LED light sources: a survey of quality-affecting factors and methods for their assessment

Abstract

Nowadays, after nearly half a century of rapid and sustained development of light-emitting diodes (LEDs), conventional light sources such as incandescent and fluorescent lamps are beginning to be replaced by LEDs. Therefore, understanding and assessing all the relevant factors that affect the quality of LEDs is becoming increasingly important for design, production and maintenance of various LED products. The most adverse quality-affecting factors are the initial variability of the optical and electrical properties in a batch of LEDs, temperature and electrical dependence and temporal degradation with corresponding variability. In this paper, we survey the most important quality-affecting factors and corresponding methods for their assessment. First, initial variability of the optical and electrical properties in a new batch of LEDs and the corresponding assessment methods are outlined. Next, the temperature stability of optical and electrical properties of LEDs is discussed. Moreover, the most frequently studied methods for spectral degradation and corresponding degradation variability prediction are reviewed according to the accuracy, applicability and specificity. Finally, the advantages and disadvantages of the established models of part stress analysis are pointed out. In this way, all the major factors affecting the quality of LED products are summarized and the corresponding methods for their assessment are outlined.

Far-field condition for light-emitting diode arrays

“…A common system error is LED source variability. The optical properties of different LEDs such as color, radiation pattern, and radiant (or luminous) flux may vary due to thermal variations and errors in the manufacturing process [12][13][14]. The change of performance is around the average values given in the technical data sheets.…”

“…10(a) we show the variability of the far-field condition for directional LEDs with β 1=2 ¼ 10°. Recent experimental measurements indicate that the LED flux variation is approximately normal [12,13]. Therefore we use the normal distribution as the probability function to generate random numbers for weighting the LED intensity with w i < 1, where IðθÞ i → w i IðθÞ i is the weighted angular intensity distribution of the ith LED.…”

“…Therefore we use the normal distribution as the probability function to generate random numbers for weighting the LED intensity with w i < 1, where IðθÞ i → w i IðθÞ i is the weighted angular intensity distribution of the ith LED. Considering that the flux may vary 50% in this type of LEDs, we use a variance of σ ¼ 0:25 [12,13]. As can be noted for the 2 × 2 array, the far-field distance is more sensible in lowdensity arrays than for high-density clusters.…”

A review on light-emitting diode based automotive headlamps

“…Arrhenius based LED lumen maintenance model [31] and bathtub curve probability distribution [87] based catastrophic failure model are the most popular mathematical models for projecting the LED lifetime and catastrophic failure rate under different factors. For instance, the probability of a catastrophic Luxeonrebel LED failure, impacted by the junction temperature of 100 ppm (or reducing 0.001%), is estimated to 4000 h [43].…”

Multiple-chip package embedded on compound board for light emitting diode

“…However, to our knowledge, all their efforts focus so far on the single chip. For gaining greater market penetration of high-power LED lighting it is favourable to use the COB based multiple-chip LED array with high opticial-thermal performance as well as low cost [1,6]. This Letter proposes that the multiple-chip COB package structure can be obtained by making a compound platform with special copper reflectors embedded in the Al-PCB.…”