Light size: Integrating spheres are hollow spheres with an inside diffusely reflecting surface. They are used as measuring tools for measuring the luminous flux of light sources and luminaires. For LED measurements according to the CIE-S 025/2015 standard, specific size requirements are defined. For example, DUT (Device Under Test) devices installed in the center of the Ulbricht sphere (named after the German engineer Richard Ulbricht) – this is referred to as a 4π measurement geometry – can only have a total area of no more than two percent of the inside surface of the sphere. This corresponds to a cube-shaped DUT with a side length of 1/10 of the inside diameter of the ball. In a 2π geometry, that is to say when the DUT is attached to the ball opening, the diameter of the measuring opening may not be more than 1/3 of the inside diameter of the ball.
In short, with a 4π measurement geometry and an integrating sphere of 1 m diameter, the DUT may have a maximum side length of 10 cm. Consequently, only small light sources can be measured in a comparatively large integrating sphere. Large lamps and luminaires require an even larger integrating sphere, in some cases more than 3 m in diameter, or the use of goniometers to measure CIE-compliant. Both variants are expensive and space-intensive and therefore unprofitable for many companies. So what happens if the “golden rule” regarding the ball diameter is not respected?
Expected measurement uncertainty
If the status of an accredited laboratory is not being sought, in practice companies check their guidelines and specifications. So it happens that luminaires are measured, which account for up to 30 percent of the ball diameter. The expected measurement uncertainty in this procedure is then between three to four percent. With comparatively small spheres and large measuring objects, this measurement error increases because the DUT restricts the ball reflection and thus the measuring accuracy.
The tests listed below show a low uncertainty of measurement if the recommended ball and DUT size deviates from CIE.
Several DUTs were prepared for the measurement series, which represented different luminaire sizes. For all DUTs, the same LED was mounted on a variety of foam-molded chassis imitations. The spectroradiometer GL Spectis 1.0 with a spectral range from 340 to 780 nm and an integrating sphere GL Opti Sphere 1100 with a diameter of 1100 mm were used. The complete measuring system was initially calibrated spectrally. Relative measurements were made for the series of measurements. The result of the analysis of the smallest DUT that is the single LED without housing was taken as a reference value. The picture shows the different DUT configurations.
The measurements for each configuration were repeated several times. The results could be reproduced. The reference DUT was a white LED with a power dissipation of 5.6 W at I = 0.6 A.
Results were better than expected
The results were better than expected. Although the specifications described in the CIE standard were exceeded many times over, the deviations were only about two percent.
The DUTs in sizes 15 × 25.15 × 55.15 × 67.15 × 67.15 × 80 and 50 × 67 cm 2 were made of black foam, while around DUT was made of a light cardboard box, which contained most of the Ball diameter filled. This produced a smaller Fluxmessfehler than the smaller black foam DUTs. The table shows the measurement results.
The measurements showed that, although differently sized luminaires correlate with luminous flux and color values, the differences are very small.
As the size of the DUT increases so does the luminous flux value, which necessitated significant adjustments. After these recalculations, the measurement results could be reproduced. For example, integrating sphere measuring systems with a dispersion layer index of more than 97 percent provide dozens of reflections in the sphere despite relatively large DUTs.
It should be noted that the self-absorption coefficient of the measuring system for each wavelength is defined so that its sum determines the luminous flux absorbed by the DUT. Depending on the size and color of the measurement object, the coefficient may vary throughout the measurement range, requiring a precise spectrometer to obtain genuinely accurate measurements.
Further conclusions regarding the measuring principle defined in the standard require additional examinations and comparisons of different measuring systems. The results show, however, that using a suitable and accurate measuring system, repeatable and meaningful measurement results are obtained from light sources that have significantly larger dimensions compared to the integrating sphere used than the standard requires.