Luminous Flux Calibration of LEDs at NIST - cie

Abstract for the 2nd CIE Expert Symposium on LED Measurement
Luminous Flux Calibration of LEDs at NIST
C. Cameron Miller and Yoshi Ohno
National Institute for Standards and Technology
Gaithersburg, MD 20899 USA
The luminous flux (unit, lumen) is one of the most important characteristics of Light
Emitting Diodes (LEDs) along with the luminous intensity, and is commonly measured
using integrating sphere photometers. However, large variations of measurement
results are reported among different manufacturers and users of LEDs, which is causing
difficulties in trade. CIE 127 [1] gave a good solution to the problems in luminous
intensity measurements, but does not give sufficient treatment for the problems in
luminous flux measurement. There are issues such as the LED mounting geometry (4π,
2π), treatment of backward emission, the best integrating sphere geometries, etc., which
are being addressed in CIE Technical Committee 2-45. While these issues are to be
resolved, there is an urgent need for standard LEDs in order to certify measurement
accuracy in industrial laboratories. To address such needs, NIST has developed
measurement procedures to calibrate total luminous flux of LEDs using the existing NIST
2.5 m integrating sphere [2].
Figure 1 shows the arrangement of the NIST integrating sphere for LED measurement.
This arrangement is for integration over 4p – the normal definition of total luminous flux –
therefore including the backward emission. The LED is mounted on the center of the
sphere, aiming at the azimuth angle q = 130°, illuminating the portion of the sphere wall
where the spatial nonuniformity of the sphere response is minimum. The LED is
mounted using a special LED mount, as shown in Fig. 2, to minimize the near-field
absorption around the test LED. The baffle (normally 20 cm) in front of the photometer
head is replaced by one with a 10 cm diameter in order to reduce the area of shadowed
region and thus minimize the spatial nonuniformity errors [3]. Other sphere
LED
Baffle 1
(10 cm)
Baffle 2
(15 cm)
( )=(90°,0°)
Photometer
head
External source (FEL lamp)
laser
sphere joints
Aperture &
photometer
wheel
Figure 1. NIST integrating sphere system arranged for calibration of LED total luminous flux.

Abstract for the 2nd CIE Expert Symposium on LED Measurement
arrangements remain the same as the normal
configuration for calibration of lamps. The test
LED is measured using our detector-based
method [2] in comparison with the beam flux
introduced from the external source, which is
approximately 2 lm. In spite of the size of the
sphere, we have enough signals for typical
high-intensity LEDs having luminous flux of 0.1
lm to 1 lm, in part to the high reflectance (~98
%) of the coating. The photometer head is a
temperature-monitored type with a built-in
amplifier having gain ranges up to 10 9 V/A.
The spectral mismatch correction is applied by
measuring the spectral power distribution of
each test LED and the total spectral Figure 2. The LED mount in the sphere.
responsivity of the sphere system. The
measurement uncertainty arising from the spatial nonuniformity of the sphere response
has been analyzed for this particular sphere geometry using the measured spatial
response distribution function (SRDF) and an LED model with a variable beam angles
(Fig.3). Figure 4 shows the calculated correction factors with respect to an isotropic
point source, as a function of the LED beam angle. The measurement errors, if not
corrected, are shown to be within 0.15 % for beam angles from 20° to 120° (Lambertian),
which is taken into account in the uncertainty budget. The calibration service for
luminous flux of limited types of LEDs is now available at NIST. The expanded
uncertainty (k=2) of the calibration is from 0.6 % to 2 % depending on the color and
other characteristics of the reference LEDs. A dedicated integrating sphere system for
LED measurements and standard LEDs for luminous flux are to be developed.
References
[1] CIE Publication No. 127, Measurement of LEDs (1997).
[2] Y. Ohno and Y. Zong, "Detector-Based Integrating Sphere Photometry,"
Proceedings, 24 th Session of the CIE Vol. 1, Part 1, 155-160 (1999).
[3] Y. Ohno and R. O. Daubach, “Integrating Sphere Simulation on Spatial
Nonuniformity Errors in Luminous Flux Measurement”, J. IES, Winter 2001 issue (to
be published).
325 330335340345350355
20°
320
0.8
315 40°
310
305 60° 0.6
300
80°
295
0.4
290 100°
285
280
260
255
250
245
240
235
120°
1
0.2
230
225
220
215
210
205
200
195190185
0
0
180
5 10 15 20 25
3035
40
45
50
130
135
140
145
150
155
160
165
175170
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
Figure 3. LED model of varied beam angle.
Correction factor
1.0015
1.0010
1.0005
1.0000
0.9995
0.9990
0.9985
Series1
0 30 60 90 120 150 180
LED beam angle (degree) FWHM
Figure 4. Correction factor for the spatial nonuniformity of
the sphere, for an LED model of varied beam angle.