精密定位編碼器之光學感測器設計與驗證

Abstract

隨著工業精密儀器科技的發展以及近年來物聯網的興起，伺服系統之應用越趨重要，本研究針對可應用於伺服系統中之編碼器進行光學感測器的研發。首先，利用輻射幾何理論探討編碼器架構之感光訊號變化，並建立感測器感光數學模型與使用光學模擬軟體ASAP (Advanced System Analysis Program)分析感測器感光訊號，依據輻射幾何與光學理論提出新創之兩種形式光學感測器(反射式及穿透式)設計並實現，最後利用實驗比對完成之感測器其感光訊號與輻射幾何理論及ASAP光學模擬之精確性。 穿透式感測器以光學遮罩產生弦波輸出為主要設計，依光學遮罩設計可分為矩形、鋸齒形、鑽石形以及水滴形，在模擬階段以RMSE (root mean square error)誤差分析方法比較設計感測器之感光訊號與理想弦波之差異，在數學模擬中與理想弦波之RMSE結果依上列設計排序分別為0.1507、0.0138、0.0082以及0.0069，光學模擬階段之RMSE依上列設計排序為0.0834、0.0188、0.0239以及0.0133，最後將實現之感測器應用於光學編碼器之架構中，其輸出感光訊號與理想弦波之RMSE結果依序為0.110、0.805、0.658以及0.641。 反射式感測器之設計目的為搭配反射隔柵(grating)下能判斷平面移動距離。設計之反射式感測器由多組感光二極體組成，依感光區之感光二極體排列設計不同共有三種新創反射式感測器，可分為矩陣形、同心圓形以及輻射形。對應反射格柵之設計(4 µm光反射面/6 µm光吸收面)，感測器在光學模擬階段可得每10 µm一周期之訊號，實現之感測器應用於搭載反射格柵之線性移動平台進行實驗，實驗階段可得每10 µm一周期之訊號，並且透過同感測器內不同組感光訊號之相位差可達到移動方向之判斷。 本文利用輻射幾何以及光學理論進行光學感測器之研發，雖實際應用上受安裝技術及其他實驗部件影響，但根據輻射及光學理論以模擬感光訊號方式探討並調整感測器設計與感測器之實驗結果仍有不錯準確度。

With the development of industrial precision instrument and the rise of Internet of Things in recent years, the application of servo system becomes more important. This study is aimed at research and development of the optical sensors which can be applied into positioning encoder for servo system. First, the sensitometric signals of optical sensor in encoder are discussed with theory of radiometry. Then, the sensitometric signals are analyzed through mathematical simulation and optical simulation. Based on theory of radiometry and optics, we design and implement reflective type and transmissive type of optical sensors which can be applied into linear encoder and rotary encoder respectively. Finally, the performances of sensitometric signals from the implemented sensor are verified with experiment. The main design of the transmissive sensors is generating sinusoidal sensitometric signals with an optical mask. According to the patterns of design for optical mask, the transmissive sensors can be subdivided into four categories and named as trapezoid shape, jagged shape, diamond shape and drop shape. The difference between sensitometric signals from the designed sensors and ideal sinusoidal signal is analyzed by the analytical method of root-mean-square error (RMSE) in every phase of simulation. In the phase of mathematical simulation, the results which is sorted by the above categories of comparison with ideal sinusoidal signal are 0.1507, 0.0138, 0.0082, and 0.0069 in RMSE. Further, in the phase of optical simulation, the results as the same sort of comparison with ideal sinusoidal signal are 0.0834, 0.0188, 0.0239, and 0.0133. Finally, the designed transmissive sensors are implemented and applied into the architecture of rotary optical encoder. The RMSEs sorted by the above between the actual sensitometric signals generating from implemented sensors and ideal sinusoidal signal are 0.11, 0.805, 0.658, and 0.0641 in the experiment. The main objective of the designed reflective sensors is that displacement can be measured by the sensitometric signals of designed sensors with the reflective grating. The designed reflective sensors are composed of multiple photosensitive diodes. In addition, the designed reflective sensors can be subdivided into three categories and named as matrix shape, concentric circular shape, and radial shape according to the arrangement of the photosensitive diodes in sensor. By collocating with the reflective grating which is composed of 4 µm reflecting surfaces and 6 µm non-reflecting surfaces, the sensitometric signals with 10 µm cycle can be generating from the designed sensor in the phase of optical simulation. Finally, the designed sensors are implemented and applied into a linear mobile platform with the reflective grating. In experiment, the actual sensitometric signals with 10 µm generating from the implemented sensors can be verified. Furthermore, the moving direction of mobile platform can be distinguished through the difference of phases between the two sensitometric signals in the sensor. In this study, with the theory of radiometry and optics, the research and development of optical sensors for encoder can be executed. Although the performances of the designed sensors are influenced by the technique of installation and other components in experiment, the methods of design and results of simulation in stage of design are fairly accurate.