Jiaye Li and Yoshi Ohno, National Institute of Standards and Technology, and Andrew Bierman, Consultant

Abstract

Recently TLM (temporal light modulation) related measurements and calculations have regained high interest, with new regulations by European Commission and other regional regulations requiring limits for TLA (temporal light artefacts) quantities for LED lighting products. In these regulations, P_st^LM (IEC TR 61547-1-2020) for flicker and SVM (TR 63158-2018) for stroboscopic effects are commonly used. However, due to complexity of P_st^LM, attentions are paid to another metric for flicker, M_p (ASSIST vol. 11, 2015). This metric works in frequency domain and requires a much shorter measurement time (>2 sec) compared to 3 minutes for P_st^LM. M_p is already used in some commercial TLM measuring instruments. However, for some types of lamps, large variations in M_p results are found for repeated light measurements using different waveform sampling parameters.

Both simulated waveforms and measured waveforms were used to evaluate M_p. The impact of waveform duration (from 2 sec to 180 sec), sampling rates, and starting phase of the measured waveform on the M_p values were investigated.

The results show that the waveform duration and starting phase have significant impact on the M_p values when using the original formulae in the ASSIST document. The variation of calculated M_p due to these two factors was largely reduced after adding a Hann window before the FFT transform. However, applying a Hann window led to a different M_p value from the one obtained with the original formulae, since the Hann window reduces the effect of spectral leakage while enhancing the FFT spectrum peak due to the added window function. The sampling rate can also have a noticeable effect on the M_p value depending on the waveform. Further investigation into the cause of these and possible solutions is in progress to improve the metric.