Makeless Exposure System Based On DMD Key Technology Research

Exposure is the key technology in the PCB board making process for integrated circuit industry.With the miniaturization of electronic products and the high integration of integrated circuit,more stringent requirements for the exposure technology are put forward.The traditional mask exposure technology has been unable to meet the production needs of the exposure industry,because the line width is more and more thin and the line spacing is getting smaller and smaller.In order to solve this problem,the maskless exposure technology based on DMD is proposed.There are several hundred thousand micro mirrors in the DMD and each micro mirror corresponds to a pixel.Therefore,it can meet the demands of meticulous exposure.The maskless exposure technology saves the mask plates and improves the production efficiency.It is a great innovation and change in the exposure industry.At the time of maskless exposure,focusing needs to use CCD and focal exposure is by judging the clarity of the CCD acquisition images.The positioning holes should be collected by CCD in order to achieve the distortion correction of exposure image in a certain range.The line width and distance of the exposure image should be detected.Finally,clear image is collected by focusing of CCD.Thus,in the case of maskless exposure,how to achieve auto focus is a key point to be solved about CCD.During the development period about exposure machine platform,considering how to implement exposure platform design is more simple and the cost of the late realization is less,an innovative solution is another key point.According to the two points,this paper has done the following research.1.The definition evaluation algorithm based on the spatial and frequency domain is not the ideal anti-noise performance and stability.Therefore,the definition evaluation algorithm of wavelet packet transform based on human visual is proposed.It's compared with the definition evaluation algorithm based on the spatial and frequency domain for some individuals.The results are compared from three aspects: the width of the steep area,the ratio of sharpness and the degree of steepness.The resolution ratio is raised by 45%,and the steepness is raised by 26%.It can be clearly seen that the algorithm proposed in this thesis has better effect and stronger noise immunity.2.The traditional focusing evaluation window lack of adaptability and easy to introduce interference.Therefore,an adaptive window extraction method based on the combination of single scale and multi-scale is proposed.Through testing the different scene images,it is evident that the focusing evaluation window based on the new method is adaptive.It not only can automatically select the main target,but also can reduce the interference with other factors.The definition evaluation window has a pronounced effect on definition evaluation.3.The traditional climbing search algorithm is so easy to get trapped into local optimal it can't find the focal plane.Therefore,the improved climbing search algorithm is put forward.The proposed algorithm can greatly avoid the influence of local optimal.4.In the development of exposure platform using DLP6500 equipment,the exposure scheme based on the video stream is proposed in order to solve the problems of insufficient memory,complex operation and high cost.The video streaming solution is tested by hardware and software.Then,the result shows that it can be used for actual exposure industry.