LED spectrum optimal modelization

LED lamps market share is continuously growing since a few years, replacing traditional products due to their higher efficiency and expected lifetime. Yet some aspects such as the dependence between photometric and colorimetric parameters and the applied electrical power still require additional investigations and modeling. In lighting, every parameter is defined as some weighting function applied to the radiometric spectrum. A detailed knowledge of the spectrum sensitivity is thus the key to predict derived parameters. After measuring LED spectrum at different command levels, it was found that the whole set of curves can be predicted from a very small number of generating functions highly correlated with the current level. A new approach is thus proposed in order to predict the light spectrum from a low-order polynomial with the current as the only input.