ASSIST LEDLife Revised2007 PDF

recommends…
LED Life for General Lighting:

Definition of Life
Volume 1, Issue 1
February 2005
A publication of the Alliance for Solid-State Illumination Systems and Technologies

recommends…
Copyright © 2005 by the Alliance for Solid-State Illumination Systems and Technologies
(ASSIST).
Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union

St., Troy, New York, USA.
http://www.lrc.rpi.edu
All rights reserved. No part of this publication may be reproduced in any form, print,
electronic, or otherwise, without the express permission of the Lighting Research
Center.
This ASSIST recommends. . . was prepared by the Lighting Research Center at the
request of the Alliance for Solid-State Illumination Systems and Technologies (ASSIST).
The recommendations set forth here are developed by consensus of ASSIST sponsors
and the Lighting Research Center. ASSIST and the Lighting Research Center may
update these recommendations as new research, technologies, and methods become
available.
ASSIST Recommends Sponsors
Boeing
Cree
Federal Aviation Administration
Lite-On
GE Lumination
New York State Energy Research and Development Authority
Nichia America Corp.
Northwest Energy Efficiency Alliance
OSRAM SYLVANIA
Philips Lighting
Photonics Cluster (UK)/The Lighting Association
Seoul Semiconductor
United States Environmental Protection Agency
Lighting Research Center Technical Staff
In alphabetical order:
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Contents
Purpose........................................................................................................................... 4
Background ..................................................................................................................... 4
Components and Systems .............................................................................................. 4
Life Definition .................................................................................................................. 4
Rationale ......................................................................................................................... 4
About ASSIST ................................................................................................................. 5
3
recommends…
Purpose
This document outlines a proposed definition of useful life for light-emitting diode
(LED) components and systems used for general lighting applications. The
audience for this document is LED component and system manufacturers.
Background
LEDs exhibit very long operational life characteristics, typically 50,000 hours or
longer. Like all light sources, LEDs slowly decrease in light output over time.
Because they rarely fail, situations can occur where LEDs are emitting less light
than intended by the specifier, yet still appear to be operating. LEDs can also
undergo gradual shifts in color that result in an unacceptable appearance.
Because of these characteristics, the ASSIST program has developed a set of

proposed definitions for the useful life of white LED components and white LED
lighting systems.
Components and Systems

For the purpose of this document, an LED component is defined as the individual
LED light source. One or many LED components may be assembled with a driver
and housing to create an LED lighting system. An LED system is defined as the
integration of all necessary components into a working module, such as a light
fixture.
Life Definition
The reported life of an LED component or system is to be defined as the
operating time (L, in hours) for the component or system to reach two
performance criteria:
L70% (hours): time to 70% lumen maintenance
L50% (hours): time to 50% lumen maintenance
Within these times, the LED component or system should not exhibit chromaticity
shifts greater than those bounded by a four-step MacAdam ellipse.
These times are to be measured under specific conditions as outlined in the

accompanying documents, "LED Life for General Lighting — Measurement
Method for LED Components" and “LED Life for General Lighting —
Measurement Method for LED Systems.”
Rationale
For general lighting applications, 70% lumen maintenance, which corresponds to
a 30% reduction from the initial light output of a lighting system, is close to the
threshold for detecting gradual reductions in light output. Research shows also
that reductions to 70% of initial light output are considered acceptable by the
majority of occupants within a space. Thus, this level is unlikely to be problematic
4
recommends…
for a wide array of lighting applications. Indeed, lumen maintenance values

greater than 70% are achieved by most successful general light sources
throughout their operating life.
When the appearance and output of a particular lighting application are critical
(e.g., wall washing in a corridor where the light sources are seen side by side),
useful life based on 80% lumen maintenance should be considered. For other
applications where light output is not critical to the performance of a lighting
system, such as decorative applications, reductions of 50% might be acceptable.
Providing the operating time to reach at least two levels of light output will assist
specifiers and manufacturers in predicting useful life based on other lumen
maintenance criteria.
About ASSIST
ASSIST was established in 2002 by the Lighting Research Center at Rensselaer
Polytechnic Institute to advance the effective use of energy-efficient solid-state
lighting and speed its market acceptance. ASSIST’s goal is to identify and reduce
major technical hurdles and help LED technology gain widespread use in lighting
applications that can benefit from this rapidly advancing light source.
5
recommends…

Measurement Method for LED
Components
Volume 1, Issue 2
February 2005

recommends…
(ASSIST).
Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union St.,
Troy, New York, USA.
All rights reserved. No part of this publication may be reproduced in any form, print, electronic,
or otherwise, without the express permission of the Lighting Research Center.
This ASSIST recommends. . . was prepared by the Lighting Research Center at the request of
the Alliance for Solid-State Illumination Systems and Technologies (ASSIST). The
recommendations set forth here are developed by consensus of ASSIST sponsors and the
Lighting Research Center. ASSIST and the Lighting Research Center may update these
recommendations as new research, technologies, and methods become available.
Boeing
Cree
Lite-On
GE Lumination
OSRAM SYLVANIA
Philips Lighting
Seoul Semiconductor
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Contents
Purpose.......................................................................................................................... 4
Scope ............................................................................................................................. 4
Method ........................................................................................................................... 4
Additional Data .............................................................................................................. 5
About ASSIST ................................................................................................................ 5
3
recommends…
Purpose
This document outlines a method for measuring useful life for light-emitting diode
(LED) components used in general lighting systems. The intent of this document
is to encourage common methods of testing and data presentation. The target
audience is LED component manufacturers.
Scope
An LED component is defined as the individual LED light source. One or many
LED components may be assembled with a driver and housing to create an LED
lighting system.
The measurement method defined here is to be used to determine the

operational life in hours for an LED component to reach L70% and L50%. These
data are to be presented by LED component manufacturers and used by LED
system manufacturers.
Method
For the purpose of measuring useful life, the component should first be operated
for a 1,000-hour seasoning period at the rated current and voltage. Then the
component should be monitored while operating at rated current for at least an
additional 5,000 hours (for a total of 6,000 hours) at three different temperatures
(Ts) measured directly on the component. Ts is the temperature of the
thermocouple attachment point on the LED. This point is usually the solder joint
or the closest measurable location to the LED junction. Manufacturers should
diagram the thermocouple attachment point. For high-power LED components
(those operated above 100 mA), the recommended Ts temperatures are 45°C,
65°C, and 85°C. For low-power LEDs (those operated below 100 mA), the
recommended Ts temperatures are 35°C, 45°C, and 55°C.
It may be assumed that the spatial and spectral distributions from the component
do not change significantly over its useful life. A broadband detector measuring
radiant flux may be used as a surrogate for luminous output rather than requiring
expensive V() correction in measurement equipment, since it is the relative
output change that is of interest.
If L70% and L50% are not reached within 6,000 hours, a functional fit (provided by
the LED or system manufacturer1) to the data between 1,000 and 6,000 hours
(i.e., excluding the first 1,000 hours) can be used to extrapolate to 70% and 50%
lumen maintenance.
For the purpose of this measurement method, lumen maintenance is defined as

the light output as a percentage of the component’s output after the 1,000-hour
seasoning period. The 1,000 hour data value is normalized to 100%. The reason
for omitting the initial 1,000 hours is because for most LEDs, the light output
increases during this period. After this period, the light output begins to decrease.
1
No single type of functional fit (for example, an exponential fit) is applicable to all LED packages because package
configuration variations can result in different degradation mechanisms, which affect the shape and rate of light
output depreciation.
4
recommends…
The time it takes for an LED to reach this maximum depends on the device and
the temperature.
As product performance improves in the future, the initial seasoning period may
get longer, and a measurement period greater than 6,000 hours may be
necessary to develop reasonable predictions of light output.
Additional Data
To aid the successful integration of LEDs into a system, system manufacturers
require additional measurements and documentation. The following data
pertaining to the conditions of measurement should also be reported:
 Number of product samples tested
 Heat sink description (size, shape, materials, etc.)
 Ambient temperatures corresponding to the Ts temperatures
 r: thermal resistance coefficients of the LED junction to the location
where the thermocouple is mounted in °C/W, corresponding to the three
test temperatures
 V: voltage across the leads, reported at the three test temperatures
 I: current through the device, reported at the three test temperatures
These data permit estimation of the junction temperature (Tj) that corresponds to
the life data.
Tj = Ts + V·I·r
About ASSIST
5
recommends…

Measurement Method for LED
Systems
Volume 1, Issue 3
February 2005
Revised August 2007

recommends…
Copyright © 2005, 2007 by the Alliance for Solid-State Illumination Systems and Technologies
(ASSIST).
Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union St.,
Troy, New York, USA.
All rights reserved. No part of this publication may be reproduced in any form, print, electronic,
or otherwise, without the express permission of the Lighting Research Center.
This ASSIST recommends. . . was prepared by the Lighting Research Center at the request of
the Alliance for Solid-State Illumination Systems and Technologies (ASSIST). The
recommendations set forth here are developed by consensus of ASSIST sponsors and the
Lighting Research Center. ASSIST and the Lighting Research Center may update these
recommendations as new research, technologies, and methods become available.
Boeing
Cree
Lite-On
GE Lumination
OSRAM SYLVANIA
Philips Lighting
Seoul Semiconductor
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Contents
Purpose.......................................................................................................................... 4
Scope ............................................................................................................................. 4
Method ........................................................................................................................... 4
Additional Data .............................................................................................................. 5
About ASSIST ................................................................................................................ 5
3
recommends…
Purpose
This document outlines a method for measuring useful life for light-emitting diode
(LED) systems used for general lighting applications. The intent of this document
is to encourage common methods of testing and data presentation. The target
audience is LED system manufacturers.
Scope
An LED system is defined as the integration of all necessary components into a
working module, such as a light fixture. An LED system may contain one or more
LED components along with a driver and housing.
The measurement method defined here is to be used to determine the

operational life in hours for an LED system to reach L70% and L50%. These data
are to be presented by LED system manufacturers and used by lighting
specifiers.
Method
For the purpose of measuring useful life, the system should first be operated for
a 1,000-hour seasoning period at the rated current and voltage. Then, the system
should be monitored while operating at rated current for at least an additional
5,000 hours (for a total of 6,000 hours) in a ventilated environment at an ambient
temperature of 25°C. For systems intended to operate in environments where the
heat buildup causes the LED junction temperature to increase, the LED system
should be tested at temperatures (Ts) corresponding to those environments for
which they are rated. Ts, measured directly on the LED component, is the
temperature of the thermocouple attachment point on the LED or the LED board.
This point is usually the solder joint or the closest measurable location to the LED
junction. Ts must be first determined by operating the system under conditions
similar to the one or more environments where it would be used. For example,
indoor directional lighting systems typically operate in three different
environments, namely open air, semi-ventilated, and enclosed.1
It may be assumed that the spatial and spectral distributions from the system do
not change significantly over its useful life. A broadband detector measuring
radiant flux may be used as a surrogate for luminous output rather than requiring
expensive V() correction in measurement equipment, since it is the relative
output change that is of interest.
If L70% and L50% are not reached within 6,000 hours, a functional fit (provided by
the LED or system manufacturer2) to the data between 1,000 and 6,000 hours
(i.e., excluding the first 1,000 hours) can be used to extrapolate to 70% and 50%
lumen maintenance. For the purpose of this measurement method, lumen
maintenance is defined as the light output as a percentage of its output after the
initial 1,000-hours. The 1,000 hour data value is normalized to 100%. The reason
1
ASSIST recommends… Recommendations for Testing and Evaluating Luminaires Used in Directional Lighting. Vol.
3, Issue 3. Internet: http://www.lrc.rpi.edu/programs/solidstate/assist/recommends/directional.asp
2
No single type of functional fit (for example, an exponential fit) is applicable to all LED packages because package
configuration variations can result in different degradation mechanisms, which affect the shape and rate of light
output depreciation.
4
recommends…
for omitting the initial 1,000 hours is because for most LEDs, the light output
increases during this period. After this period, the light output begins to decrease.
The time it takes for an LED to reach this maximum depends on the device and
the temperature.
As product performance improves in the future, the initial seasoning period may
get longer, and a measurement period greater than 6,000 hours may be
necessary to develop reasonable predictions of light output.
Additional Data
In addition to providing L70% and L50% life data, LED system manufacturers should
also provide the number of product samples tested.
About ASSIST
5
recommends…

Sample Data Sheet for High-power
LEDs
Volume 1, Issue 4
February 2005

recommends…
(ASSIST).

Center.
available.
Boeing
Cree
Lite-On
GE Lumination
OSRAM SYLVANIA
Philips Lighting
Seoul Semiconductor
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Data Sheet – High-power LED Life

This document outlines the life data and testing information that component
manufacturers of high-power LEDs should provide. This data is to be measured
under specific conditions as outlined in the accompanying document, "LED Life
for General Lighting — Measurement Method for LED Components,” Vol. 1,
Issue 2.
Product Information
Product Name ________________________________________________________________________
Product Number ______________________________________________________________________
Test Date(s) _________________________________________________________________________
Testing Procedure
[Briefly outline the procedure and equipment used to derive life data presented in this document. Include
a diagram showing the thermocouple attachment point.]
Life Data
Ts 45°C 65°C 85°C

L70% (hours)
L50% (hours)
* Ts is the temperature of the point on the LED where the thermocouple is attached. This point is usually
the solder joint or closest measurable location to the LED junction.
45°C Life Graph 65°C Life Graph 85°C Life Graph

[Insert life graph here] [Insert life graph here] [Insert life graph here]
Additional Data
Number of product samples tested ________________________________________________________
Heat sink description (size, shape, materials, etc.)____________________________________________
Ambient temperatures corresponding to Ts temperatures:

@Ts 45°C ___________ @Ts 65°C ___________ @Ts 85°C ___________
Thermal resistance coefficients of the LED junction to the Ts measurement point, in °C/W (θr):
@Ts 45°C ___________ @Ts 65°C ___________ @Ts 85°C ___________
Voltage across the leads (V):

@Ts 45°C ___________ @Ts 65°C ___________ @Ts 85°C ___________
Current through the device (I):

@Ts 45°C ___________ @Ts 65°C ___________ @Ts 85°C ___________
3
recommends…
Junction temperature (Tj = Ts + V·I·θr):

@Ts 45°C ___________ @Ts 65°C ___________ @Ts 85°C ___________
About ASSIST
4
recommends…

Sample Data Sheet for Low-power
LEDs
Volume 1, Issue 5
February 2005

recommends…
(ASSIST).

Center.
available.
Boeing
Cree
Lite-On
GE Lumination
OSRAM SYLVANIA
Philips Lighting
Seoul Semiconductor
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Data Sheet – Low-power LED Life

This document outlines the life data and testing information that component
manufacturers of low-power LEDs should provide. This data is to be measured
under specific conditions as outlined in the accompanying document, "LED Life
for General Lighting — Measurement Method for LED Components,” Vol. 1,
Issue 2.
Product Information
Product Name ________________________________________________________________________
Product Number ______________________________________________________________________
Test Date(s) _________________________________________________________________________
Testing Procedure
Life Data
Ts 35°C 45°C 55°C

L70% (hours)
L50% (hours)
* Ts is the temperature of the point on the LED where the thermocouple is attached. This point is usually
the solder joint or closest measurable location to the LED junction.
35°C Life Graph 45°C Life Graph 55°C Life Graph

Additional Data
Number of product samples tested ________________________________________________________
Heat sink description (size, shape, materials, etc.)____________________________________________
Ambient temperatures corresponding to Ts temperatures:

@Ts 35°C ___________ @Ts 45°C ___________ @Ts 55°C ___________
Thermal resistance coefficients of the LED junction to the Ts measurement point, in °C/W (θr):
@Ts 35°C ___________ @Ts 45°C ___________ @Ts 55°C ___________
Voltage across the leads (V):

@Ts 35°C ___________ @Ts 45°C ___________ @Ts 55°C ___________
Current through the device (I):

@Ts 35°C ___________ @Ts 45°C ___________ @Ts 55°C ___________
3
recommends…
Junction temperature (Tj = Ts + V·I·θr):

@Ts 35°C ___________ @Ts 45°C ___________ @Ts 55°C ___________
About ASSIST
4
recommends…

Sample Data Sheet for LED Systems
Volume 1, Issue 6
February 2005
Revised August 2007

recommends…
Copyright © 2005, 2007 by the Alliance for Solid-State Illumination Systems and
Technologies (ASSIST).

Center.
available.
Boeing
Cree
Lite-On
GE Lumination
OSRAM SYLVANIA
Philips Lighting
Seoul Semiconductor
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Data Sheet – LED System Life

This document outlines the life data and testing information that manufacturers of
LED systems should provide. This data is to be measured under specific
conditions as outlined in the accompanying document, "LED Life for General
Lighting — Measurement Method for LED Systems,” Vol. 1, Issue 3.
Product Information
Product Name ________________________________________________________________________
Product Number ______________________________________________________________________
Test Date(s) _________________________________________________________________________
Testing Procedure
Life Data
Operating Ts at 25°C ambient Ts1 as determined for Ts2 as determined for

Temperature temperature additional additional
environment environment
L70% (hours)
L50% (hours)
25°C Life Graph Ts1 Life Graph Ts2 Life Graph
Additional Data
Number of product samples tested________________________________________________________
About ASSIST
3
recommends…

Recommendations for the Definition
and Specification of Useful Life for
Light-emitting Diode Light Sources
Volume 1, Issue 7
February 2005
Revised April 2006

recommends…
Copyright © 2005, 2006 by the Alliance for Solid-State Illumination Systems and
Technologies (ASSIST).

Center.
available.
Boeing
Cree
Lite-On
GE Lumination
OSRAM SYLVANIA
Philips Lighting
Seoul Semiconductor
John Bullough
Yimin Gu
Nadarajah Narendran
Jennifer Taylor
2
recommends…
Contents
Introduction ..................................................................................................................... 4
Background ..................................................................................................................... 4
Proposed Definition ......................................................................................................... 4
Rationale ......................................................................................................................... 5
Comparison Among Technologies ........................................................................................................... 5
Human Response ..................................................................................................................................... 6
Conclusions..................................................................................................................... 7
References...................................................................................................................... 8
About ASSIST ................................................................................................................. 8
3
recommends…
Introduction
This document outlines a recommendation for the definition and specification of
useful life for light sources and lighting systems using light-emitting diodes
(LEDs). It was developed by the Lighting Research Center (LRC) at Rensselaer
Polytechnic Institute in collaboration with members of the Alliance for Solid-State
Illumination Systems and Technologies (ASSIST).
Background
Traditionally, the definition of life for light sources is based on the median
operating time that can be expected under specified conditions (Rea, 2000). For
example, fluorescent lamps are operated on a 3-hr-on, 20-min-off cycle at a
temperature of 25°C. After a certain amount of time, the light source is expected
to fail and no longer function. Based on this definition, the life of LED sources
would often be in excess of 100,000 hr because the solid-state components have
no obvious failure mechanism. However, it has been shown that nearly all light
sources, including LEDs, experience both gradual reductions in light output and
gradual shifts in color appearance.
The term lumen maintenance refers to the output of a light source at a given
instance relative to its initial light output when brand new. Narendran et al. (2000,
2001) demonstrated that some LED sources (e.g., 5-mm packages designed for
indicator applications), when operated under rated conditions, reach less than
50% lumen maintenance well before 10,000 hr—much shorter than the 100,000+
hr life until failure. More recent LED configurations designed to provide
illumination maintain lumen output for much longer periods of time, with
approximately 80% lumen maintenance at 20,000 hr.
Color shifts as a function of time have also been demonstrated in LEDs.

Narendran et al. (2001) demonstrated shifts in chromaticity of individual 5-mm
indicator packages approaching the size of a four-step MacAdam ellipse (thought
to be detectable by well over 99% of the population; see MacAdam, 1942) after
only 1100 hr of operation. Similar shifts were observed for individual illumination-
grade devices after 2000 hr of operation, demonstrating improvements in control
of color shifts (Narendran et al., 2003).
Since a light source that produces only a fraction of its initial light output, or that
shifts considerably in color appearance over time, is arguably not a very useful
source, a definition of useful life that is an alternative to the conventional time
until failure is proposed. The alternative definition takes into account the longevity
of LEDs as well as their lumen maintenance and color shift characteristics. For
lumen maintenance issues, two levels of useful life are proposed based on the
variety of applications for which LEDs might be used. For color shifts, it is
proposed that the useful life definition be invoked only when color appearance is
important for the intended application.
Proposed Definition
It is proposed that the useful life of an LED light source (e.g., a bare LED
component) or an LED lighting system with regard to lumen depreciation be
defined in two ways: one is the length of time that source or system takes, when
operated at a specified temperature, to reach 70% of its initial light output (L70%);
4
recommends…
and the other, the length of time to reach 50% (L50%) of its initial light output
under the same operating conditions. With regard to color shift, it is proposed
that throughout the useful life, the chromaticity coordinates of the source or
system should not fall outside a four-step MacAdam ellipse (L4M) containing the
initial chromaticity coordinates. It should be noted that the useful life of an LED
light source is unlikely to remain the same when it is integrated into a lighting
system. End users of a particular LED lighting system should seek the system's
life value.
Rationale
Comparison Among Technologies
A comparison of lumen maintenance characteristics of light sources (Rea, 2000;
Bullough, 2003) demonstrates that most conventional light sources (e.g., filament
lamps, fluorescent lamps and high pressure sodium lamps) exhibit greater than
70% lumen maintenance by the time they have operated for their rated life, or
20,000 hr, whichever comes first (Figure 1). Indeed, much of the trade literature
in the lighting industry (e.g., Anonymous, 1998, 2002; Knisley, 2001; Simpson,
2002) points to lumen maintenance values in excess of 80% or 90% as being
desirable characteristics of general lighting sources.
Lumen Maintenance Values for Various Light Sources
100% 400W HPS

400W MH
lumen maintenance (%)
90%
400W MV
80% 28W T5
32W T8
70% 5-m m LED
Hi-LED
60%
100W Inc.
50% 50W TH
0 5000 10000 15000 20000
operating time (hr)
Figure 1. Lumen maintenance curves for different light sources out to 20,000 hours (Rea,
2000; Bullough, 2003).
Important exceptions to this analysis are mercury vapor and metal halide lamps,
which regularly reach less than 70% of initial light output during their rated
lifetime. Since mercury vapor lamps have relative low luminous efficacy and color
rendering indices compared to metal halide lamps, these sources are not
considered further in this document. Regarding metal halide lamps, their
relatively low lumen maintenance attributes are recognized in the literature
(Brates and Hrubowchak, 1999; Fetters, 1999), and lumen maintenance values
of 80% to 85% are being cited as appropriate short-term target values for these
lamps in the future.
Relatively few data are available on color shifts with different light sources over
time, with the exceptions of LEDs, described above (Narendran et al., 2001,
2003), and metal halide lamps (Carleton et al., 1997; Krasko et al., 1998).
Carleton et al. (1997) showed that metal halide lamps with quartz arc tubes
experienced shifts in chromaticity greatly exceeding a four-step MacAdam ellipse
5
recommends…
after 5000 hr of operation, and lamps with ceramic arc tubes remaining well
within a four-step MacAdam ellipse in terms of chromaticity. Krasko et al. (1998)
did not report chromaticity shifts but rather shifts in correlated color temperature,
demonstrating shifts of 200-300 K by 4000 hr, which approximates in magnitude
the color shift described by a four-step MacAdam ellipse.
Human Response
The Illuminating Engineering Society of North America (IESNA) (Rea, 2000)
states that a difference in illuminance of about one-third (~33%) from a given
value constitutes a dramatic difference. If, as is often the case, the goal of a
general lighting installation is not to look dramatically different from the norm, a
difference of 30% in light level (e.g., no less than 70% of the initial level) would
provide a reasonable basis for avoiding dramatically noticeable differences.
Changes in light output caused by lumen maintenance reductions are very slow
and gradual. While there is extensive research on the visual response to very
rapid fluctuations in light level (e.g., flicker), the lighting and vision science
literature is much sparser regarding the response to slow changes in light level.
With recent developments in the electric utility industry in North America and
Europe, however, the concept of using reductions in light levels through dimming
during periods of peak electricity use has become a potentially attractive option
for avoiding blackouts. In that context, several studies of the responses of
occupants to relatively slow (over several seconds or minutes) reductions in light
output are of interest and could also provide insight as to the detectability of light
level reductions over the life of a lighting installation:
• Knau (2000) measured the ability of observers to detect differences in

light level viewed as a perfectly uniform, full field of view. Knau (2000)
found that when the initial luminance was 10 or 100 cd/m2 and was
reduced slowly, the change in level was detected after a reduction of
about 20% to 40%.
• Kryszczuk and Boyce (2002) measured the time taken by observers to
notice illuminance reductions in a simulated office environment.
Regardless of the initial illuminance (about 500 or 1000 lx), the speed of
reduction (from about 4 to 300 lx/s), or whether the occupant was given a
mental distraction task, the difference was typically noticed when the
illuminance reduction was between 17% and 22%.
• Shikakura et al. (2003) found that observers could notice a 10%
immediate reduction in illuminance (from 450 or 700 lx) if they were
focused only on trying to notice that difference. When observers were
given typical office-type tasks, the noticeable difference increased to
20%. Increasing the length of time for the decrease from 2 to 8 s tended
to increase the noticeable difference only when an office-type task was
simultaneously performed.
• Akashi and Neches (2004) also measured observers' ability to detect
slow (over a period of 10 s) reductions in illuminance (from 500 lx) and
found that reductions of about 15% to 30% were needed before they
could be detected by a majority of subjects. In addition, a majority of
subjects found a 30% reduction acceptable (for the purpose of reducing
peak load), even if they could detect it.
In general, these studies suggest that reductions in light level to about 70% to
80% of its initial value are the likely minimum change that would not be detected
reliably by observers. Likely, the characteristics of the specific luminous
6
recommends…
environment and the nature of the tasks being performed impact the magnitude
of the detectable reduction, given differences among the studies. The evidence
from Akashi and Neches (2004) indicates that even when reductions in light level
are detected, they can still be considered acceptable.
Regarding color shifts, Narendran et al. (2000) reported that differences in

chromaticity approximating a four-step MacAdam ellipse appeared to be a
reasonable tolerance for perceptions of color differences in a display containing a
variety of colored objects. Since color shifts over time would not be seen in a
side-by-side but rather in sequential mode, such a color shift magnitude appears
to be a reasonable one for minimizing detection of such shifts.
Lumen Maintenance at 20,000 Hours Range of Just Detectable Illuminance Reductions

100% 100%
Illuminance at detection (% of initial)

Lumen maintenance (%)
80%
90%
60%
80%
40%
70%
20%
60%
0%
T5
T8
TH*
Inc.*
Hi-LED
MH
5-mm LED**
HPS
MV
50%
Knau Kryszczuk Shikakura Akashi
Light source Study

*rated life less than 20,000 hours
**lumen maintenance <50% before 20,000 hours
Figure 2. (a; left) Lumen maintenance values for light sources at 20,000 hours or end of life.
(b; right) Summary of detectable illuminance reductions from several studies.
Conclusions
Based on the performance characteristics of conventional lighting technologies
regarding lumen maintenance (summarized in Figure 2a), as well as the
responses of people to the lighted environment (Figure 2b), a lumen
maintenance value of 70% is justified as the basis for determining the end of
useful life for an LED light source in general lighting applications (L70%). Of
course, this recommendation is applicable only to general illumination
applications such as those found in commercial interior lighting. For some
applications where uniform appearance of the lighted environment is one of the
primary criteria (e.g., wall washing in a corridor), a more stringent requirement,
such as 80% lumen maintenance, could be necessary. Yet other applications,
such as decorative lighting, could well have much less stringent requirements,
provided by a 50% criterion for light output (L50%). Applications such as
emergency lighting, for example, are subject to specific regulations and
requirements and are considered outside the scope of this definition.
When color appearance of lighting is an important consideration, a further useful

life value is proposed based on the tolerances of human perception for detecting
color shifts. Using the time for a light source to shift in chromaticity outside a four-
step MacAdam ellipse containing the source's initial chromaticity coordinates
(L4M) is therefore proposed as a basis for useful life when color appearance
consistency over time is important.
7
recommends…
References
Akashi Y, Neches J. 2004. Detectability and acceptability of illuminance
reduction for load shedding. J Illum Eng Soc 33(1): 3-13.
Anonymous. 1998. Demand and regs lift lamp demand slightly. Purchasing
124(4): 96-98.
Anonymous. 2002. Buildings magazine: Editors' choice top 100 products
2002. Buildings 96(12): 25-56.
Brates N, Hrubowchak DM. 1999. Investigations of sodium-scandium and
sodium-rare earth chemistries for 4000K, high CRI metal halide lamps. J
Illum Eng Soc 28(1): 103-116.
Bullough JD. 2003. Lighting Answers: Light Emitting Diode Lighting Systems.
Troy: National Lighting Product Information Program, Lighting Research
Center, Rensselaer Polytechnic Institute.
Carleton S, Seinen PA, Stoffels J. 1997. Metal halide lamps with ceramic
envelopes: A breakthrough in color control. J Illum Eng Soc 26(1): 139-
145.
Fetters JL. 1999. Lighting efficiency: Making it happen, part I. Elec Const
Maint 98(4): 40-42.
Knau H. 2000. Thresholds for detecting slowly changing Ganzfeld
luminances. J Opt Soc Am A 17(8): 1382-1387.
Knisley J. 2001. Standard knowledge. Elect Const Maint 100(4): 84-86.
Krasko Z, Brates N, Nortrup E. 1998. Optimization of metal halide lamp
design for better color rendering and color control. J Illum Eng Soc 27(2):
152-159.
Kryszczuk K, Boyce PR. 2002. Detection of slow light level reduction. J Illum
Eng Soc 31(2): 3-10.
MacAdam DL. 1942. Visual sensitivities to color differences in daylight. J Opt
Soc Am 32(5): 247-274.
Narendran N, Vasconez S, Boyce PR, Eklund N. 2000. Just-perceivable
color differences between similar light sources in display lighting
applications. J Illum Eng Soc 29(2): 68-77.
Narendran N, Maliyagoda N, Bierman A, Pysar RM, Overington M. 2000.
Characterizing white LEDs for general illumination applications Proc
SPIE Int Soc Opt Eng 3938, pp. 240.
Narendran N, Bullough JD, Maliyagoda N, Bierman A. 2001. What is useful
life for white light LEDs? J Illum Eng Soc 30(1): 57-67.
Narendran N, Deng L, Pysar RM, Gu Y, Yu H. 2003. Performance
characteristics of high-power light emitting diodes. Proc SPIE Int Soc Opt
Eng 5187.
Rea MS (ed.). 2000. IESNA Lighting Handbook: Reference and Application,
9th ed. New York: Illuminating Engineering Society of North America.
Shikakura T, Morikawa H, Nakamura Y. 2003. Perception of lighting
fluctuation in office lighting environment. J Light Vis Env 27(2): 75-82.
Simpson T. 2002. Fluorescent and HID duke it out. Plant Eng 56(10): 60-62.
About ASSIST

ASSIST LEDLife Revised2007 PDF

Uploaded by

Copyright:

Available Formats

ASSIST LEDLife Revised2007 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ASSIST LEDLife Revised2007 PDF

Uploaded by

Copyright:

Available Formats

recommends…

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies

Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union

ASSIST Recommends Sponsors

Lighting Research Center Technical Staff

Because of these characteristics, the ASSIST program has developed a set of

Components and Systems

L70% (hours): time to 70% lumen maintenance

L50% (hours): time to 50% lumen maintenance

These times are to be measured under specific conditions as outlined in the

for a wide array of lighting applications. Indeed, lumen maintenance values

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies

ASSIST Recommends Sponsors

Lighting Research Center Technical Staff

The measurement method defined here is to be used to determine the

For the purpose of this measurement method, lumen maintenance is defined as

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies

ASSIST Recommends Sponsors

Lighting Research Center Technical Staff

The measurement method defined here is to be used to determine the

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies

Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union

ASSIST Recommends Sponsors

Lighting Research Center Technical Staff

Data Sheet – High-power LED Life

Ts 45°C 65°C 85°C

45°C Life Graph 65°C Life Graph 85°C Life Graph

Heat sink description (size, shape, materials, etc.)____________________________________________

Ambient temperatures corresponding to Ts temperatures:

Voltage across the leads (V):

Current through the device (I):

Junction temperature (Tj = Ts + V·I·θr):

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies

Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union

ASSIST Recommends Sponsors

Lighting Research Center Technical Staff

Data Sheet – Low-power LED Life

Ts 35°C 45°C 55°C

35°C Life Graph 45°C Life Graph 55°C Life Graph

Heat sink description (size, shape, materials, etc.)____________________________________________

Ambient temperatures corresponding to Ts temperatures:

Voltage across the leads (V):

Current through the device (I):

Junction temperature (Tj = Ts + V·I·θr):

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies

Published by the Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union

ASSIST Recommends Sponsors

Lighting Research Center Technical Staff

Data Sheet – LED System Life

Operating Ts at 25°C ambient Ts1 as determined for Ts2 as determined for

25°C Life Graph Ts1 Life Graph Ts2 Life Graph

LED Life for General Lighting:

A publication of the Alliance for Solid-State Illumination Systems and Technologies