| Abstract | Array-based spectrometers, with their compact size, low weight, low cost, and fast measurement time, are now frequently used in place of both conventional single-channel scanning monochromators, and broadband meters. Their rapid measurement capability makes them an attractive option for routine solar UV spectral measurements, where shortterm variability in signal is a challenge. However, compactness, portability, low cost and high speed are achieved at the expense of the spectrometer's optical and electronic performance. Thus such spectrometers are more prone to measurement error from environmental changes, and more prone to other intrinsic sources of error such as stray light and detector non-linearity, which significantly affect solar UV measurements, than a scanning monochromator. The effects of stray light and non-linearity can be reduced either by improved optical and detector design or by a detailed spectrometer characterization. We present in this paper our investigation of the performance of three different commercial array spectrometers: two mini-spectrometers, and a more elaborate array spectrometer with an on-board image amplifier device. These were tested for a subset of performance parameters: their wavelength accuracy and stability, electronic linearity, responsivity linearity, stray light sensitivity, and mechanical stability and repeatability. With all three spectrometers we found that these parameters, particularly but not limited to stray light, had a significant impact on the measurement of the incoming optical radiation. This meant that, without characterization, the instruments would be unable to accurately measure the UV component of any source with significant visible radiation. We discuss various simple and low-cost solutions for improving the performance of these instruments, and providing a rigorous calibration using a straightforward set-up including optical filters and the quasi-monochromatic light from a double monochromator. © 2013 AIP Publishing LLC. |
