of the gas
between the electrodes and the type of gas. Examples
for using these principles include the use of auxiliary
electrodes or the use of Penning gases (see chapter
14.4 “Ignition at low ignition voltage (Penning effect)”).
The sockets and cables must be suitably designed for
the high ignition voltages. In particular with the E27
sockets for single-ended screw base discharge lamps,
care must be taken that a similar socket (E27) for incandescent lamps is not used, which does not meet
the requirements.
When lamps are defective or no lamp is inserted,
continuous operation of the ignition unit can possibly
damage the ignition unit or the luminaire. It is therefore
advisable to switch the ignition unit off for a period of
time after failed ignition, or to use an ignition unit with
timer function.
Preference should be given to using a timer
ignition unit.
Semi-parallel
ignition unit
choke
Luminaire
US
Capacitor for
PFC
Lamp
Fig. 20: Simplified circuit diagram for conventional
operation of high intensity discharge lamps with a
semi-parallel ignition unit
In the semi-parallel ignition unit, part of the choke
windings is used to transform the ignition pulses. This
means the choke must be adequately insulated for the
high voltage and have a tap for the ignition unit. As
with pulse ignition units, the ignition unit can generally take 1000 pF or approx. 15 m lead length, and
the connection lead between choke and lamp must
be insulated for the corresponding voltage levels. A
capacitor with minimum capacitance depending on the
unit must be provided for compliance with the EMC of
the ignition unit.
19
4.1.3 Superimposed ignitor
with symmetrical and asymmetrical ignition pulses. In
the asymmetrical units, care must be paid to correct
polarity of the lamp connections!
Superimposed
ignition unit
choke
Luminaire
US
Capacitor for
PFC
Lamp
Fig. 21: Simplified circuit diagram for conventional
operation of high intensity discharge lamps with a
superimposed ignition unit
In a superimposed ignition unit, the high voltage is
only present at the lamp outputs of the unit. Depending on cable and structure, ignition units of this type
can normally take loads of 100 pF, corresponding to
a lead length of about 1.5 m.
4.2 Warm re-ignition
Normal ignition voltages in the range up to 5 kV do
not permit immediate re-ignition of a lamp which is
still hot. The high operating pressures demand ignition
voltages of e.g. 25 kV. If a lamp goes out for instance
because of a brief interruption in the supply voltage, it
must cool down for a few minutes (for lamp wattages
< 150 W) until the falling pressure in the arc tube permits re-ignition for normal ignition units in the 5kV
range. Higher wattage levels require considerably
longer cooling down periods because of the higher
thermal capacity. The cooling down process depends
also on the ambient temperature and the luminaire. A
narrow, hot luminaire prolongs the cooling down procedure, delaying re-ignition of the lamp.
This cool down time must be considered for ignition
units with a timer cut-out which are designed to shut
off after a certain period of time with failed ignition.
The design intent assumes that the lamp is defective
or not inserted. The selected timer period must be
sufficient to allow the lamp enough time to cool down
following a power interruption so that lamp re-ignition
is again possible. The warm re-ignition times of the
POWERBALL ® are under 10 minutes, far shorter than
those of the cylindrical version.
The timer period for ignition units with warm
re-ignition must be appropriately long enough.
4.3 Hot re-ignition
High ignition voltages of 16 to 60 kV are necessary for
immediate re-ignition of hot metal halide lamps (hot
re-ignition) on account of the high vapor pressures.
The lamp, sockets and luminaire must be designed for
these high voltages, and suitable ignition units must
be used. There are two versions of hot re-ignition units
20
At present, hot re-ignition is permitted for doubleended quartz lamps (with a few exceptions). As far
as ceramic lamps are concerned, the single-ended
HCI®-TM series with GY22 socket are approved for
hot re-ignition. Approval of other double-ended ceramic lamps is in preparation.
4.4 Ignition at low ignition voltage (Penning effect)
Various methods can be used to reduce the voltage
necessary to ignite the lamp. One such method uses
the so-called Penning effect. When the energy stored
in a meta-stable excitation level of the basic gas is
larger than the ionization energy of the admixture, volume ionization begins at lower field strengths, resulting in a larger number of charge carriers at the same
voltage than in pure gas. Examples for the Penning
effect include mercury in argon for metal halide lamps
and argon in neon for some discharge lamps.
4.5 Ignition at low ambient temperatures
Most metal halide lamps with wattages < 400 W
can be operated at ambient temperatures of – 50 °C.
The usually evacuated outer bulb and the luminaire
ensure that the arc tube is thermally decoupled
from the surroundings, so that the normal operating
parameters are achieved as far as possible. For
HQI ® 2000 W NI and 2000 W DI, ignition is only
permitted to – 20 °C.
In the ignition units, the ferrite core is sensitive to temperature, that is, the rated ignition voltage is lower at
lower temperatures. Some ignition unit manufacturers
recommend the use of ignition units without cut-out,
as the ferrite core in this case is heated up by the losses during failed ignition attempts so that the specified
ignition pulse levels are reached again. There are also
ignition units with an additional integrated resistor for
heating the ignition unit so that these are approved
for temperatures down to – 50 °C. Here again, it takes
a while after switching the ignition unit on until it has
heated up enough to reach the specified ignition pulse
levels.
For specific applications such as in cold storage houses, semi-parallel ignition units can be used (which permit longer lead lengths) that are fitted in warmer zones
outside the luminaires.
4.6 Cable capacitance
4.7 Start-up behavior of metal halide lamps
The capacitance of the supply cables between lamp
and ignition unit depends on various general conditions. These include the size and structure of the
cable (diameter, distances and insulation together with
number of individual cables, dielectric coefficients of
the materials). The capacitance also depends on the
grounding and shielding of the cable and where it is
fastened, e.g. close to grounded surfaces. Commonly
used power cables are not suitable for this purpose,
because of their relatively thin PVC-isolation, where
the wires lie comparably close together. The capacitance here is about 100 pF/m. Because of the high
ignition voltage for discharge lamps the lead wires
have thicker insulations and they are normally not
placed close to each other. The capacitance of the
lead wires will therefore be lower than for the power
cable.Capacitances only form limited resistance to
high-frequency voltage components of the ignition
pulse. The capacitance attenuates the ignition pulse,
with resulting ignition pulses possibly not reaching the
amplitudes required to ignite the lamp. Certain load
capacitances must therefore not be exceeded, depending on the specifications of the ignition unit.
After igniting the lamp and heating the discharge, the
discharge runs initially only in the start gas. The mercury and the metal halides are still in liquid or solid
form on the arc tube wall. The voltage across the
discharge is initially still very low. The start gas argon
radiates a little in the visible range (weak violet light),
which is why the luminous flux in the initial phase is
still very low.
Through power consumption in the lamp, first the mercury and then also the metal halides begin to evaporate. The individual filling particles evaporate at different rates, resulting in differing ratios of the particles
during runup. The dominance of individual particles
in the start-up phase results in the colour phenomena
during this period shown in Fig. 22. Only after a few
minutes, having reached the steady state, is the required composition achieved, producing the full luminous flux and the required light colour.
after 40s
after 20s
CRI=66
after 80s
after 60s
after 100s
CRI=70
CRI=36
CRI=29
after 120s
CRI=85
CRI=92
Fig. 22: Course of light parameters of a HCI®-T 150 W/NDL during start-up
21
The new round ceramic arc tube (POWERBALL ®) has
a uniform wall thickness without thick ceramic plugs
as in the cylindrical ceramic type. The mass is therefore only about half that of the cylindrical version. This
means less energy and therefore less time is needed
to bring the POWERBALL ceramic arc tube up to operating temperature. The times required to achieve the
lit-up status are therefore clearly shorter than in the
cylindrical version, as shown in Fig. 23.
The time it takes to reach a steady state depends on
the start-up current and the associated wattage input.
If the current is too high, the electrodes will be damaged, causing the walls to blacken. The standard for
metal halide lamps (IEC 61667) therefore limits the
start-up current to twice the nominal lamp current.
With OSRAM POWERTRONIC®, the start-up is faster
than with a conventional ballast, as shown in the following Fig. 24.
HQI-T 150 W at OSRAM PTU
HQI-T 150 W at CCG
Luminous flux in lm
Luminous flux in lm
cylindrical HCI-T 150 W
Time in s
Fig. 23: Start-up behavior of luminous flux in various
metal halide lamps operating with an OSRAM electronic ballast
22
Time in s
Fig. 24: Start-up behavior of luminous flux of a HQI®-T
at various ballasts
5 Reducing the wattage of high intensity discharge lamps
5.1 Introduction
5.2 Wattage reduction techniques
High intensity discharge lamps generate light by exciting mercury and other metals within an arc tube into
a plasma generated by the current flow between two
electrodes.
The following dimming methods are generally known
(by conventional means or electronic ballast):
• Reduction in supply voltage
• Phase control: leading edge, trailing edge
Discharge lamps must be operated with a ballast and
are rated for a certain lamp wattage. Either conventional chokes or electronic ballasts can be used.
• Increase in choke impedance or decrease in lamp
current (amplitude modulation)
• Change in frequency for high-frequency operation
To change the nominal lamp wattage of a lamp, the
following general physical conditions are significant for
the resulting effects:
• The electrodes of discharge lamps are rated for
a certain lamp current. If the current is too high,
parts of the electrodes melt and evaporate. If
the current is too low, the electrode is operated
in cold state. This changes the mechanisms for
releasing electrons from the electrode with more
electrode material being deposited on the tube
wall. Deviations in lamp current from the nominal value in both directions can therefore cause
blackening of the arc tube wall with a decline
in luminous flux, together with negative effects
on the light colour and possibly also on the service life.
• The partial vapor pressure of the filling particles
responsible for generating light depends on the
temperature of the arc tube wall. A change in
the arc tube wall temperature resulting from a
change in lamp wattage influences the composition of the filling in the plasma arc and thus the
electrical and photometric properties of the
lamp.
5.2.1 Reducing the supply voltage
A reduction in supply voltage beyond recommended
limits (see sections 3.1.2 and 3.1.3) will decrease the
lamp wattage. Reducing lamp wattage results in decreased lamp voltage and re-ignition peak voltage, and
is generally to a lesser extent than the supply voltage.
This reduction in the gap between the re-ignition peak
and the supply voltage makes it more probable that
the lamp will go out. This applies particularly to aged
lamps where the lamp voltage and re-ignition voltage
have already increased.
Fig. 25 shows, as an example, the behavior of certain
lamp types on reducing the supply voltage. Here the
ratio of re-ignition voltage to effective supply voltage
(ULS/US) has been standardized to 1 for 220 V supply
voltage. It can be seen that when the supply voltage decreases, this ratio generally assumes values of
greater than 1. This also means that the gap between
ULS/US referred to the ratio at 220 V
HCI-TM 250 W/WDL
HCI-TS 70 W/WDL
HCI-TT 150 W/WDL
HQI-TS 150 W/WDL
• At higher arc tube wall temperatures, the metals
do not recombine with the iodides and the pure
metals can migrate into the wall (applies to quartz
arc tubes).
• Drop in luminous flux through blackening
of thearc tube
• Change in color properties
• Reduction in service life
1,4
ULS/US referred to the ratio at 220 V
Wattage reduction has the following side effects:
HQI-TS 150 W/NDL
1,3
1,2
1,1
1
0,9
140 150 160 170 180 190 200 210 220 230 240 250 260
Supply voltage US in V
Fig. 25: Relative change in the re-ignition peak (ULS )
to supply voltage (US ) referred to the ratio at 220 V for
various metal halide lamps
23
re-ignition voltage and the current supply voltage decreases. If the re-ignition voltage exceeds the supply
voltage, the lamp goes out (see also chapter 6.2.2
“Increase in re-ignition peak”).
This means that POWERBALL HCI® must not be
dimmed by reducing the supply voltage, as the
re-ignition peak can cause earlier extinguishing of
the lamp or flicker.
5.2.2 Phase control: leading edge, trailing edge
Fig. 26 and 27 show the decrease in effective supply
voltage by phase control with leading edge or trailing
edge. There are also variations in which the supply
voltage is reduced in the middle and not before or
after the zero crossing. In other versions, the supply
voltage in the leading edge phase is only decreased
and not reduced to zero.
U,I
␣
␭
UB
UL
␣
␭
UB
UL
Supply voltage
Lamp voltage
Ignition angle
Current flow angle
␣
UB
UL
UB
UL
␣
␭
Supply voltage
Lamp voltage
Ignition angle
Current flow angle
t
The least disadvantages are to be expected by reducing current in rectangular mode. The steep zero crossings mean that lower re-ignition peaks and less blackening from sputtering can be expected.
5.2.4 Change in frequency for high-frequency mode
t
Fig. 27: Principle of phase control with trailing edge
(idealized diagram)
A change in wattage when using an inductive ballast
can also be achieved by varying the frequency of the
power supply, as the inductive resistance of the choke
depends on frequency. The change in choke impedance at low frequencies has been dealt with in the
preceding chapter 5.2.3.
For the phase control with trailing edge or other
methods where supply voltage is temporarily switched
off or reduced, suitable means are required to provide
an uninterrupted, “smooth” lamp current to prevent the
lamp from flickering and going off.
If the change in impedance is caused by changing the
frequency in radiofrequency operation, in discharge
lamps the possible occurrence of acoustic resonances
has to be considered. Resonance in the discharge
tube can cause the plasma to start to oscillate depending on the arc tube geometry and plasma temperature. This can cause the lamp to flicker or go off,
and in extreme cases, destruction of the lamp should
the arc attach to the arc tube wall due to severe resonance. This is why a currently proposed standard for
electronic operation of metal halide lamps limits the
amount of high-frequency oscillations.
Increased blackening and therefore a drop in luminous
flux must be expected in all versions compared to fullload operation.
It is difficult to find reliably resonance-free operating windows, for various reasons: the resonance
frequencies change during the start-up and also dur-
For the phase control with leading edge, the resulting intervals with no current result in a greater cooling
down of plasma end electrodes, thus increasing the
re-ignition peak, causing the lamp to go off earlier.
24
l
If a switchover to other chokes is used for dimming
lamps with a wattage > 400 W, they must be left to
burn at 100% for at least 1 hour.
Fig. 26: Principle of phase control with leading edge
(idealized diagram)
␭
Increasing choke impedance reduces the current
through the lamp. The supply voltage remains the
same so that the voltage is still high enough to reignite the lamp. The flatter zero crossing of the current
can however be expected to cause greater cooling
down of plasma and electrodes, with greater blackening as a result of the processes at the electrode during
re-ignition. The blackening therefore causes a greater
drop in luminous flux compared to full-load operation.
Fig. 28: Amplitude modulation e.g. by choke changeover
t
U,I
5.2.3 Increasing choke impedance or decreasing
lamp current
ing the course of the service life. Lamps of differing
geometries and filling also show different resonance
frequencies. A reduction in wattage also changes the
resonance frequencies due to the change in plasma
temperature.
5.3 Recommendations for reducing the wattage in
discharge lamps
The PTo with squarewave operation and optimised ignition runs the POWERBALL HCI® lamps ideally down
to 60% of the lamp output (rated value). No significant
negative effects arise even when the output is reduced
to 85% of the rated output.
Even when operated at between 85% and 60% of the
rated output, this does not negatively effect the failure
rate. However, increasingly the lamps have a slightly
green touch and the colours may deviate from each
other (colour spread).
5.3.1 Metal halide lamps:
Operation of OSRAM POWERSTAR HCI® (cylindrical burner) and OSRAM HQI® lamps with reduced
power is not permitted as this can cause considerable colour deviations, much poorer luminous flux
maintenance and shorter service lives
Basically it is possible to dim POWERBALL HCI®.
While the higher thermal load capacity of the round
ceramic arc tube permits an improved dimming behaviour in terms of luminous efficacy and colour rendering
compared to metal halide lamps with quartz arc tube
or with the normal cylindrical ceramic tube, dimming
still causes a shift in the chromaticity coordinate.
Dimmed lamps show a greater drop in light output and
wider colour spread throughout the service life.
It is important to avoid these effects for interior lighting. They are more apparent when operated on conventional control gear than on electronic control gear.
OSRAM therefore recommends not to reduce power
with currently available lamps on conventional control
gear or for indoor lighting.
The type of dimming has a great impact on the results.
It is recommended to use only adjustable Electronic
ballasts with squarewave operation and to completely
avoid dimming via reducing the supply voltage or
leading-edge or trailing-edge phase dimmers. It is not
possible to guarantee that dimmed lamps will be able
to meet the product properties.
In any case, the lamp should run for at least 15 minutes
with 100% wattage after being switching on so that
the lamp can light up correctly.
A warranty regarding lifetime can only be given
when approved POWERBALL HCI® units (cf. online
catalogue) on the POWERTRONIC® PTo are dimmed.
Operation of POWERBALL HCI® on the
POWERTRONIC® PTo:
The combination of POWERBALL HCI® and
POWERTRONIC® PTo allows energy-saving operation
everywhere where optimised colour rendering is not
important, for example outdoor lighting.
The luminous flux drops slightly more throughout the
service life in dimming mode than when operated at
100% on the PTo. This effect can be reduced if the
lamps are operated via a combination of dimming and
100% operation.
Dimming causes a reduction in light output and
colour change.
Squarewave operation is recommended for
dimming.
For outdoor lighting: optimised operation
of the approved POWERBALL HCI® on the
POWERTRONIC® PTo.
There is no warranty for dimmed POWERBALL
HCI®. A warranty regarding lifetime can only
be given when approved POWERBALL HCI ®
units (cf. online catalogue) are dimmed on the
POWERTRONIC® PTo.
5.3.2 Dimming of other discharge lamps
High pressure mercury lamps:
These lamps can be dimmed to 50% of the rated
wattage, whereby they must be started up with 100%
wattage. Dimming is possible by voltage reduction,
phase control and amplitude modulation.
High pressure sodium lamps:
It is possible and allowed to reduce the power of all
VIALOX® NAV® down to 50% of the lamps rated values
without impact on the service life
• via step switching by changing to inductive
control gear with the next lower rating or
• via step switching with additional inductance,
whereby in both cases electronic power switches must
be used.
When reducing the power, ensure that the lamps are
started and operated at their rated values for approx.
10 minutes before dimming.
It is not permitted to reduce the power by leading
edge phase control or reducing the mains voltage.
OSRAM recommends the electronic ballast
POWERTRONIC® PTo for dimming operation.
25
6 Lamp service life, aging and failure behavior
6.1 Lamp service life and aging behavior
All lamp-specific electrical and photometric data are
ascertained after operating for 100 hours under laboratory conditions using reference ballasts (according to
IEC). The service life data are determined under controlled laboratory conditions with a switching rhythm of
11 h on/1 h off. In practice, noticeable deviations can
occur due to deviating supply voltage, ambient temperature and other general conditions. For metal halide
lamps, there can be individual differences in colour
from lamp to lamp, caused by external influences such
as supply voltage, control gear, burning position and
luminaire design.
Unless stated otherwise, the specifications apply to
TS types for horizontal burning position and to T and
E types up to 250 W for base up burning position. For
lamps > 400 W, the horizontal burning position applies
for the T-lamp. For lamps with 400 W, the burning position depends on the type (as stated in the catalogue).
Should deviating burning positions be used in practice, this could cause changes in luminous flux, colour
temperature and service life. The POWERBALL HCI®
with its round arc tube is less critical than conventional
cylindrical ceramic tubes.
The luminous flux is generally independent of the ambient temperature itself outside the luminaire. However,
too high ambient temperatures can cause increased
arc tube blackening in the long run. Also, special ignition units are required for lower ambient temperatures
down to approx. – 50 C°. HQI®-2000 W lamps with integrated auxiliary discharge are only permitted to – 20 °C.
For measurement of the electrical, photometrical and
colour characteristics, HQI®-TS and HCI®-TS lamps
shall be operated within a luminaire simulator. Details
on luminaire simulator (Quartz tubes around the lamp)
for determining the lamp data for HQI®-TS and HCI®TS can be found in IEC 61167, Annex B.2.
The mean service life stated in the documents (B50 value) is the burning time within which maximum half of the
lamps can have failed, i.e. the survival rate at this point
in time is at least 50%. This is a value normally indicated
by all lamp manufacturers. Apart from the B50 value,
it is also common practice to indicate the times e.g. at
which 10% or 3% of the lamps failed (B10 or B3).
Mean service life (B50):
max. half the lamps have failed.
Economic life:
on account of the decrease in luminous flux and
the increasing failure rate, the illumination level of
the installation has fallen below a required value.
26
The economic life is obtained by including the decrease in luminous flux over the service life in the
calculation. Multiplying the survival rate by the maintenance of the luminous flux provides the decline in
luminous flux of the installation. These factors are
considered when preparing a maintenance schedule
according to EN12464 (see also 7.5 Maintenance of
lighting systems with metal halide lamps).
Data on lamp survival behavior and luminous flux behavior can be found in the corresponding Technical
Information.
One main reason for the reduction in luminous flux is
blackening of the arc tube by electrode material which
has settled on the tube wall throughout the service life.
Frequent switching, overload operation, use in confined luminaires or high ambient temperatures can add
to this blackening process and thus clearly reduce the
service life. Operation at reduced wattage also causes
increased tube blackening, as explained in chapter 5
“Reducing the wattage of high-intensity discharge
lamps”.
6.2 Storage of metal halide lamps
Incorrectly stored lamps (e.g. damp and warm) may
suffer corrosion on the contacts after a while; this oxidation must be completely removed before the lamps
are used. In unfavourable conditions, this may even
result in ignition problems. Lamps with filler material
may lose this filler material if stored incorrectly and the
socket contacts may then become exposed. There is a
risk of arc-overs during the ignition process or the risk
of touching live parts.
6.3 Failure mechanisms of metal halide lamps
The following failure mechanisms are possible for
metal halide lamps and the probability increases as the
lamp gets older.
• Leaking arc tube
• Increase in re-ignition peak, finally the lamp goes out
• Broken leads
• Leaking outer bulb
• Ignition failure
• Breakage or wear of the electrodes in the
arc tube
• Scaling of the base contacts by arcing in
the socket
• Bursting of the lamp
Glow discharge
Arc discharge
Incandescent lamp mode
Fig. 29: Various states of outer bulb discharge
6.3.1 Leaking arc tube
High temperatures and pressures in the arc tube, the
aggressive chemical substances in the tube and the
thermal cycling of a lamp place extreme strains on
the arc tube. This can cause the tube to leak, allowing
starting gas and filling particles to enter the outer bulb.
Depending on the size of the leak, this effect is usually
a gradual process. It is initially noticed by a considerable change in the light colour. Increasing leaks of
starting gas into the outer bulb can result in the discharge process moving from the arc tube to outer bulb
discharge.
– For lamps with evacuated outer bulb, various abnormal discharge states can occur, depending on
tube filling pressure and outer bulb volume.
– For lamps with gas-filled outer bulb, usually
lamps > 400 W, glow discharge and incandescent
mode do not occur. Particularly in lamps operating with ignition units, the described faults result
in a direct arc discharge. In extreme causes, this
can cause the lamp to burst.
In the case of glow discharge, the voltage across
the lamp is high but only very low current. Sputtering
causes material to be deposited on the outer bulb. It is
possible for glow discharge to precede arc discharge.
The temperatures in the pinched area are lower than in
normal operation.
In the case of arc discharge, the voltage across the
lamp is low and the current is limited by the choke.
The attachment of the arc onto the leads in the outer
bulb can cause these to melt. The high temperatures
cause the material of the leads to evaporate and then
settle on the outer bulb. The hot arc near the pinching area can result in high temperatures (in extreme
cases they can exceed 800 °C). At the contact between
socket and lamp holder and at the electrical contact, the
temperatures are naturally much lower. Here in extreme
cases 300 °C were measured at the contact between lamp
and lamp holder and 250 °C at the electrical contact of the
lamp pin to the lamp holder. The electrical contact is also
relevant for the Temperature Code of the socket (see also
chapter 7.3 lamp holder).
If metallic coatings in the pinching area form through
material deposition from the leads so that they form a
continuous conductive layer between the leads, then
the result in the so-called incandescent mode. The
metal coating offers sufficient resistance that power is
consumed and the coating begins to glow. It is hereby
possible that electrical values similar to normal operation are reached, which would make it impossible for
an electronic ballast for example to detect this abnormal situation. This also causes high temperatures in
the pinching area.
Glow and arc discharges can be detected by current
and voltage values deviating from the normal levels, so
that an electronic ballast with a corresponding automatic
cut-out feature can switch off such lamps. In addition,
the luminaire design must use components resilient to
high thermal loads so that the possibly high temperatures will not lead to harmful situations for the operator.
6.3.2 Increase in re-ignition peak
The re-ignition peak is a peak in the lamp voltage after
the zero crossing of current and voltage. For sinusoidal lamp current, the current decreases gradually before the zero crossing. As the plasma is heated by the
current flow, a decrease in current causes the plasma
to cool down and reduces its conductivity. After the
zero crossing, the cooled plasma can initially no longer
conduct the current through the lamp. As the current
does not rise through the lamp, an increasing amount
of supply voltage falls across the lamp. The rise in
voltage causes the ionization of the plasma and therefore the current to increase again, meaning the plasma
is reignited, hence the name “re-ignition peak”. If the
re-ignition peak exceeds the level that can be provided
by the supply voltage, the lamp goes out.
27
This is one of the advantages of the rectangular electronic ballast. As the zero crossing for current is very
steep, the events of limited current availability are
very short and the plasma has little chance to cool
down.
Lamp voltage
Supply voltage
Supply voltage
Lamp voltage
Lamp current
Lamp current
Re-ignition peak
Fig. 31: A lamp goes out because the re-ignition peak
is too high
Fig. 30: Re-ignition peak, supply voltage and lamp current
The lamp voltage and the re-ignition peak increase
with progressing lamp age; in addition, this parameter
also depends on the ambient temperature and increases while the lamp is heating up. This results in what is
known as cycling, where the lamp periodically goes off
and on again. The re-ignition voltage increases while
the lamp is heating up and continues to increase until
the luminaire is completely heated through. This is why
it can happen that the lamp does not go off until after
several or even many minutes of burning time.
Fig. 31 shows a lamp with high re-ignition peak. After
the zero crossing, the current barely starts to flow. This
is why the voltage loss across the choke is low and
nearly the entire voltage supply falls across the lamp,
with lamp voltage following supply voltage. The current flow decreases even further from period to period,
so that conductivity continues to fall; in the end, the
voltage required to re-ignite the plasma is higher than
the supply voltage → the lamp stays off after the zero
crossing.
A decline in the supply voltage can also cause the
lamp to go out. It is only when the lamp has cooled
down sufficiently that re-ignition is possible with the
normal ignition units. After “cycling” for a long time,
it is possible that the lamp will not ignite at all.
This fault is not critical if the ignition unit does not
suffer from the frequent ignition attempts.
6.3.3 Broken lead or broken weld
This can be caused by material fatigue or extreme
mechanical load. Normally this is a non-critical fault;
in very rare cases, a loose contact can cause high
induced voltages.
Lamps with gas-filled outer bulb for supply voltages of
400 V can form an arc when a lead is broken or a weld
comes loose. Due to the current-limiting choke, this
arc can persist for a longer period of time and cause
the lamp to burst. Such arcing occurs both in lamps
with an ignition unit and lamps with auxiliary starter
electrode (lamps for ignition at supply voltage with a
wattage of 2000 W).
6.3.4 Leaking outer bulb
Mechanical impacts can cause the outer bulb to leak
so that air penetrates. Given the high temperatures,
leads oxidize when oxygen is present, causing an
open in the circuit. This is a non-critical fault; the lamp
no longer ignites. Ignition units without cut-out can,
however, fail prematurely due to permanently generating ignition pulses.
6.3.5 Lamps that do not ignite
This can result from open electrical connections within
the lamp or extreme aging and is actually a non-critical
fault. Ignition units without cut-out can, however, fail prematurely due to permanently generating ignition pulses.
28
6.3.6 Breakage or differing wear of the electrodes
Breakage of an electrode or differing wear in the electrodes with choke operation can cause a flow of asymmetric current with DC components, which can result
in the choke overheating. This effect of asymmetrical
conductivity is dealt with in greater detail below.
A broken electrode in a ceramic lamp can cause leaks
in the arc tube as a result of overheating capillaries,
with the effects described above. In rare cases, a discharge attachment near the arc tube wall in ceramic
lamps can cause the arc tube to burst.
A broken electrode in a lamp with quartz arc tube can,
after a longer period of time, cause the arc tube wall to
bulge and possibly leak or burst, if the discharge still
persists.
6.3.7 Scaling of the base / socket
Particularly in the case of old ignition units without
automatic cut-out and aged lamps or soiled contacts,
high transition resistances can cause oxidation and
thus overheating of the contacts. When ignition pulses
persist for a longer period of time, if lamps have gone
out because the re-ignition peak is too high or if the
lamp has not ignited, it is possible for arc-over to occur in the socket. If scaling has occurred, the socket
must be replaced as well as the lamp.
Vibrations can cause the lamp to become loose with
the possibility of arc-overs and scaling caused by the
resulting poor contact. In these cases the usage of a
lamp holder with retention device is recommended, as
described in the standard IEC 60238 “Edison screw
holders”, section 2.23 “Lamp holder with retention device”. The test conditions are described in section 12.14.
Suspending the luminaire on a chain attenuates
vibrations compared to suspending it on a rope.
6.3.8 Bursting of the lamp
It is generally possible for the arc tubes of metal halide
lamps to burst. This is very rare for ceramic metal halide lamps; the probability is greater in lamps with a
very old quartz arc tubes. With progressing age, the
quartz crystallization increases, making it brittle. However, the lamps normally fail by going out.
During operation, the arc tube is under great pressure. When the arc tube bursts, fragments can fly at
great speed, destroying the outer bulb when they hit it.
When the outer bulb is broken after the tube has burst,
very hot fragments of the arc tube come into contact
with the luminaire.
OSRAM therefore strictly differentiates between lamps
for open and closed luminaires. Lamps for open luminaires have a mechanical safeguard around the arc
tube to ensure that all fragments remain intact within
the outer bulb should the arc tube burst. Compliance
with this requirement is ensured through inhouse tests
at OSRAM, which are much more stringent than in
actual operation and in some published standards, for
example, ANSI Standards.
This is the corresponding pictogram for
lamps of this kind as per IEC 62035.
As it is generally not possible to rule out the possibility of the lamp bursting for all other lamps, metal
halide lamps must be operated in closed luminaires,
which are designed to contain all hot
fragments of the lamp in the case of it
bursting. The corresponding pictogram
for lamp and luminaire according to
IEC 62035 is shown on the right.
Silicate glass panes are recommended as a cover
screen. When plastic screens are used, it is important
to ensure that the hot parts of the lamp will not melt or
set fire to the screen should the lamp burst.
The cover screen must be both resistant to temperature change and break-proof.
6.3.9 Rectifying effect
High intensity discharge lamps can assume an asymmetrical mode (rectifying effect). There are various
possible causes:
• Differently heated electrodes:
This is typical when the lamp starts, but is normally only short-lived. The DC component sends
the choke into saturated state, the magnetic
resistance decreases and current is limited to a
lesser extent, shown as an example in Fig. 32.
This effect is described in the standards as „inrush current“ (IEC 61167).
• Malfunction of one of the electrodes:
This can be caused by differently worn electrodes, or in rare cases by a broken electrode.
The result is longer asymmetrical lamp current,
or if an electrode has broken off, a permanently
asymmetrical lamp current.
29
The effects are similar to rectifying effect at the
start, but the longer persistence can cause overheating of the choke and ignition unit.
• Discharge in the outer bulb:
As the leads are not geometrically the same, the
discharge generated between them can be asymmetrical, with the effects described above.
Fig. 32: Asymmetrical conductivity with lamp current and lamp voltage during a normal lamp start. This is only a
transient effect which causes no harm.
30
Rectifying effect causes a high DC current component.
As a result, the choke goes into saturated state with
a marked decrease in choke impedance. In extreme
cases, the lamp current is only limited by the choke’s
ohmic resistance.
Permanently excessive current causes a dramatic increase in the temperature of the choke windings until
the insulation is destroyed and short circuits occur
between the choke windings.
These phenomena can occur with metal halide lamps
(see warning in IEC 61167) so that the standards have
stipulated safety measures for luminaires (see IEC
60598-1 Paragraph 12.5.1). Similar regulations exist
for high intensity sodium vapor lamps in the IEC 60662
standard.
A safety measure in the circuit such as a thermal
switch or a thermal fuse, integrated into the magnetic
ballast, protects the circuit from such damage.
In accordance with a declaration issued by lamp manufacturers in reaction to standard EN 62035, published
by the LIF (Lighting Industry Federation Ltd) in Technical Statement No. 30, and by the ZVEI in the “Lamp
manufacturers statement regarding EN 62035”, certain
lamps do not require any safety measures to prevent
asymmetrical conductivity. As far as OSRAM’s metal
halide lamps are concerned, this refers to lamps with
wattage levels of 1000 W and more.
Although asymmetrical conductivity is generally possible in lamps with wattages > 1000 W both during the
start and in steady state, the dimensions of the arc
tube and lamp components means that the tendency
to asymmetrical conductivity is far less than with
smaller wattage levels, and is weak enough in steady
state that no safety measures against asymmetrical
conductivity are required.
OSRAM PTi is not affected by asymmetrical conductivity, as current and voltage are monitored and controlled, which is why it is recommended for the operation of discharge lamps.
6.3.10 Conclusions
• Safe operation of metal halide lamps depends
on the use of luminaire parts (lamp holder, leads
etc.) which can withstand the high temperatures
that can possibly occur in the case of outer bulb
discharge.
• Apart from burst-protected lamps for operation
in open luminaires, all lamps must be operated
in closed luminaires.
• All metal halide lamps with small wattage must
be equipped with a safeguard to protect the
lamp from the effects of asymmetrical conductivity (e.g. chokes with thermal protection).
• It is advisable to use ignition units with time
cut-out.
• The use of electronic ballasts is beneficial if the
electronic ballast has a corresponding cut-out
mechanism.
It is wise not to operate metal halide lamps right up
to the end of their natural service life, but to replace
them at the end of the economic life. This is appropriate because the luminous flux decreases noticeably
on exceeding the economic life, and the probability of
undesirable effects increases at the end of the service
life.
Lamps should be replaced before reaching the
economic life if
• the light colour of the lamp changes noticeably,
• the luminous flux decreases considerably,
• the lamp will not ignite,
• the lamp goes on and off intermittently (“cycling”).
On complying with all safety measures, metal
halide lamps are safe to operate and provide a
brilliant, efficient light.
31
7 Luminaire design and planning of lighting systems
7.1 Measuring temperatures, ambient temperature
7.1.1 General physical conditions for temperature
limits for outer bulbs and pinches in metal
halide lamps
Outer bulb
Foil
Radiation
Heat flow
When the limit values for pinch temperature or outer
bulb temperature for hard and soft glass are exceeded,
the following can be expected:
Arc tube
Capillary
• the foil oxidizes
• the evacuated outer tube collapses and the
gas-filled outer tube blows out when the glass
turns soft
Fig. 33: Diagram to show thermal coupling of arc tube
and outer bulb
7.1.2 2 Measurement with thermocouple
• the cement in screw-base lamps crumbles
Although quartz glass can withstand far higher temperatures than the stated limit values, it is possible for the
arc tube to overheat when the stated limit values for
the outer bulb are exceeded, with the following effects:
Measurement with thermocouple is a simple, practical
method of obtaining measured values.
• change in colour properties
• leaking arc tube
• blackening of the arc tube and deterioration in
luminous flux maintenance
The outer bulb temperature is only an indirect indicator for arc tube load! Outer bulb and arc tube are
coupled by radiation and to a small extent by thermal
conduction via the leads.
Fig. 34: Affixing the thermocouples to the outer bulb
and the lamp base
While it is important to limit the outer bulb temperature, it is worthy to note that an inadequately designed
reflector can still cause the arc tube to overheat without any major change in the outer bulb temperature
(see also chapter 7.9 “Optical design of reflectors”).
One indication that the design of a luminaire is insufficient comes from comparing the lamp voltage
measured when burning freely outside the luminaire,
and inside the luminaire after being left long enough
to stabilize. The increased lamp voltage measured in
lamps < 400 W in the luminaire should not exceed 5 V.
There is no notable difference in the lamp voltage
inside the luminaire for lamps with wattage input
> 1000 W, outer bulb and free-burning operation. But
soiling of the outer bulb caused by evaporation from
luminaire components can cause the lamp voltage
over its service life to increase more than in free-burning operation. This increase depends on the level of
surface soiling in the lamp and can therefore not be
put in figures. To avoid this effect, it is recommended
to use temperature- and UV-resistant materials in the
luminaire.
32
Fig. 35: Clamping the thermocouple on the outer bulb
using a spring element
Please adhere to the following to properly measure
with a thermocouple:
• Good contact with the surface being measured
• Low heat dissipation from the connection point
and therefore
– thermocouple wire with a small diameter
– thermocouple wire parallel to the surface being
measured
• When measuring the outer bulb temperature,
corresponding evaluation must consider the
radiation (cooling-down curve)
7.1.3 Measuring points for thermocouples in
different lamp types
The radiation of the arc tube heats up the thermocouple on the outer bulb above the temperature of the
quartz glass on which it is positioned. Because of the
low thermal capacity, the thermocouple cools down
quickly to the temperature of the quartz glass, after
switching off the lamp, and then cools down slowly
with the quartz, as shown in Fig. 36 and Fig. 37. The
flat part of the cooling curve can be extrapolated back
to the switch-off point to ascertain the temperature of
the outer bulb during operation.
The measurement must be made under worst case
conditions, i.e. for the pinch temperature: base up
burning position, for the outer bulb temperature:
horizontal burning position (if permitted).
Compliance with the outer bulb temperature is not
always sufficient to design a “good” luminaire.
500
T m1 temperature read when switching off, includes additional heating up of
the thermocouple by radiation
measured temperature in °C
T m2 extrapolated temperature of the quartz glass when switching off
The specific limit values are stated in the Technical
Information.
The measured values are to be ascertained under the
worst case conditions.
Both supply voltage and choke impedance influence
lamp wattage. A lower choke impedance and higher
supply voltage each cause an increase in lamp wattage. As the limit temperatures also increase with increasing lamp wattage, the worst case is to be ascertained at the maximum possible lamp wattage. To cover all possible parameters such as choke impedance,
spread of lamp wattage and supply voltage tolerance.
Lamp wattage should be set to approximately 20%
above the normal wattage for measurement purposes.
The operating temperatures of a lamp differ according
to the burning position. The worst case operation is
the base up burning position for pinch or base edge
temperature measurements and the horizontal burning
position for the outer bulb temperature measurement,
insofar as this is possible with the permitted burning
positions.
400
Temperature measurement lamps prepared with thermocouples are available from OSRAM on request for
a fee.
300
200
100
0
0
50
100
200
150
cooling time in s
250
300
350
Fig. 36: Cooling curve for HCI-T 150 W/830 PB in
closed luminaire, lamp wattage 180 W
°C 530
520
510
500
490
480
470
460
450
440
430
420
410
400
390
380
0
Tm1 temperature read when switching off, includes additional heating up of
the thermocouple by radiation
T m2 extrapolated temperature of the quartz glass when switching off
5
10
15
20
25
30
35
sec
Fig. 37: Enlarged detail of figure 36
33
7.1.3.1 HCI®-TC G8.5
7.1.3.4 HCI®-E and E/P / HQI® E27 and E40
Pinch temperature
(in base up burning position)
•
Base edge temperature
(in base up burning position)
•
Outer bulb temperature
(in horizontal burning position)
Outer bulb temperature
(in horizontal burning position)
•
•
7.1.3.2 HCI®-T / HQI®-T G12
(similar for the HCI®-TM and HQI®-TM G22)
Pinch temperature
(in base up burning position)
7.1.3.5 HCI ®-T and TT / HQI®-T E27 and E40
Base edge temperature
(in base up burning position)
•
•
Outer bulb temperature
(in horizontal burning position)
Outer bulb temperature
(in horizontal burning position)
•
7.1.3.3 HCI®-TF, GU 6.5
•
7.1.3.6 HCI®-TX/P
Pinch temperature
(in base up burning position)
•
Base edge temperature
(in base up burning position)
•
Outer bulb temperature
(in horizontal burning position)
•
Outer bulb temperature
(in horizontal burning position)
•
34
7.1.3.7 HCI®-TS, RX7s, RX7s-24 and HQI®-TS Fc2
Outer bulb temperature
(in horizontal burning position)
7.1.3.10 HQI®-T, ≥ 1000 W
Base edge temperature
(in base up burning position)
•
•
••
Outer bulb temperature
(in base up burning position)
Pinch temperature
(in horizontal burning position)
•
Outer bulb temperature
(in horizontal burning position)
•
•
Base edge temperature
(in horizontal burning position)
7.1.3.8 HCI®-PAR E27
Base edge temperature
(in base up burning position)
7.1.3.11 HQI®-TS Excellence
•
•
Pinch temperature
(in horizontal burning position)
Reflector neck temperature
(in base up burning position)
•
•
Reflector edge temperature
(in horizontal burning position)
7.1.3.9 HQI®-TS long arc and short arc ≥ 1000 W
•
Pinch temperature
(in horizontal burning position)
•
Outer bulb temperature
(in horizontal burning position)
Measurement with radiation pyrometry
The temperature limits for lamps where only one burning position is permitted are determined in the allowed
burning position.
35
7.2 Influence of ambient temperature on ballasts
and luminaires
As the ambient temperature increases, the temperature of the luminaire components also increases at
the same rate. The lamp reacts to a higher ambient
temperature with an increase in lamp voltage and
lamp wattage. This can accelerate corrosion and aging
processes. An increased re-ignition peak can result in
failure by the lamp going off at an earlier point in the
service life.
Higher temperatures at choke and ignition unit mean
a reduced service life for these parts or also earlier
failures. The limit temperatures for chokes is generally
130 °C, and 70 °C to 105 °C for ignition units (consider
manufacturer’s instructions). A higher ambient temperature means that luminaires are increasingly switched
off, triggered by the thermal protection.
It can therefore be assumed that the high ambient
temperature has a distinctly negative influence on the
service life of lamps and luminaires.
Luminaire design has a great influence on the temperature of the parts. Heat-generating parts such as
chokes and filter coils can be mounted on materials
with good heat conducting properties and adequate
ventilation openings to ensure there is ample heat
dissipation. The greatest possible spacing should be
kept between heat-sensitive parts such as ignition unit
and capacitors, and heat-generating parts. If necessary, forced cooling should be provided by ventilation
elements.
At the end of the lamp service life, far higher temperatures than normal can occur in the pinch area caused
by outer bulb discharges (see also chapter 6.2 “Failure
mechanisms of metal halide lamps”). The socket in the
immediate vicinity of this point must be rated for that
temperature.
7.3 Lamp holder
Metal halide and high-pressure sodium lamps have
many different bases. These include for example RX7s,
Fc2, G8.5, GX10, GX8.5, GU6.5, G12, G22, GY22, E27,
E40 and K12s, depending on whether the lamps are
single- or double-ended. All sockets must be rated for
the typical conditions for discharge lamps, i.e. high
ignition voltage and high temperatures. It is up to the
user to make an appropriate selection and to ensure
that the lamp holders are installed correctly according
to the corresponding regulations (e.g. IEC 60598 / VDE
0711, IEC 60335 / VDE 0700). Sockets consist of several
parts, each with their own function limits. Exceeding
these limits causes premature failure of the sockets.
36
Temperature and ignition voltage are critical, because the effects of exceeding the limit values are
often only detected after a longer period of time.
This results in a sudden, not gradual, decrease in
service life.
Please note that lamps vary in wattage (permitted
up to +12%) and temperature, and that according
to IEC 60926, ignition units may generate output
voltages of up to 30% above the nominal value.
• Ignition voltages:
The socket must be rated for the corresponding
ignition voltage.
When mounting the socket and the supply leads
in the luminaire, it is important to consider the required creepage distance and clearance, as well
as distances in the insulation. The luminaire IEC
60598-1 standard corresponding to EN 60598-1
defines the safety requirements for ignition voltage regarding creepage and clearance distances.
Particularly when using high-pressure discharge
lamps with Edison bases E27 and E40, care is
required to ensure that the sockets are approved
for discharge lamps. Suitable sockets are marked
with the value to max. “5 kV” and comply with
the increased creepage and clearance distances
required in the socket standards IEC 60238 or
EN 60238 (VDE 0616 Part 1). In the same way,
the other base systems are subject to the socket
standards for special sockets IEC 60838-1 or
EN 60838-1 (VDE 0616 Part 5).
CAUTION : Do not use sockets for incandescent
lamps, e.g. E27 or R7s. Sockets for discharge
lamps must be used to handle corresponding
ignition voltage.
• Temperature code Txxx (continuous use
temperature)
This is the highest temperature for which the
socket was designed. The temperature of special
holders is measured at the socket contact according to 60838-1 (all holders for high-pressure
discharge lamps except Edison sockets). If the
heat resistance of insulation, terminals and leads
deviate from this temperature limit, then separate limit values are stated. In the case of Edison
sockets (according to IEC 60238 ), the rated
temperature is valid for every point in and on the
socket.
At the end of the lamp service life, higher temperatures than normal can occur in the pinch area
caused by outer bulb discharges. The socket must
be rated accordingly (see also chapter 6.2.1 Leaking arc tube).
When replacing such lamps, the socket must
always be checked for signs of damage and replaced if necessary, because a damaged socket
would also damage the new lamp.
• Rated current and rated voltage
The socket must be chosen according to the
lamp parameters. The rated current in this case
is the highest continuous load current, and the
rated voltage is the highest voltage for which the
socket is designed.
CAUTION! Certain sockets such as G12 and E27
are used for different wattage levels. When inserting or exchanging lamps make sure that the right
lamp for the respective ballast is chosen. Otherwise the lamp is operated incorrectly, and the
socket may possibly not be rated for the deviating
operating conditions.
• Fastening parts
The connection parts, e.g. blade terminals, must
be chosen according to the requirements (e.g.
temperature, current load, corrosion resistance).
• Connection leads
The connection leads must be rated accordingly
for the conditions of use with regard to heat and
UV-resistance, mechanical strength, electric
strength and current carrying capacity.
PTFE leads are normally not suitable to handle
ignition voltage. In practice, silicone-insulated
leads with 3.6 mm outer diameter have proven
effective for discharge lamps. For lamps with immediate hot re-ignition, silicone insulation 7 mm
thick should be used together with fiberglass
inlay.
While the lamp is starting up, it is possible for the
start-up currents to briefly exceed the nominal
values, which must be taken into consideration
when rating the socket. Up to 1.5 to 2x the operating current can flow during the start-up phase
(within the first 5 minutes of operation).
• Isolation of Contact Connections
Care must be taken to electrically isolate the
connection contacts in the installation procedure
if this is not safeguarded by the socket alone.
• Lamp pins
Only use lamps with clean metallic contacts.
Oxidized contacts result in high transition resistances and generate high operating temperatures.
The surface of the lamp pins must be smooth and
must not show any visible traces of mechanical
machining in the area of contact with the socket
contact, as otherwise the socket contacts can be
damaged.
7.4 Leads to luminaires
The lead cables to the luminaires must be rated for
their conditions of use, taking account of adequate
heat and UV-resistance, mechanical strength, electric
strength and current carrying capacity, as well as giving due consideration to the effect of cable lengths
(e.g. when remote mounting is required). Cable resistance grows linear to cable length. The resulting voltage drop across the cable reduces the effective supply
voltage. The effects are described in chapter 3.1.3
“Influence of deviations in supply voltage”.
Various factors must be considered when choosing
leads in the lamp circuit:
• The voltage drop across the lead depends on
the flowing current and can be reduced by using
cable with a larger cross section.
• It should also be borne in mind that cable resistance increases with higher ambient temperature.
The resistance of a copper cable rises by about
10% for an increase in temperature of 25 °C.
• Consideration must be given to the voltage drop
in the outgoing and incoming cable.
• 230 V systems are more sensitive to additional
line resistance than 400 V systems.
In applications demanding the lowest possible colour
scattering, the supply conditions should be approximately the same, i.e. supply voltage or line resistance
should be equivalent.
7.5 Maintenance of lighting systems with metal
halide lamps
Since March 2003, the EN 12464-1 standard applies
to interior lighting systems throughout Europe. If a
lighting system is being planned according to this
standard, it is necessary to draw up a maintenance
plan. This has to take into account influences causing
a drop in luminous flux in the system during the course
of the service life, such as dirt depreciation of luminaires and the room itself, together with the aging of
the lamps and lamp failures. The maintenance factor
replaces the previous planning value.
Maintenance factor MF = LLMF x LSF x LMF x RMF
LLMF = lamp luminous flux maintenance factor
LSF = lamp survival factor
LMF = luminaire maintenance factor
RMF = room maintenance factor
37
Example for a maintenance plan
Maintenance plan
Only regular maintenance can ensure compliance with the stipulated illuminance levels in the EN 12464
standard for the lighting system. The following maintenance intervals must therefore be followed.
Room
Type of surrounding:
Normal
Maintenance interval:
every 2 Years
Luminaire XXX
Influence of reflections from the room surfaces:
medium (Room index 1.1 < k < 3.75)
Luminaire characteristics:
direct
Reflector type:
B – Reflector open at the top
Lamp type :
Metal Halide lamp (CIE)
Ballast:
CCG
Operating hours per year:
3000
Maintenance interval (Luminaire):
every 2 Years
Maintenance interval (Lamp):
every 3.5 Years
Failed lamps are replaced immediately:
Yes
Maintenance factor:
0.61
Maintenance instructions:
Lamps must be replaced with suitable replacement lamps of the same characteristics (luminous flux, light
color, color rendering). Any existing starters should also be replaced during relamping.
The room and surfaces deflecting light are to be maintained so as to preserve the original reflection
characteristics.
Always comply with the manufacturer’s cleaning instructions.
Measures to comply with a required minimum lighting level include regular cleaning of the room and the
luminaire, as well as timely replacement of the lamps.
Timely replacement of the lamps also ensures avoidance, for the most part, of undesirable effects at the
end of the lamp service life.
When compiling the maintenance plan, consideration
must be given to the decrease in luminous flux over
the course of the lamp service life in the form of the
lamp luminous flux maintenance factor (LLMF).
The CIE has stated a general luminous flux curve for
metal halide lamps which results in a maintenance
factor of 0.68, for example, for 9000 h.
Given the lesser drop in luminous flux in POWERBALL
HCI® over the service life compared to standard metal
halide lamps with cylindrical tube, the maintenance
factor for 9000 h is 0.8.
38
In practice, this results in the following application
cases (see also table 2):
To achieve a constant luminous flux of min. 500 lx
throughout the service life, taking cleaning and change
intervals into account, the following applies initially:
• Standard lamp and lamp change after 3 years,
i.e. for an LLMF of 0.68 (case 1):
962 lx
approx. 17% higher illuminance,
i.e. more luminaires
• POWERBALL HCI® and lamp change after
3 years, i.e. for an LLMF of 0.8 (case 2):
820 lx
• Standard lamp and lamp change for an LLMF of
0.8, here after 14 months (case 3):
820 lx lamp change required after 14 months,
as 80% of the initial lumens achieved
after 3500 h.
Table 2: Comparison of change intervals for different lamp types
Lamp
HCI-T 70
W/830 PB
Case 1
Case 2
Case 3
Maintenance intervall
3 Years
3 Years
1 Year 2 Months
Operating hours /year
3000
3000
3000
yes
yes
yes
LLMF
according CIE
POWERBALL HCI®
according CIE
RMF
0,95
0,95
0,95
LWF
0,8
0,8
0,8
LSF
1
1
1
LLMF
0,68
0,8
0,8
MF
0,52
0,61
0,61
Immediate exchange of defect lamp
Diagrams showing luminous flux behaviour and survival rate can be found in the Technical Data of the lamps
in the online catalogue.
7.6 Standards and directives for discharge lamps
7.6.1 Standards
The international body for issuing standards relating to
electrical engineering is the IEC (International Electro
Technical Commission). European standards (EN) are
usually identical with the IEC standards. In addition
to the contents adopted by the IEC, the EN standards
also include the requirement to withdraw contradicting national standards within an appropriate period of
time. Furthermore, safety standards are listed in the
low voltage directive, which is mandatory for the CE
symbol and for test marks.
The standards for lamps and accessories are broken
down into safety and performance standards. While the
safety standards stipulate tests regarding electrical,
optical and thermal hazards, the performance standards
look at aspects such as dimensions, electrical description, luminous flux, service life and stipulation of test
procedures.
OSRAM products are constructed according to the
relevant standards and in compliance with the valid
directives.
The following table provides an overview of the key
standards for operating high-pressure discharge
lamps. It features the IEC standards; the corresponding EN standards bear the same number.
39
Table 3: IEC standards for discharge lamps and accessories
Lamp
Safety
62035
Performance
Discharge lamps (excluding
fluorescent lamps) –
Safety specifications
60188
High-pressure mercury vapour lamps –
Performance specifications
60192
Low-pressure sodium vapour lamps –
Performance specifications
60662
High-pressure sodium vapour lamps
61167
Metal halide lamps
61549
Miscellaneous lamps
Bases, sockets and gauges
60061-1
Lamp caps and holders together with gauges for the control of interchangeability and safety;
Part 1: General requirements and tests
60061-2
Lamp caps and holders together with gauges for the control of interchangeability and safety;
Part 2: General requirements and tests
60061-3
Lamp caps and holders together with gauges for the control of interchangeability and safety;
Part 3: General requirements and tests
60061-4
Lamp caps and holders together with gauges for the control of interchangeability and safety;
Part 4: General requirements and tests
60238
Edison screw lamp holders
60399
Barrel thread for lamp holders with shade holder ring
60838
Miscellaneous lamp holders
Accessories
Safety
Performance
60155
Glow-starters for fluorescent lamps
60155
Glow-starters for fluorescent lamps
61048
Auxiliaries for lamps – Capacitors for
use in tubular fluorescent and other
discharge lamp circuits – General and
safety requirements
61049
Capacitors for use in tubular fluorescent and other discharge lamp
circuits – Performance requirements
61347-1
Lamp control gear – Part 1:
General and safety requirements
–
61347-2-1
Lamp control gear – Part 2-1:
Particular requirements for starting
devices (other than glow starters)
60927
61347-2-4
Lamp control gear – Part 2-4:
Particular requirements for d.c.
supplied electronic ballasts for
general lighting
60925
61347-2-5
40
Lamp control gear – Part 2-5:
Particular requirements for d.c.
supplied electronic ballasts for public
transport lighting
Auxiliaries for lamps – Starting
devices other than glow starters) –
performance requirements
DC supplied electronic ballasts for
tubular fluorescent lamps –
Performance requirements
Accessories
Performance
Safety
61347-2-6
Lamp control gear – Part 2-6:
Particular requirements for d.c.
supplied electronic ballasts for
aircraft lighting
61347-2-9
Lamp control gear – Part 2-9:
Particular requirements for ballasts
for discharge lamps (excluding
fluorescent lamps)
61347-2-12
Lamp control gear – Part 2-12:
Particular requirements for d.c. and
a.c. supplied electronic ballasts for
discharge lamps (excluding fluorescent lamps)
60923
Auxiliaries for lamps – Ballasts for
discharge lamps (excluding tubular
fluorescent lamps ) – Performance
requirements
Luminaires
60598-1
Luminaires – Part 1: General requirements and tests
EMC
IEC/CISPR15
Limits and methods of measurement of radio disturbance characteristics of electrical lighting
and similar equipment
IEC 61547
Equipment for general lighting purposes – EMC immunity requirements
IEC 61000-3-2
Electromagnetic compatibility (EMC) – Part 3-2: Limits – Limits for harmonic current
emissions (equipment input current ≤ 16 A per phase)
Not all lamps are covered by data sheet in the lamp
standards, but the application range of every standard
applies to all lamps of the corresponding type.
OSRAM lighting products marked with “CE” fulfill the
safety and EMC standards where applicable for the
specific product (see table 3).
Vibration and impact tests are covered by IEC
60068-2-6 Fc and IEC 60068-2-29 Eb.
7.6.3 Certificates
7.6.2 Directives
“CE” stands for “Communauté Européenne” (European
Community) and indicates compliance of a product
with the corresponding European Directives. The
CE mark addresses authorities and is applied by the
manufacturer. CE marking was created primarily to
warrant safe products for the end consumer in the free
movement of goods within the European Economic
Area (EEA) and its European Community (EC). The CE
mark is frequently called a “passport” for the European
Single Market. The key directives for lighting products
with corresponding compliance confirmed by application of the CE mark are the electromagnetic compatibility directive (EMC, 89/336/EEC) and the electrical
equipment directive (73/23/EEC), also called the “low
voltage directive”. The low voltage directive demands
that the product does not cause any harm to persons,
animals and things. Compliance with the low voltage
directive can also be verified by compliance with the
safety standards.
On the initiative of European manufacturer’s associations, European test and certification bodies have
agreed to provide a uniform European evaluation of
electrical products in order to indicate to the buyer
y
of a product that it is safe and complies with stateof-the-art technology. This resulted in the ENEC
agreement and the ENEC mark (ENEC = European
Norms Electrical Certification). Prerequisite for being granted an ENEC certificate is compliance of the
product with the corresponding European safety and
performance standards. The production procedure
must have a quality management system (e.g. based
on DIN EN ISO 9002). The corresponding certification
body carries out regular audits to monitor whether the
system requirements are being met. The number next
to the ENEC symbol identifies the certification body.
All current certification bodies that have signed the
ENEC agreement can be found together with the
corresponding countries and the register of issued
ENEC approvals on the ENEC internet website
www.enec.com.
41
If an ENEC mark is issued for a product by a certification body, then the European certifying bodies participating in the ENEC agreement treat this product as
if they had tested and certified it themselves. Further
testing and certification by one of these bodies is no
longer necessary.
The ENEC mark can be obtained for luminaires for
which a European standard exists. Luminaire accessories such as ballasts, ignition units, lamp sockets and
capacitors can also be issued an ENEC mark if they
satisfy the corresponding EN standard.
7.9 Optical design of reflectors
7.9.1 Condensate in the lamp
While the mercury in metal halide lamps evaporates
completely when operated at full power, the metal halides are in saturated state. There is therefore always a
surplus of condensed metal halides at the “cold spot”
in the arc tube.
Burning
position
45° angle
7.7 Radio interference
Selected luminaires must comply with the international
requirements such as CISPR 15 and CISPR-22 A or B,
and in practice, radio interference is low enough so
that no negative effects are expected on the environment.
Even though ignition pulses from an ignition unit without cut-out can cause radio interference if with a defective lamp, there are no regulations
g
for this case. The
interferences can be extensive. One solution is swift
replacement of the defective lamp or use of ignition
units with a cut-out feature. These detect the defect or
the absence of the lamp and switch off the ignition unit
after a limited time period of futile ignition attempts.
The unit has to be disconnected from the grid power
supply to reset the timer.
Condensate
Horizontal
burning
position
Figure 38: Example for condensate precipitation
in the lamp
The balance between condensed and evaporated part
of metal halides depends on the temperature of the
arc tube wall. The coldest spot of the tube where the
metal halides have condensed is usually at the bottom
of the tube.
7.8 RoHS conformity
7.9.2 Projection of the condensate
All products brought onto the market in Member States
of the European Union by OSRAM since 1 July 2006
comply with the requirements of the EC directive
2002/95/EC “on the restriction of the use of certain
hazardous substances in electrical and electronic
equipment” (RoHS).
As a fundamental rule, our products contain no cadmium, hexavalent chromium, polybrominated biphenyls
(PBB), polybrominated diphenylether (PBDE) or lead,
and fulfill the requirements of the directive for the use
of mercury.
42
The light radiated from the plasma projects the condensate of the lamp so that the reflector needs to mix
the emitted light to ensure homogenous radiation. In
particular when the burner is horizontal, the radiation
components from the upper half of the burner and the
lower half of the burner have to overlap and be mixed
during projection. If the reflector cannot ensure this,
the condensate of the lamp is projected and appears
on a white wall as a yellow mark.
on the outer bulb and the socket or the pinch area first
need to be measured as stated in the catalogues, see
also section 7.1.
However it should be noted that even if the temperatures measured on the outside of the lamp lie within
the defined tolerance values, this does not necessarily
mean that there is no overheating inside the lamp.
Those surfaces closely surrounding the lamp, such as
the reflector neck, diffuser tube and glare shield caps,
reflect back on the lamp. Likewise, elliptical reflectors
radiate back onto the lamp if the burner is not positioned correctly in the burner point of the reflector.
In these cases the lamps may also suffer damage even
if the temperatures measured on the outside of the
lamp lie within the defined tolerance values.
The following recommendations are made:
Figure 39: Projection of the condensate by the reflector
7.9.3. Back reflection on the lamp
The luminaire design needs to ensure that no radiation is reflected back onto the lamp as this can cause
thermal loads on sensitive parts of the lamp which
normally leads to unusually early failures. For a lamp
with a quartz arc tube, this may lead to the expansion
of the tube or to a leak in the pinch area.
In lamps with ceramic arc tubes, the so-called sealing
area at the ends of the capillaries is particularly sensitive: here overheating can lead to increased chemical
reactions and then to cracks and lamp failures. A further sensitive component is the lamp’s getter.
To avoid glare, the glare shield caps often used for
halogen bulbs or the glare shield rings on lamps
with double-sided burners such as HCI® may not be
used. Instead, the glare protection for these lamps is
effected by using e.g. honeycomb filters or “anti-glare
baffles” or “hoods” attached to the outside of the
luminaire.
The glare can also be reduced by using protected
lamps, such as HCI®-TX/P, because the fact that there
is no front glass at the luminaire means that there is no
reflex glare on the front glass.
To achieve a homogenous colour radiation, facetted
and matt reflectors should be used. Diffuser tubes
around the lamp are not suitable.
In lamps with a reflector casing, the reflector should
be smoothly cut at the opening and should not have a
neck.
Sealing area
Reflector nec k
It is more difficult in cases where the reflector itself
comprises the outer part of the luminaire. If a reflector
neck is used here, e.g. to prevent the emission of light
scatter, this leads, e.g. in ceramic lamps to a higher
temperature load in the socket-based capillaries.
Capillaries
Getter
e
The extent of the damage depends on the following
parameters:
• Extent of the spatial covering of the reflector neck
and the capillaries: less is better
• The diameter of the reflector neck: bigger is better
• Level of reflection of the reflector neck: matt is
better than mirrored
• Overall volume of the reflector: bigger is better
Figure 40: Example of a reflector with a reflector neck
over the thermally critical parts of the lamp.
To establish whether or not the luminaire design could
cause impermissible thermal damage, the temperature
The service lives stated by OSRAM only apply to
lamps operated in luminaires that do not reflect back
on the lamp. They are based on a switching rhythm
11 h ON, 1 h OFF.
43
If back reflecting construction elements are used in the
luminaire design, the guarantee for the lamps can be
restricted or even completely suspended.
It is therefore recommended contacting OSRAM if
there is any doubt during the design stage.
In the case of lamps with back reflecting construction
elements, tests should always be carried out to ascertain whether the extent of the lamp damage can at
least be assessed as minimal.
44
It is useful to do comparative burning tests with non
back reflecting luminaires. If in this case of e.g. ceramic lamps, visible deposits in the outer bulb of the
lamps occur at an early stage in the tested luminaires,
the burner has overheated due to back reflection.
Due to the fact that the failure rate of ceramic lamps
depends on the switching frequency, a lamp test of
this kind can be accelerated by increasing the switching frequency to e.g. 3h ON, 1h OFF.
8 Light and colour
Light is the part of the electromagnetic spectrum
which can be seen with the eye. By definition, the perceptible wavelength range is 380-780 nm, although
radiation can also be perceived as colour in the near
infrared range. Similar to visible light, ultraviolet and
infrared variation belong to the electromagnetic spectrum.
V(λ)
V’(λ)
(Night vision)
L < 0,1 cd/m2
(Day vision)
L > 30 cd/m2
Fig. 43: Spectral brightness sensitivity
V(λ
(λ) for photopic vision and
V'(λ
(λ) for scotopic vision
Light intensity
Light intensity
Light intensity is the measure
of lightoutput in a specified
direction.
Unit: 1 candela (cd)
Fig. 41: Visible light as part of the electromagnetic
spectrum
Fig. 44: Definition of luminous intensity
Different wavelengths can be perceived to different
extents. The maximum of the sensitivity curve for
photopic vision is at 555 nm. The light output (luminous flux) is ascertained by multiplying the physical
radiation output with the eye sensitivity curve V(λ)
(see Fig. 43).
If the entire radiation output is emitted monochromatically in the wavelength of maximum eye sensitivity
(555 nm), then the theoretical maximum luminous efficacy is 683 lm/W. With uniform distribution of radiation
over the range of 380 – 780 nm, approx. 196 lm/W is
possible.
Illumination
Average illumination of a
surface is luminous flux
per unit area.
Lux
Fig. 45: Definition of illuminance
Luminance
Luminous flux
Luminous flux is the light output
of a light source.
Unit: 1 lumen (lm)
Fig. 42: Definition of luminous flux
visible surface
Light intensity
illuminated
surface
Luminance is the measure of
the brightness that the eye has
of a surface.
Unit: 1 candela/m2 (cd/m2)
Luminance depends on the
surface of the area seen by the
eye and on the luminous
intensity, radiated from the
surface towards the eye.
Fig. 46: Definition of luminance
Luminous = Irradiated light in Lumen (lm)
efficacy
spent electrical power in watt
(Gl. 9.1)
45
8.1 Night vision
The luminous flux, measured in lumens, is the irradiated output of a light source evaluated by the eye. It
is defined by multiplying the physical radiation output
with the eye sensitivity curve V(λ). Standard luminous
flux measurements only consider the reaction of the
eye at high illuminance levels (photopic vision) as is
typical for daylight and indoor illumination. Luminous
flux measurements measure photopic light as perceived by the central region of the eye.
When the illumination level is very low, for example at
night by star light, the vision conditions are said to be
scotopic. The reaction of the eye changes under these
circumstances. The eye sensitivity curve for low illumination levels (less than 0.1 cd/m²) is the V'(λ) curve, as
shown in the figure 43.
The change in eye sensitivity comes from the presence
of two types of light receivers on the retina: rods and
cones. The rods are responsible for vision under low
illuminance and are located in the peripheral field of
vision. The rods are sensitive to scotopic light while
the cones react to photopic light. When the illumination
level decreases, the rods are therefore more active,
while the cones become inactive.
The effective, seen “lumen” will differ from the measured photopic luminous flux. When the illumination
level falls, the effective “luminous flux”, e.g. of yellow
high-pressure sodium lamps, decreases while the
effective “luminous flux” of white light with a higher
share of green/blue light increases.
Figure 47 shows the radiation output of a HCI ®-TC
70 W/NDL and a NAV®-T 400 W Super 4Y, normalized
in the interests of comparability to a luminous flux of
1000 lm. The diagram shows the relative distribution
of the radiation in the spectrum.
Weighting factor for eye sensitivity curve
Sensitivity for red and yellow light decreases, while
there is better perception of blue light. When luminous
flux is measured under photopic conditions, this does
not correspond to what the eye perceives at low light
levels. The reaction of the eye does not change suddenly from high to low illumination levels. The change
is gradual when the illumination level decreases to
twilight and typical street lighting conditions. This is
called mesopic vision which lies between photopic and
scotopic vision.
Fig. 47: Physical radiation output in W per 1000 lm and per 5 nm
In Fig. 48, the physical radiation output has been multiplied by the V(λ) curve to ascertain the luminous flux
per 5 nm in each case. Integration of the values for all
wavelengths between 380 nm and 780 nm results in
the specified 1000 lm for both light sources.
46
The NAV® lamp radiates more light in the range around
580 nm, which is near the maximum of the V(λ) curve.
This contributes to a high luminous efficacy. On the
other hand, there are some gaps in the spectrum,
particularly in the blue part of the spectrum, which is
responsible for the poorer colour rendering compared
to the metal halide lamp.
Weighting factor for eye sensitivity curve
Fig. 48: Relative luminous flux in lumen per 1000 lm and per 5 nm
Illumination levels in street lighting are higher than
0.1 cd/m², resulting in a sensitivity between photopic
and scotopic vision.
Weighting factor for eye sensitivity curve
In Fig. 49, the radiation output has been multiplied by
the V'(λ) curve for illumination levels below 0.1 cd/m².
The diagram shows that the perceived illumination
level of the metal halide lamp is far higher (in this example about three times higher) than the high-pressure
sodium vapor lamp.
(λ)
Fig. 49: Equivalent to luminous flux, taking account of an eye sensitivity curve at a low illumination level V'(λ
47
8.2 Colour rendering
One way of showing the colour impression is the standard chart as per DIN 5033 – basic stimulus.
Colour is a sensory impression conveyed by the eye.
The evaluation of a colour stimulus by the eye causes
a uniform effect (colour stimulus specification). This
can be described by colourimetric numbers (e.g. x,
y and z in the CIE 1931 or CIE 1976 colour space or
L, a and b in the CIE 1976 (L*a*b*) space or W, U and
V in the CIE 1964 colour space (W*, U*, V*)). But the
perceived colour (the subjective impression) depends
on the general conditions (colour mood, surrounding
surfaces, luminance).
The primary colours, i.e. saturated monochromatic colours, run around the periphery of the colour triangle.
An ideal black body (or Planck radiator) radiates an
electromagnetic spectrum depending on its temperature. The colour thus depending on temperature is
depicted in the Planck curve, this is the so-called
“colour temperature”.
Colours on the Planck curve are marked with the
corresponding colour temperature; chromaticity coordinates deviating only slightly from the Planck curve
(within the range of the Judd straight lines, corresponding to a distance of approx. 5.4 threshold
value units) are marked with the correlated colour
temperature.
Fig. 50: Standard colour chart as per DIN 5033
Reference color temperature 4000K
HQL-Standard
HCI - NDL
R1: 46
R1: 99
R2: 61
R2: 97
R3: 54
R3: 88
R4: 46
R4: 96
CRI: 50
R5: 43
R5: 97
R6: 36
R6: 95
R7: 66
R7: 96
R8: 44
R8: 93
Fig. 51: Defining the colour rendering indices in comparison for two light sources
48
CRI: 95
Larger deviations are associated with a clear tint. The
distance to Planck is also known as the chromaticity
gap Δc.
Colour rendering is specified by irradiating defined test
colours in succession with a reference source (an ideal
Planck radiator with the temperature and therefore colour temperature of the test light source) and with the
test light source. The specific resultant colour shift ΔEi
is defined for every test colour i in the uniform colour
space CIE 1964 (W*, U*, V*).
The specific colour rendering index R i is defined as
follows
Ri = 100 – 4.6 Δei
Every special colour rendering index can therefore
reach a maximum value of 100 when the test colour
appears identical under reference and test light
source. Negative values are also possible with greater
deviations (and hence larger ΔE i ).
Table 4: Test colours from DIN 6169
Testcolours
R1 – Dusky pink
R2 – Mustard yellow
R3 – Yellow green
R4 – Light green
R5 – Turquoise blue
R6 – Sky blue
R8 – Syringa violet
R7 – Aster violet
Saturated colours and additional test colours
R9 – Red
R10 – Yellow
R11 – Green
R12 – Blue
R13 – Skin colour
R14 – Leaf green
Apart from the first 8 colour rendering indexes,
DIN 6169 also defines other test colours, which are
four saturated colours and additional test colours. The
further test colours permit a more precise description
of the colour rendering properties of the light source.
In principle it is possible to define any random number
of many different test colours.
Table 5: Colour rendering levels
Evaluation Colour rendering level Colour rendering index CRI
8.2.1 Test colours from standard DIN 6169
The arithmetic mean from the first 8 test colours
(see Table 4) shows the general colour rendering index
CRI or Ra.
The general colour rendering index results in the
colour rendering levels for light sources given in
Table 5.
Very good
1A
≥ 90
Very good
1B
80 – 89
Good
2A
70 – 79
Good
2B
60 – 69
Suboptimal
3
40 – 59
Suboptimal
4
20 – 39
Thanks to a higher possible wall load, the colour rendering properties when using POWERBALL® technology
have been visibly improved compared to the lamp with
cylindrical ceramic arc tube. A further improvement
specific Color rendering index
HQI-T POWERSTAR
Typical lamp with
cylindrical ceramic
Fig. 52: Comparison of the specific colour rendering indices for various metal halide lamps
49
has become possible due to additional adaptation of
the HCI® Shoplight, which achieves the best colour rendering properties of all metal halide lamps. Figure 49
shows the values of the colour rendering indices 1 to
14 for four different lamp types with the correlated colour temperature of 3000 K. The advantages can best
be seen for colour rendering index R9 for saturated
red, but the superiority of POWERBALL ® technology is
also apparent for the other colour rendering indices.
8.3 Light and quality of life
It has been a known fact for many years that as well as
its known visual effects, light also has other biological
effects on the human body. The most acknowledged
effect is the way light influences the day-and-night
cycle. This influence is also perceived by the eyes, not
however via the vision center in the brain but via other
nerve cells that affect the pineal gland and hence the
forming of the sleep hormone melatonin. Bright light
in the night suppresses the formation of melatonin,
Pineal gland
reducing the level of melatonin in the bloodstream.
This is called melatonin suppression. (see Fig. 53).
Scientific studies on how light forms or suppresses the
sleeping hormone melatonin have shown that together
with the visual path which is responsible for vision,
there is also a non-visual path which, independent of
the visual system, controls melatonin production and
therefore the circadian rhythm (daylight rhythm).
While the visual path leads directly from the eye to the
vision center of the brain via the optic nerve, the nonvisual path is coupled via the suprachiasmatic nucleus
(SCN) to the pineal gland and controls melatonin production. This process is relatively slow, in time constants of several minutes, while the process of vision
takes place within a few 10 ms.
The SCN is a collection of several thousand nerve
cells, located above the intersection of the optic
nerves (chiasma). This is deemed today to be the main
regulator of the inner clock (master clock).
Suprachiasmatic
Nucleus (SCN)
Visual center
Light
pineal gland
release of
melatonin
Ganglion cell
in upper cervical
vertebra
Retinohypothalamic
tract (RHT)
Spinal cord
Fig. 53: How light affects on the human brain
50
circadian function
visual sensitivity
spectral sensitivity
And so we distinguish between the visual path, responsible for all visual tasks such as recognizing pictures, perceiving brightness, contrast, shapes, etc.,
and the non-visual path, or also “biological path”,
which controls in particular the circadian rhythms and
also influences in the daytime our alertness and mental performance and also biological functions such
as hormone production, the blood circulation and the
metabolism.
The non-visual path is essentially independent of the
visual path.
wavelength [nm]
Light
Fig. 54: Definition of a circadian function C(λ
(λ) by Gall
et. al. in analogy to the photometric function V(λ
(λ) as
standardized by the CIE.
Non-visual path
Cirkadian rhythm
• Alertness
• Tiredness
• Hormone production
• Vitality
• Recovery
• Blood circulation
• Metabolism
•
Visual path
Recognizing pictures
• Brightness, illuminance
• Contrast
• Shapes
• Movement
• Perception
• Information
•
Scientific studies (Prof. Brainard, Thomas Jefferson
University, Philadelphia) have established that melatonin suppression depends not only on brightness but
also on the wavelength of the light used. Light in the
blue spectral range of about 460 nm has the strongest
effect on suppressing melatonin.
The course of the sensitivity curve measured by Brainard for melatonin suppression shows no correlation to
the course of the previously mentioned eye sensitivity
curves for the red, green or blue photoreceptors in the
eye.
This made it apparent that there is a further previously
unknown type of light-sensitive cell in the eye responsible for the circadian effect of light.
Prof. Gall from the “Lichttechnisches Institut” at the
University of Ilmenau has recognized that the sensitivity curve for melatonin suppression published by
Brainard is very similar to the known curve V(λ) which
describes eye sensitivity for seeing brightness. Only
the spectral position is shifted towards blue.
The curve C(λ), suggested by Gall, serves today as the
foundation for a measuring system for circadian lighting data, defined in DIN V 5031-100. These weighting
factors can be taken into account in order to consider
the biological effect of light sources.
8.4 UV radiation
The metal halide lamp standard IEC 61167 describes
the effective UV-radiation output and specifies limit
values in the respective lamp data sheets. This
means that the UV radiation of the lamp in the range
250 – 400 nm is weighted with a so-called evaluation
function (see figure 55) (similar V(λ)-evaluation of
visible radiation).
This evaluation curve shows the generalized sensitivity
of human tissue to UV radiation over wavelength and
has been defined by the ICNIRP (International Commission on Non-Ionizing Radiation Protection).
This evaluation curve is used today by nearly all national and international bodies (standardization, professional associations, etc.). The ACGIH (American
Conference of Governmental Industrial Hygienists)
uses this evaluation for workplace guidelines.
The NIOSH (National Institute for Occupational Health
and Safety) is an American federal authority that researches occupational health and safety and issues
corresponding recommendations.
The maximum daily dose (8h working day) permitted
according to the ICNIRP is 30 J/m². With a mean illuminance of 500 lx, this dose is achieved with an effective UV radiation of approx. 2 mW/klm. The IEC 61167
data sheets for metal halide lamps indicate the maximum values of the generated effective UV radiation.
IEC 62035 states the limit values for UV radiation
(2 mW/klm resp. 6 mW/klm) for high-intensity discharge
lamps as an indication for the luminaire manufacturer.
51
OSRAM metal halide lamps comply with the limit
values of 2 mW/klm or even go below the limit
considerably.
Exceptions are the HQI® lamps without outer bulb
with the power of 1000 W and 2000 W. Here special
safety precautions have to be met by the luminaire.
Standardization of UV variables per “klm” or “lm” offers the advantage of being able to make direct comparison of the relative radiation shares of various lamp
types and wattage classes with regard to the same
application illuminances.
As a comparison:
• Tubular Fluorescent T8 & T5 have an ACGIH UV
value of approx. 0.2 mW/klm (with possible minor
fluctuations depending on wattage class and light
colour).
• Compact lamps have a lower 0.03 mW/klm
Fading occurs not only due to UV but also with shortwave visible light, depending on the spectral object
sensitivity (effect function) of the irradiated object.
There is plenty of information on this subject in the
Division 6 report of the CIE (CIE technical collection)
entitled “On the Deterioration of Exhibited Objects by
Optical Radiation”. Although this deals with objects in
museums, the results are also applicable for example
to shop window lighting. A stronger fading effect could
be achieved by stronger focusing of the light or by a
higher luminous flux in the lamp.
A numerical definition of colour change generated by
irradiation must be expressed in the form of colourimetric differences ΔE*ab. In this way, it is possible to
express exactly every fading, blackening and yellowing
or basically every colour change. Effective radiation
resulting in a colour change of exactly ΔE*ab = 1, is
called threshold effective radiant exposure. This value
is important, as experience shows that colour changes
in this magnitude can be perceived by the average
observer on comparing unexposed areas of a sample
with exposed parts.
Other limit values can be used (ΔE*ab = 2, 3, 4, etc.)
if the correspondingly larger colour differences are
acceptable.
8.4.2 Protective measures to reduce fading
Every protective measure must refer to a reduction in
effective radiant exposure Hdm. Effective radiant
exposure Hdm is the product of the radiation time
tdm and effective irradiance Edm.
Fig. 55: Evaluation function for the sensitivity of human
tissue to UV radiation as per ICNIRP
8.4.1 Fading effect
The colour change in light-sensitive materials resulting
from irradiation with light sources depends on
• irradiance or illuminance,
• spectral distribution of the radiation from the
light source,
• spectral object sensitivity (effect function) and
• irradiation time.
If daylight contributes to the lighting e.g. via skylights
or shop windows, this has to be considered as part of
the irradiation as well. Daylight contains considerable
amounts of light in the UV range und short-wave visible light.
In new objects, colour change is strongest during the
initial period of light exposure. Old wall carpets for example which have been exposed to light for centuries,
show hardly any remaining sensitivity to radiation.
52
Reduction can consist of:
• avoiding the critical wavelengths by using
corresponding filters according to the spectral
sensitivity of the irradiated object
• reducing the irradiance
• reducing the exposure time
• enlarging the distance to the luminaire
Remarks on filtering the critical wavelengths:
Relative spectral sensitivity for most samples is not
only very high in the ultraviolet range of irradiation,
and also still fairly high in the visible range for many
exhibits. This would mean that the short-wave visible
range also has to be filtered. To what extent this is feasible depends on the colour rendering characteristics
and the changed colour temperature of the remaining
visible radiation.
9 Disposal of discharge lamps
High-pressure discharge lamps contain small quantities of mercury as an environment-relevant substance.
Metal halide lamps can also contain thallium iodide as
an additive. This is why discharge lamps must be disposed of separately from domestic waste and industrial waste similar to domestic waste. The last owner
is obliged to dispose of the discharge lamp using the
correct procedure.
Breakage of high-pressure discharge lamps emits
traces of toxic mercury and thallium halides.
More information on handling discharge lamps is
available at
http://www.osram.com/weee
In any case the legal regulations of the respective
country have to be respected.
General information on disposal can be found at
http://www.osram.com/weee
9.2 Collection, transport and disposal of discharge
lamps at end-of-life
During transport to disposal or collection points, please
make sure that the lamps are adequately protected
from breakage, which would result in the emission of
mercury.
Transport of the used discharge lamps by the last
owner is not subject to transport permission. The
lamps do not constitute dangerous cargo in accordance with the corresponding regulations GGVS,
GGVE and ADR and RID.
9.3 Ordinance on Hazardous Substances
9.1 Statutory requirements
Directive 2002/96/EC WEEE (waste of electrical and
electronic equipment) came into effect on 13 February
2003. It applies in all Member States of the European
Union.
Discharge lamps (fluorescent lamps, compact fluorescent lamps, high-pressure mercury vapor lamps, metal
halide lamps, high-pressure and low-pressure sodium
vapor lamps) are not subject to mandatory marking
according to the Ordinance on Hazardous Substances.
Similar systems are also in use in some non-European
countries.
The main aim of this EU directive is the re-use, material recycling and other forms of recycling of such waste
products in order to reduce the quantity of waste and
to protect resources, particularly by means of reuse
and recycling.
All manufacturers and importers of electrical and
electronic equipment are obliged to take back their
products and to ensure that they are treated, reused or
recycled.
OSRAM lamps intended for recycling are
marked with this symbol.
53
10 List of abbreviations
54
ACGIH
American Conference of Governmental Industrial Hygienists
AGLV
Arbeitsgemeinschaft Lampen-Verwertung (Lamp recycling consortium)
ANSI
American National Standards Institute
CE
Communauté Européenne (European Community)
CIE
Commission Internationale de l‘Eclairage (International Lighting Commission)
DALI
Digital Addressable Lighting Interface (communications standard for lighting systems)
CISPR
Comité international spécial des perturbations radioélectriques
(Special International Committee for Electromagnetic Interference)
ELMAPS
European lamp Manufacturers association for the preparation of standards
EMC
Electromagnetic Compatibility
EN
European standards
ENEC
European Norms Electrical Certification
ECG
electronic control gear
ICNIRP
International Commission on Non-Ionizing Radiation Protection
IEC
International Electro technical Commission
CCG
Conventional Control gear (choke)
LIF
Lighting Industry Federation Ltd
LSF
Lamp survival factor (as per standard EN 12464)
LLMF
Lamp luminous flux maintenance factor (as per standard EN 12464)
LMF
Luminaire Maintenance factor (as per standard EN 12464)
NIOSH
National Institute for Occupational Safety and Health
PCA
polycrystalline Alumina
RMF
Room Maintenance factor (as per standard EN 12464)
SCN
suprachiasmatic Nucleus
WEEE
Waste Electrical and Electronic Equipment
MF
Maintenance factor (as per standard EN 12464)
ZVEI
Zentralverband Elektrotechnik- und Elektronikindustrie e.V.
(Central Federation of the Electrical and Electronic Industry)
11 Literature
[1] Kelly, D. H. (1961) Visual Response to Time-Dependent Stimuli. I. Amplitude Sensitivity Measurements.
JOURNAL OF THE OPTICAL SOCIETY OF AMERICA Vol. 51, Nr 4 On Pages: 422-429
Henger, U. (1986) Untersuchungen zur Entwicklung eines Messgerätes zur Bestimmung des Flickerfaktors.
Licht 86 7. Lichttechnische Gemeinschaftstagung.
[2] Afshar, F. 2006. Light Flicker-Factor as a Diagnostic Quantity for the Evaluation of Discharge Instabilities
in HID Lamps. LEUKOS Vol. 3 No 1, July 2006
[3] Sturm: Betriebsgeräte und Schaltungen für elektrische Lampen“, Siemens AG Verlag
”
55
OSRAM AG
Head Office
Hellabrunner Strasse 1
81543 Munich
Germany
Phone +49 (0)89-6213-0
Fax +49 (0)89-6213-20 20
www.osram.com
106T020GB 04/12 pd OSRAM HID Subjeect to change without notice. Errors and omission excepted.
www.osram.com
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement