How is a light source’s spectrum measured?

May 1, 2019

With all the buzz about how the output spectrum of LED lighting can be varied for applications like street lighting, human-centric lighting, and horticultural lighting, the question occurs: how is the spectrum determined anyway? With a nifty little device called a spectroradiometer.

The typical test set up looks like this:

The light source is mounted in an integrating sphere, which is basically a hollow ball, white on the inside, with a port for the fiber optic cable that runs to the spectroradiometer. The sphere diameter needs to be large enough that the test source does not appreciably absorb light that is reflected off the interior wall, which is why integrating spheres come in lots of sizes (as a point of interest, a test lab I worked with several years ago built a sphere out of the tops of two farm silos, connected by a hinge at one end and painted white on the inside).

Spectroradiometers measure the spectral radiance of a light source – that is, the emitted radiation flux. When the source is turned on inside the sphere, the light is reflected off the sphere’s white surface many, many times, so that regardless of the form factor of the source, the measured result is ideally the total or “integrated” amount of light emitted. The light that hits the detector port is routed to the spectroradiometer, where the characteristics of the light are determined. Because the light output is integrated, any information about directionality is lost. To determine the distribution of the emitted light, a separate test using an instrument called a goniometer is required, but that’s a discussion for another day.

A spectroradiometer consists of four main components:

• input optics
• a monochromator that separates the light into its spectral components
• detectors that receive the spectrally separated light and convert the light flux into an electrical signal
• embedded intelligence for calculating the various characteristics of the light emitted by the test source

Light from the source passes through a slit (1) placed at the focal point of a curved mirror (2) such that the light reflected from the mirror is collimated. The collimated beam makes its way to the grating (3) where it is diffracted; the directions of the reflected beams depend on the groove frequency of the grating. This diffracted light then reflects off another mirror (4), which refocuses each narrow bandwidth segment onto one or more exit slits (5). The light is now split into its narrow-bandwidth (typically 1 nm) components.

Figure 2  A monochromator separates light into its spectral components.

The diffracted light is then projected onto a detector for conversion to digital signals that will be used to generate the spectral power distribution (SPD). There are two detection configurations: The first consists of a fixed diffraction grating and a CCD detector array. This configuration has the advantage of being able to simultaneously capture the entire spectrum of the source, achieving millisecond measurement times, and is most commonly used. The other configuration consists of a single detector with a rotating grating, which results in much longer measurement times, but can also provide a wider wavelength range when several grating/detector pairs are incorporated into one instrument.

Spectroradiometers can be used to determine the SPD of any light source and, indeed, have been used for years to characterize conventional sources (e.g., incandescent, fluorescent). Figure 3 shows typical relative SPDs for four different sources: halogen, fluorescent, an RGB LED (which creates white light by combining red, green, and blue), and a PC LED (that creates white light by passing blue light through a phosphor coating). Notice the three peaks – corresponding to red, green, and blue – for the RGB LED, and the two peaks at blue and orange for the PC LED.

Figure 3  Relative spectral power distribution of light sources (Source: US Dept. of Energy Solid State Lighting Technology Fact Sheet, April 2016)

In addition to determining SPD and total lumen output, spectroradiometers use algorithms to calculate the chromaticity coordinates, correlated color temperature (CCT), and color rendering index (CRI) of white light sources, but that’s also a discussion for another day.

Stray light caused by imperfections in the diffraction grating may be a significant contributor to measurement errors. Varying environmental conditions in the test lab itself, such as fluctuating temperature, humidity, air flow, and even AC & RF interference, can also contribute to measurement errors.

There are several standards that provide the test method for LED light sources, including spectroradiometer performance requirements. Most commonly used is: Illuminating Engineering Society LM-79, Approved Method: Electrical and Photometric Measurements of Solid State Lighting.

Original Source