Crosstalk Analysis and Scanning Uniformity Improvement of Electrostatic MEMS Scanner

Abstract

MEMS scanners have been widely used in various imaging systems, such as light detection and ranging (LiDAR) and optical coherence tomography (OCT) and display owing to advantages such as compact size, low power consumption, high speed, and cost-effectiveness over conventional Galvano mirror. However, scanning uniformity is degraded by undesired oscillation induced by the harmonics of the input signal and mechanical or electrical coupling. This thesis proposes methods to improve scanning uniformity and analyze crosstalk for MEMS scanners. Input shaping was performed based on an experimental transfer function to effectively obtain the desired scan output with high uniformity for an electrostatic gimbaled MEMS scanner driven in a quasi-static mode. The improvement method features possible driving extended to a higher frequency, whereas the conventional control needs dynamic modeling and is still ineffective in mitigating harmonics, sub-resonances, and/or higher modes. The performance of the input shaping was experimentally evaluated in terms of the usable scan range (USR), and its application limits were examined with respect to the optical scan angle (up to ±4°) and frequency (up to 300 Hz). The proposed method was experimentally confirmed to show good performance because of its simplicity and operable range. For the crosstalk analysis, the two-axis gimbal-less MEMS scanner was characterized to estimate the mechanical crosstalk due to imperfect decoupling of hinged linkages and/or fabrication error. Two types of geometrical crosstalk, caused by rotational alignment angle and fan-shaped distortion, were theoretically derived, and total crosstalk was measured at two rotational alignment angles of the MEMS scanner. The mechanical crosstalk in a quasi-static mode was estimated as 0.9% by comparing the theoretical results with experimental ones based on the least squared error. The proposed method was validated for one-axis driving over a wide range of optical scan angle and extended for two-axis driving to examine the feasibility and limitation of the proposed method. The maximum discrepancy of crosstalk between theoretical calculations and experimental results was 0.72% for one-axis driving and 2.22% for two-axis driving, respectively. This result can be used as new guidance to reduce mechanical crosstalk in the design of MEMS scanner. Since the mechanical crosstalk is a general character of a gimbal-less MEMS scanner, the input shaping using estimated mechanical crosstalk was applied to mitigate the crosstalk and/or find shaped input for desired scanning pattern.