Shen-Whan Chen

System Identification of TP Film Evaporation by Using Nearly Equivalent NN Model

This paper presents a technique, called “nearly equivalent neural network (NN) model” in the application of nonlinear system identification. This technique is expected to adequately to catch the behavior of the nonlinear system. To demonstrate the new technique proposed, the evaporation system of TP decoration film was analyzed. The complex relationship between the film’s transmittance and its possible influencing factors was identified. For the comparison, the same simulations were also performed by using the conventional neural network with the standard steepest descent error back-propagation (BP) learning algorithm.

Design of a PA controller for a quad-band cell phone

A controller which was applied in a power amplifier module for a quad-band cell phone was designed and fabricated in this paper. This controller provides the required DC bias voltage of the quad-band amplifier so that the output of the amplifier reaches the specific power level under various battery output. In addition, a decoder circuit was also included. The outputs of the decoder were used to manage switches so that the power amplifier can operate at different frequency bands, DCS or GSM, and modes, transmission or receiving.

A High P1dB GSM RF Switch in Electronic Engineering

In order to meet the high RF power capabilities by the requirements of mobile communication system, two types of high P1dB GaAs PHEMT quad-band RF switch are designed and implemented in this paper. One has adopted three-gate-finger method and the other uses only single gate but with extra gate length. Both switches possess qualities of low insertion loss, high degree of isolation, and low noise. Their output P1dB at the bands of 820 to 920 MHz as well as 1720 to 1920 MHz are measured and compared.

An intelligent data mining technique for the estimations of TP film's chromatic aberration

This paper presents a novel intelligent data mining technique to estimate the optical properties of touch panel (TP) with different layers coating. The neural network (NN) model is developed to be the intelligent tool for the data analyzer. The new computational method based on well-trained NNs weights is used to analyze the influencing factors of TP films chromatic aberration, i.e. L.a.b. values. The influence degree of each input variable to the net output could be determined through the computational results. The useful and important information for TP films chromatic aberration then can be obtained. In this research, the data of TP film with Cr and Cr2O3 coating are studied. All possible influencing factors are collected and considered. An artificial intelligent (Al) estimator using the best influencing factors is expected to be developed so that the optical properties of TP decoration film could be precisely estimated before the evaporation process is taken.

Analysis of acoustic wave property on ZnO/Diamond/Si thin film substrates

Multi-layer piezoelectric thin-film substrates which may have the properties of good temperature characteristic, high reflectivity, high coupling coefficient, high phase velocity, and integratible with semiconductor, have the potential to be used on acoustic devices. The properties of piezoelectric thin-film substrates are determined by the structures of the substrates, piezoelectric material, the metal film thickness and etc. Usually the acoustic wave parameters of the substrates are not readily obtainable from literature. In this paper, the acoustic wave properties of ZnO/Diamond/Silicon substrate was analyzed using finite element method. The data of this paper would be applied to design high frequency RF acoustic devices such as resonators and filters.

A wide-band small-signal model from 2 to 85 GHz

In this paper, a new extraction method is developed for creating a very wide band, from 2 to 85 GHZ, small-signal model for interdigitated, low noise, pseudomorphic high electron mobility transistor (pHEMT). By applying the scalability nature of transistor, a physically related small-signal model is obtained through curve-fitting optimization. This new extraction method also eliminates the need of a de-embedding procedure.

Determination of Coupled-of-Modes Parameters by Applying Neural Network Technique

The coupling-of-modes (COM) theory has been widely used in simulation of surface acoustic wave devices. The accuracy of simulation strongly depends on accurate COM parameters of the piezoelectric substrate. The COM parameters of most single crystal piezoelectric substrates can be obtained in literatures. However, for thin film and multilayer substrate, the COM parameters are often not accessible. In this paper, a method that uses neural network technique to extract COM parameters was proposed. The validity of the method was verified by simulation of a two-port SAW filter with uniform interdigital transducer structure.

A switch IC implemented by four depletion-mode field effect transistors for dual-band GSM switch diplexer module

A new switch design method by applying depletion-mode field effect transistor for dual-band GSM switch diplexer module is presented here. If compared with switch module implemented by using PIN diode, the new module would not only save more currents, consuming only around 0.4 mA in transmitting mode, but also match totally in terms of logic control function and pin-to-pin replacement. If compared with other switch modules also by using depletion-mode field effect transistor, the new module needs only four transistors instead of eight transistors and excluding the use of any voltage supply.