Professional Documents
Culture Documents
Determination of Maintenance Factors - GN11 ILP
Determination of Maintenance Factors - GN11 ILP
Maintenance factor
determination and its
impacts on the
performance and overall
efficiency of LED
luminaires
Copyright © 2020 ILP
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system
or transmitted in any form or by any means: electronic, electrostatic, magnetic tape,
mechanical, photocopying, recording or otherwise, without permission in writing from the
Institution of Lighting Professionals.
Website: www.theilp.org.uk
Acknowledgements
The development of this document has involved many people and ILP wishes to express its
appreciation to them all. In particular, the following have had a significant input to the
completion of this work:
David Lodge, CU Phosco Ltd Stuart Morton, Jacobs
Allan Howard, WSP Nigel Parry, Orangetek
1.1
calculated accurately.
1.0 It is clear that the architecture of the LED,
the LED package, the circuit arrangement
0.9
on the LED board, the driver type and
0.8
Ts = 55°C, lf = 700mA method, the number of switching cycles,
Ts = 85°C, lf = 700mA
Ts = 105°C, lf = 700mA the power supply quality and the
0.7 TM–21 limits probability of being struck by lightning etc
TM–21 L70(10k) will all have an effect on the actual useful
> 60,000hrs
0.6 life of the LED products5. The median
103 104 105 useful life based solely on TM-21
Hours
extrapolations of lumen depreciation is
Figure 1: Typical TM–21 graphs
likely to overstate the life of the luminaire
consistent values without spikes or by omitting these effects. However, TM-21
variations, humidity is controlled, and no is currently the best available and most
chemical contamination or air pollution is commonly used method for projecting long
present. The tests also exclude any effects term lumen maintenance of LED light
of switching cycles on the LED sources.
performance or life. Switching cycles may
Luminaire maintenance factor, covering
have a life limiting effect on the LED
dirt build-up and cleaning cycles, should
module and LED driver.
be chosen in accordance with BS 5489-
LSF is impractical and uneconomic to 1:2013 Annex B. The original research on
determine accurately due to the time which the dirt build-up figures was based
required for testing a statistically used HID luminaires but they are expected
significant population of luminaires to end to be representative for LED luminaires
of life. LSF = 0.98 or LSF = 0.97 are until the research is repeated with LED
common but these figures are typically luminaires.
estimated by LED component suppliers
using statistical methods. LSF of 0.97 BS EN 62717 LED lamps and LED
representing 3% failures of individual LEDs
modules6
in a working luminaire shall be used unless
evidence is provided by the LED supplier The publication of BS EN 62717 LED
or luminaire manufacturer to show an modules for general lighting is used to
alternative value of LSF is appropriate. The define the performance of LED modules by
3% has been selected arbitrarily as carrying out tests and considering the
representing values typically used by change in performance of the luminaire
manufacturers in the industry. It provides between the start and end of the test.
a consistent benchmark that avoids results
While BS EN 62717 relates only to LED
being skewed by under-estimating or
modules, it is directly aligned with BS EN
5 Possible LED PCB architectures and the impacts of 62722-1 Luminaire Performance and the
LED failures resulting from these are discussed in definitions for Lx are used throughout to
Appendix E.
represent performance at the end of the
6 This standard is published as a voluntary method of 6,000 hour test period.
establishing a presumption of compliance with the
eco-design requirements of: Clause 10.2 discusses lumen maintenance
Commission Regulation (EC) No. 244/2009 (see BS
EN 62717 Annex ZZA) for Non-directional Household
over the test period and defines a code to
Lamps represent the luminaire performance
Commission Regulation (EC) No 1194/2012 (see BS based on the testing period. Notes 1 to 4
EN 62722 Annex ZZB) for Directional Lamps, LED
Lamps and related equipment should be referred to for the limitations of
Commission Delegated Regulation (EU) No 874/2012 the guidance provided. Note 1 recognises
(see BS EN 62717 Annex ZZC) for Energy Labelling that it is impractical to measure the lumen
of Electrical Lamps and Luminaires.
reduction over the life of the LED module of 10,000 to 30,000 hours and no
and recommends that the testing is suggestion is made that this would
carried out up to 25% of the rated life of adequately predict the performance at
the luminaire, to a maximum of 6,000 50,000 hours or 100,000 hours.
hours (see also BS EN 62717 clause 6.1).
The methods set out in BS EN 62717 do
For luminaires with more than 24,000
not suggest Lx,By represents the
hours rated life, this defines the test time
performance of the luminaire at the end of
as 6,000 hours.
the scheme life, and so the Lx,By figures
Defining the maintained light level, Lx, in should not be used for calculated
BS EN 62717 as the maintained flux at the maintenance factors. For the same reason
end of the 6,000-hour test period creates Lx,B10 should not be used by customers to
confusions because BS 5489-1 refers to specify the minimum performance at the
the maintained light levels at the end of end of the rated life or by manufacturers
the useful life of the luminaire. As such the to make claims of luminaire performance
Lx parameters as defined in BS EN 62717 at the end of the rated life. These figures
are not suitable for use in determining the are typically given as marketing material
maintenance factor of luminaires. and are not based on evidence of lifetime
testing.
BS EN 62717 Annex A sets out normative
guidance on the use of IES LM-80 (see When to use Lx,B50 and Lx,B10
page 4 above) for lumen maintenance, but
again this only considers the maintained In cases where LED optical distributions
lumen output at the end of a test of a are using ‘overlay’ optics and all LEDs in
single LED. the module are lighting the same area, for
example street lighting or area flood
Annex C introduces By in the term Lx,By lighting, the use of Lx,B50, the median
giving a representation of the tightness or useful life from IES TM-21, is sufficient
spread of light output levels across a and significantly simplifies the calculation
population of LED modules7. By defines the of maintenance factors. Note: the validity
proportion, y (in %), of the population of of the prediction is limited to 5.5 or 6
LEDs that has fallen below the threshold times the test duration, and so any
light output Lx where x is the remaining extrapolation from TM-21 graphs beyond
percentage of the initial light output. Lx that limit should be treated as indicative
B50 defines the median useful life as the rather than claimed performance.
time taken for 50% of the population of
LEDs of the same type to have fallen Where each LED or group of LEDs is
below the lumen maintenance defined by illuminating a different section of the lit
Lx. The spread of light output levels can be object, for example linear luminaires
illustrated as a bell curve (normal providing wall washing, any difference in
distribution) of maintained light output illuminance level would be obvious and
levels of a population of LEDs or LED would detract from the artistic effect. In
modules at the stated time. Unfortunately, these cases, the distribution of lumen
this provides only informative explanations depreciation across the products supplied
of LED product full lifetime metrics as no will need to be more tightly controlled and
standardised methods of calculating or the customer may specify the luminaires
testing the parameters are offered, and so using Lx,B10 to prevent a visible shift of
claimed performance of different the light colour across the wall. That said,
manufacturers’ products might vary and the light levels required at the end of the
therefore may not be comparable. Also, useful life of the installation should still be
the life of the LED modules given in the calculated based on Lx,B50 for the median
example is typically indicating a rated life useful life.
Note that there are some unexpected
7 See Appendix B regarding uncertainties related to consequences from specifying tighter
using By.
the initial LED module flux (excluding a. Initial luminous flux of the complete
optical losses and depreciation over life) luminaire (including optics and any
and comparing it with initial LED power diffuser, bowl or glass front) shall be
consumption (reduced by the CLO factor). measured on a goniophotometer or an
Figures that appear to be more than integrating sphere with the drive
150lm/W are almost certainly not current set at the designated nominal
representative of the real world energy value and the CLO factor set at 100%
consumption over the life of the product. representing the end-of-life power
Care should be taken during the consumption condition.
preparation of tender specifications and b. Testing shall be completed at ambient
comparisons of tender responses to air temperature, Ta = 25˚C.
determine if such figures are reasonable, c. The initial luminous flux multiplied by
and evidence of the calculations and the BS 5489-1 (LLMF × LSF) elements
testing behind the numbers should be of the maintenance factor to give the
requested if in doubt; see Appendix D. maintained luminous flux at end of
life.
In an attempt to provide a representative
d. LLMF based on IES TM-21 prediction
and fair comparison between different
using the nearest higher temperature
products from different manufacturers, an
curve to the LED reference (solder
overall efficiency parameter is proposed.
point or junction) temperature taken
The overall efficiency can be specified in
from the in-situ temperature
the tender documents requiring each
measurement test at ambient air
manufacturer to claim this to give a
temperature, Ta=25˚C.
standard assessment process.
e. LSF taking account of both the
number of LED failures expected
Overall efficiency during the life of the luminaire (use
actual figures where available but no
To make it easier for customers and
less than the greater of 3% of LEDs or
purchasers to compare luminaires using
one LED failure per LED module).
datasheets, the Institution of Lighting
f. The effects of the LED module circuit
Professionals is asking luminaire
architecture taken into account when
manufacturers to publish, and purchasers
assessing the impact of a single LED
to specify publication in their contracts, a
failure on the performance of a
single standard comparison efficiency
luminaire. This may result in adverse
figure for each luminaire model that we
drive current distribution around the
will call the overall efficiency.
circuit, leading to sequential failure of
The overall efficiency shall be the further LEDs or in the worst case, the
maintained luminous flux divided by the whole circuit board going out of light.
average CLO power consumption of the 3. The average CLO power consumption of
complete luminaire determined as follows: the complete luminaire using the Elexon
1. the maintained luminous flux and guidance on determining the CLO
average CLO power consumption of the power, using a straight-line power
luminaire measured with the luminaire increase between the initial power and
drive current greater than or equal to end-of-life power, including driver
700mA to each high-power LED or losses.
100mA drive current to each mid-power
The overall efficiency will allow the typical
LED regardless of the circuit
performance of two luminaires to be
architecture being series, parallel,
compared including all relevant factors
ladder, or series parallel.
including initial LED flux, LED lumen
2. Maintained luminaire flux shall be
depreciation, LED survival rates, thermal
measured using the following
management and optical efficiency.
procedure:
Conclusions
Maintenance Factors methods in BS 5489-
1 are relevant for HID luminaires and
should be calculated using the methods in
BS 5489-1.
The methodology in BS PD ISO/CIE TS
22012: 2019 is the best practice approach
for calculation of maintenance factors for
LED luminaires.
The guidance in BS EN 62717 and BS EN
62722 is not suitable for the calculation of
maintenance factors for any luminaires.
Lx,By should not be specified for the
performance over the rated life of a
luminaire as the methods in BS EN 62717
and BS EN 62722 relate to the
performance at the end of the test period
not the end of the rated luminaire life. The
median useful life, Lx,B50 shall be used for
calculating maintenance factors.
A standard method is presented for
calculating the overall efficiency of the
luminaire. This method outlines a testing
and calculation method using standardised
methods and co-ordinated operating
parameters to ensure performance and
overall efficiency claims for luminaires are
reliable.
Appendices
A. LED standards
B. Uncertainties in useful life predictions
C. Abrupt failure
D. Luminaire manufacturers performance
data
E. LED failure modes and consequences
All out
Series circuit
O/C
Total failure
I1
I1 increases dramatically
S/C Vf must be closely matched
I2 I2 decreases dramatically
I1
Maintenance factor determination of LED luminaires
I1 = 0
O/C Vf must be closely matched
I2 I2 = Full drive current
Ladder circuit
O/C
of others
Full current
Short circuit Series Light output reduces No change Improves As single LED failure.
Guidance Note 11/20
Open circuit Series Luminaire fails No current Not in light Not in light after first LED
flow failure
Open circuit Series parallel Light reduces slightly. Drive current Increase in drive Single LED failure causes chain
(several long No light output from re-routed via current to of LEDs to be out of light.
series circuits in LED chain with failed working LEDs parallel chain Increases risk of future failures
parallel) LED. Parallel chains in parallel increases Tj of remaining LEDs which in turn
increase light output chains significantly, lead to the remaining full
but don’t fully increases chains failing.
compensate due to depreciation and
increased junction reduces life.
temperature. Increase in drive
current may
lead to over-
running LED.
17
Maintenance factor determination of LED luminaires
18
LED failure Circuit Single LED failure – effect on… Multiple LED failure effect
architecture
Light output Drive Life of
current to remaining
other LEDs LEDs
Short circuit Ladder Light output will Drive current Life remains As single LED failure.
(lots of short reduce by the number to other LEDs unchanged or
series circuits in of LEDs in the failed in the same improves
parallel) parallel ladder section. rung goes to slightly (reduced
zero and in power in
other rungs luminaire will
remains mean the LEDs
unchanged run slightly
cooler)
Open circuit Ladder Light output reduces Drive current Small increase More tolerant to individual LED
(lots of short slightly. One short re-routed via in drive current failures than with open circuit
Maintenance factor determination of LED luminaires
series circuits in series chain of LEDs working LEDs to parallel LEDs in series parallel. Single
parallel) goes out but parallel in parallel chains increases LED failures can occur
chains increase output chains Tj, increases separately. Drive current
to compensate. depreciation and redistribution across many
Increase Tj will reduce reduces life but parallel chains has less impact
light output. only slightly. on individual LEDs but as
number of individual LED
failures increases the
consequences increase leading
to sequential failure of LEDs
and reduced useful life. More
tolerant to individual LED
failures than with open circuit
LEDs in series parallel.
Guidance Note 11/20