IESNA LM80 and TM21

Mark Hodapp
Philips Lumileds
Senior Application Engineer
370 West Trimble Road
San Jose, CA 95131
T 408 964-2648
F 408 964-5354
mark.hodapp@philips.com
www.lumileds.com
2011 Southeast Region Workshop Tampa, Florida

What is lumen maintenance?
Well understood that LEDs are very long life light sources.
However, the light output of SSL products gradually goes down over time.
Rated lumen maintenance life of the LED (L
p
) is the elapsed operating time over
which the LED light source will maintain the percentage, p, of its initial light output
L
70
(hours) = time to 70% lumen maintenance
L
50
(hours) = time to 50% lumen maintenance
2011 Southeast Region Workshop Tampa, Florida










Lumen maintenance and EPA
EPA is using a simple exponential model for lumen maintenance until TM-21 is
published:
| (time ) = exp( alpha time )
ln (0.7)
alpha =
time
ln (0.7)
alpha (25,000 hour ) = = 1.4267 e 5
25,000
EPA limit LM (25,000) = exp| (1.467 e 5)(6,000)| = 0.918
2011 Southeast Region Workshop Tampa, Florida






Lumen maintenance and EPA
Lumen Maintenance Projection for White >3500K LXML-PWx1
LUXEON Rebel under these conditions
85C, 0.35A (Tjunction ~ 98C) Normalized to 1 at 0 hours
N
o
r
m
a
l
i
z
e
d

L
i
g
h
t

O
u
t
p
u
t

1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
Ta within 5C of Ts, in accordance with LM80.
ENERGY STAR 25,000 and 35,000
limits after 6,000 hours of stress
35,000 hour
25,000 hour
EPA: Simple exponential extrapolations
of form:
Flux (time) = exp [ -alpha time]
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000
Hours
2011 Southeast Region Workshop Tampa, Florida






Lumen maintenance and EPA
Lumen Maintenance Projection for White >3500K LXML-PWx1
LUXEON Rebel under these conditions
85C, 0.35A (Tjunction ~ 98C) Normalized to 1 at 0 hours
1,000 10,000 100,000 1,000,000
Hours
2011 Southeast Region Workshop Tampa, Florida
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
N
o
r
m
a
l
i
z
e
d

L
i
g
h
t

O
u
t
p
u
t

Ta within 5C of Ts, in accordance with LM80.
ENERGY STAR 25,000 and 35,000
limits after 6,000 hours of stress
35,000 hour
25,000 hour
*Philips Lumileds recommends
extrapolations of s 6x the test
time
6x test time*
L70 = 162,000 hours
TM-21 protocol:
L70 > 36,000 hours
EPA: Simple exponential extrapolations
of form:
Flux (time) = exp [ -alpha time]

LM-80 Scope
Lumen maintenance test method written by
IESNA (Illuminating Engineering Society of
North America)
LED package, array or module driven by
auxiliary driver
LEDs are driven with external current sources
during operation and lumen maintenance
testing
LED Case temperature is controlled during
operation
During lumen maintenance testing, LED is
allowed to cool to room temperature and tested
at air temperature of 25C
2011 Southeast Region Workshop Tampa, Florida

LM-80 Test Method
LM-80:
IES Approved Method
Specifies
Operation at three case temperatures (55C, 85C and one
selected by manufacturer)
Air Temperature to within +/- 5C, Case Temperature to within
+/- 2C
RH less than 65%
Minimum 6,000 hours, data collected every 1,000 hours
Data collection at 25C
Constant current, rated voltage
Record Lumen Maintenance, Chromaticity, Catastrophic Failures
Reporting Format
2011 Southeast Region Workshop Tampa, Florida

What LM-80 Does Not Specify
LM-80 does NOT specify:
Pass/Fail Criteria
Graphing of results
Curve fitting methods
Extrapolation and L70 prediction methods
Sample Size
How many drive currents
What changes to an LED package require new testing
2011 Southeast Region Workshop Tampa, Florida

EPA ENERGY STAR

Specific
Lumen Maintenance Requirements
Test According to LM-80, by NVLAP accredited lab
Min 25 samples per combination of Temperature and Current
LM > 91.8% at 6,000 hours for residential indoor
LM > 94.1 % at 6,000 hours for non-residential and residential outdoor
Ts point temperature below LM-80 tested temperature
Drive current below LM-80 tested current
2011 Southeast Region Workshop Tampa, Florida

EPA ENERGY STAR

Change in Reporting Requirements
NO CHANGE TO IES REPORT FORMAT. However, this format is not
accepted by EPA
Until 12/31/2010 use IESNA LM-80 prescribed report format
Starting 1/1/2011 new report format
Starting 1/1/2011 devices tested must be of equal or lower CCT than
products being used
2011 Southeast Region Workshop Tampa, Florida

TM-21 What it is, and why it is important
IESNA TM-21 (Technical Memorandum) specifies how to extrapolate the
LM-80-08 lumen maintenance data to times beyond the LM-80 test time.
For example: EPA ENERGY STAR Manufacturers Guide requires L70 of
25,000 or 35,000 hours.
TM-21 is important because it is referenced in the EPA ENERGY STAR
Guides, and customers generally are concerned about getting ENERGY
STAR approvals for their products.
It is also important as it creates a common playing field for LED competitors
to specify lumen maintenance behavior for their white LED products
intended for illumination applications.
2011 Southeast Region Workshop Tampa, Florida

TM-21 status
TM-21 is being written by a panel of industry experts:
6 LED manufacturers (Philips Lumileds, Osram, Nichia, Illumitex,
GE, and Cree).
2 US Government Labs (PNNL, NIST)
Status: Agreement has been reached by panel on basic mathematical
algorithm. First draft is expected to be completed by mid February
2011. IESNA will send draft for Ballot by end of February 2011. After
approval, TM-21 will become a new IESNA document.
2011 Southeast Region Workshop Tampa, Florida

TM-21 Mathematical Algorithm
Manufacturer should normalize the light output to 1 at 0 hours for each unit in the data set.
Data for all units are averaged together (Result represents average behavior).
All average lumen maintenance data points from 0 and less than 1,000 hours are removed.
For 6,000 hour data: average lumen maintenance data points from 1,000 hours to 6,000
hours are fit to a simple exponential extrapolation model using a least-squares curve fit.
For data s 10,000 hours: The last 5,000 hours of data should be used for the lumen
maintenance extrapolation.
For data >10,000 hours. The data points from the last 50% of the total measurement time
should be used. In the case where the last 50% of the total measurement time is not an
integer multiple of 1,000 hours, then the interval shall be rounded-up to the next 1,000
hours should be used. [e.g.: after 11,000 hours of measurement time, the last 6,000 hours
of data should be used].
The L70 extrapolation shall be the lower of: The resulting L70 time OR 6X the LM-80 test
time. [e.g. with 6,000 hours of LM-80 test data, then L70 s 36,000 hours. With 10,000
hours of LM-80 test data then L70 s 100,000 hours.
2011 Southeast Region Workshop Tampa, Florida






TM-21 Mathematical Algorithm
u(t ) = B exp| o t |
TM-21 Example LL
| B |
ln |
\
0.7
.
L70 =
o
Major changes:
B term added to simple exponential
model
Least-squares fit of data starting at
1,000 hours (6Khr data set)
Least-squares fit of data using last
5,000 hours (7K-10K data set)
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000
6X extrapolation time limit
Hours
0.6
0.7
0.8
0.9
1.0
1.1
N
o
r
m
a
l
i
z
e
d

L
i
g
h
t

O
u
t
p
u
t
LL ave
1K-6Khr extrap
Points Used
2011 Southeast Region Workshop Tampa, Florida

Lumen maintenance is not the whole story
LM-80 and TM-21 provide an industry standard for lumen maintenance
testing and for lumen maintenance extrapolations for LED components. LM
79 is the lumen maintenance standard for SSL systems.
Next Generation Industry Alliance and US Department of Energy LED
Luminaire Lifetime: Recommendations for Testing and Reporting, May
2010, discusses practical lifetime considerations for SSL luminaires.
Lumen maintenance is not the dominant failure mode for most SSL
products.
2011 Southeast Region Workshop Tampa, Florida

System Reliability LED Reliability
R
system
= R
electrical
* R
connections
* R
LEDs
* R
optical
* R
thermal
* R
mechanical
Philips Lumileds has developed reliability and lumen maintenance models in order to predict
long-term reliability performance of LUXEON Rebel products. See WP15 Evaluating the
Lifetime Behavior of LED systems for more information.
2011 Southeast Region Workshop Tampa, Florida