Dhidik Prastiyanto, Sutikno Madnasri, Esa Apriaskar, Nur A. Salim, Oky P. Pamungkas, Arindi Imanindi, Wahyu Caesandra, Tegoeh Tjahjowidodo
The purpose of this work is to assess mathematical models of a digital spin coater and find the best one to represent its rotational speed behavior. Simulation models were established using a single input-output identification system approach and involved the use of multilevel periodic perturbation signals (MLPPS). The input data was taken from a Pulse-Width-Modulation (PWM) signal for the actuator in the form of MLPPS, while the observed output was from the rotational speed of the spin coater. The prospective models were represented in state-space and transfer function, both in discrete and continuous time domain. Through this study, it was found that the fitness percentage of the models obtained with the utilized approach ranged from 72% to 92% after being validated with the output of the real system. The results also indicate that for the given operating points, candidate model with discrete transfer function TFD3 has the lowest mean squared error (MSE) in average. The findings of this research may serve as a beneficial knowledge prior to controller design of the digital spin coater. Better model may lead to better controller performance that can support to perform uniformity on the film thickness. © 2022 Lavoisier. All rights reserved.
Department of Electrical Engineering, Universitas Negeri Semarang, Semarang, 50229, Indonesia; Department of Physics, Universitas Negeri Semarang, Semarang, 50229, Indonesia; Faculty of Integrated Technologies, Universiti Brunei Darussalam, Gedong, BE1410, Brunei Darussalam; Mechanical Department, Katholieke Universiteit Leuven, Sint-Katelijne-Waver, 2860, Belgium
