Yi-Chin FANG, Yih-Fong TZENG^*, and Si-Xian LI

Department of Mechanical and Automation Engineering, National Kaohsiung First University of Science and Technology, 2 Jhuoyue Road, Nanzih District, Kaohsiung 811, Taiwan, R.O.C.

(Received October 29, 2007; Accepted July 1, 2008)

This research proposes a new extended optimization method for a miniature light emitting diode (LED) pocket-sized projection display, introducing integration of the Taguchi method and principal component analysis in order to optimize the multiple quality characteristics of an LED pocket-sized projection display. With the aid of interactive optimization, control factors with three different levels are carefully selected in the complicated preliminary experiments. A set of optimal design parameters is well selected for best results on the combined effects of the total luminous flux, illumination uniformity, and the packing size of the system. The selected control factors are inclusive of major lens and system specifications, such as lens overall length, X-CUB semi-aperture, length of light integrator, width of integrator, total internal reflection (TIR) prism entering semi-diameter for the TIR prism, air-gap of the TIR prism, and digital micromirror device (DMD) position; an L18 orthogonal array is applied and implemented in the experiments. According to experimental results, the optimal design parameters for the projection display can be determined as A1 (lens specifications: type I), B3 (lens length: overall length), C1 (X-CUB semi-aperture: 8 mm), D3 (integrator length: 36.6 mm), E2 (integrator width: 3.5 mm), F2 (TIR prism entering semi-diameter: 11 mm), G1 (TIR prism air-gap: 1.0024 mm), and H1 (DMD location: -0.5 mm). In addition, analysis of variance (ANOVA) is also employed to identify the factor A (lens specifications), factor D (integrator length), factor F (TIR prism entering semi-diameter), and factor G (TIR prism air-gap) as key parameters, which account for 71.82% of the total variance. The other factors when compared are found to have relatively weaker impacts on the process design. Furthermore, a confirmation experiment of the optimal design parameters shows that the aforesaid multiple performance characteristics are optimized to achieve the best levels. It is concluded that Taguchi method and principal component analysis (PCA) combine to optimize and then minimize the LED pocket-sized projection display system, which not only yields a sufficient understanding of the effects of control factors, but also produces an optimized design to ensure that the LED pocket-sized projection display system exhibits the best multiple performance characteristics.

Key words: optical design, optical engine, LED light source, Taguchi method, principal component analysis, LED pocket-sized projection display, optimization, analysis of variance (ANOVA), miniature projector, interactive optimization

^*E-mail address: franktzeng@ccms.nkfust.edu.tw