Pioneering e-supply chain integration in semiconductor industry: a case study

Taiwan Semiconductor Manufacturing Company (TSMC) is the largest semiconductor foundry in the world. Advanced Semiconductor Engineering Inc. (ASE) is the world’s leader in semiconductor assembly and testing. From 1998 to 2004, the two companies completed electronic integration of 11 key business processes through the Internet. The result is a seamless interface between TSMC, ASE and their joint customers. They can now obtain accurate, timely information on their product status and respond appropriately when needed. While the direct economic benefits are estimated to be around US$ 10 million through productivity increase over a total investment of about US$ 2 million, the indirect benefits of this initiative could be on the order of US $100 million if the joint customers’ benefits are considered. In collaboration with the RosettaNet organization, TSMC and ASE leveraged their pioneering experiences to define three data exchange standards which can then be widely adopted in the semiconductor industry. This case study is a demonstration of how two leading companies in their respective fields can join forces to make a difference in creating value for the entire semiconductor industry, which in turn benefits society at large. With the momentum continuing to build and the sphere of influence continuing to expand, it is anticipated that TSMC, ASE and the entire sector will upgrade their competitiveness in terms of cost, quality, responsiveness and customer orientation.     

Construction of a prediction model for the structural stability of a surface grinder using backpropagation neural network

Learning and prediction capability of the backpropagation neural network (BPNN) have been used to build the prediction model for the structural stability of a surface grinder. The Lagrange energy method is applied to derive the dynamic equation of the lumped parameter model of the surface grinder. The major factors influencing the structural stability of the system can be determined after the ratio of kinetic energy of the sub-structure and the ratio of potential energy of the sub-structure interface are obtained. An orthogonal rotatable central composite design is adopted to dispose the treatment combinations of the major factors. The BPNN model is constructed by the treatment combinations of the training patterns and verified by the treatment combinations of the test patterns. In this paper, a 3-layer BPNN model with a 10-neuron hidden layer which converged after 4,072 learning cycles is selected to predict the structural stability of a surface grinder within the planned ranges. The percentage residuals of both training patterns and test patterns are all within 3.41%, thus the prediction accuracy of the BPNN model is excellent so that the engineering demands are well satisfied.     

New multifunction filter using an inverting CCII and a voltagefollower

A new multifunction filter using minimum components is presented. This filter which possesses single input and four outputs can simultaneously generate second-order highpass, bandpass, and lowpass filtering functions at individual output terminals. The filter has the following merits; it uses only two grounded capacitors which makes it suitable for integrated circuit implementation, no matching condition is required, offering multifunction outputs and has low passive and active sensitivities. Experimental results agree very well with the theoretical result     

ID-based non-interactive zero-knowledge proof system based on one-out-of-two non-interactive oblivious transfer

In this paper, we propose an ID-based non-interactive zero-knowledge proof system based on the 1-out-of-2 noninteractive oblivious transfer protocol. This zero-knowledge proof system is secure against a newly discovered cheating attack.     

Performance of GaAs-AlGaAs V-grooved inner stripe quantum-well wirelasers with different current blocking configurations

GaAs-AlGaAs V-grooved inner stripe (VIS) quantum-well wire (QWW) lasers grown by metalorganic chemical vapor deposition with different current blocking configurations, n-blocking on p-substrate (VIPS), p-n-p-n blocking on n-substrate (VI(PN)_nS) and p-blocking on n-substrate (VINS) have been fabricated and characterized. The VIPS QWW lasers show the most stable characteristics with effective current confinement: one of the lasers shows fundamental transverse mode, lasing up to 5 mW/facet, typical threshold current of 39.9 mA at 818.5 mm, an external differential quantum efficiency of 24%/facet, and characteristic temperature of 92 K. The current tuning rate was almost linear at 0.031 mm/mA, and the temperature tuning rate was measured to be 0.14 nm/°C. Comparison of the light output versus current characteristics of the lasers with different current blocking configurations is presented here     

Estimation of collapse moment for the wall-thinned pipe bends using fuzzy model identification

In this work, the collapse moment due to wall-thinned defects is estimated through fuzzy model identification. A subtractive clustering method is used as the basis of a fast and robust algorithm for identifying the fuzzy model. The fuzzy model is optimized by a genetic algorithm combined with a least squares method. The developed fuzzy model has been applied to the numerical data obtained from the finite element analysis. Principal component analysis is used to preprocess the input signals into the fuzzy model to reduce the sensitivity to the input change and the fuzzy model are trained by using the data set prepared for training (training data) and verified by using another data set different (independent) from the training data. Also, three fuzzy models are trained, respectively, for three data sets divided into the three classes of extrados, intrados, and crown defects, which is because they have different characteristics. The relative root mean square (RMS) errors of the estimated collapse moment are 0.5397% for the training data and 0.8673% for the test data. It is known from this result that the fuzzy models are sufficiently accurate to be used in the integrity evaluation of wall-thinned pipe bends and elbows.     

Spray deposition process study of SnO2silicon solar cells using orthogonal experimental design

With the “orthogonal experimental design” (OED), the dependence of conductivity and transparency of tin oxide films and open circuit voltage V_oc of SIS solar cells on the spray deposition process factors was studied. The OED technique was found to be a powerful method for realizing the best factor combination. Using OED, curves were obtained which clearly depict the effects of each factor on the SIS characteristics. The results of the experiments also illustrate exactly which variation in fabrication technique most affects the sheet resistance, V_oc, and film transparency. Under the optimal combination of conditions, a thin film doped with NH₄F was obtained, with about 90% transmission and a sheet resistance of about 80 Ω/□, along with an SIS structure with a V_oc of about 0.61 V.     

P-V-T-C equation for epoxy molding compound

The isothermal and isobaric volume shrinkage is measured by a single-plunger-type dilatometer for epoxy molding compound (EMC). This device has been found suitable for measuring volume change of thermosetting materials such as commercial EMC under isothermal and isobaric conditions. Moreover, the degree of cure (conversion) was determined by a differential scanning calorimetry (DSC). Combining volume change and conversion, a mathematic pressure-volume-temperature-cure (P-V-T-C) model is proposed to describe the relationship between volume shrinkage, pressure, temperature and conversion. The P-V-T-C equation can be simply expressed as VS(P,T,C)=F/sub 1/(P,T)/spl middot/C/sup F2(P,T)/. This equation can well describe historical profiles of volume shrinkage under specified isothermal and isobaric states. From the predicted results, volume shrinkage under different pressure levels in any specified temperature can be approximated as and it obeys the principle of linearity. With the help of this model, together with three-dimensional mold filling simulation, engineers will be able to predict warpage and residual stresses for a package after molding.     

Explosive Random‐Coefficient AR(1) Processes and Related Asymptotics for Least‐Squares Estimation

Abstract. Large sample properties of the least‐squares and weighted least‐squares estimates of the autoregressive parameter of the explosive random‐coefficient AR(1) process are discussed. It is shown that, contrary to the standard AR(1) case, the least‐squares estimator is inconsistent whereas the weighted least‐squares estimator is consistent and asymptotically normal even when the error process is not necessarily Gaussian. Conditional asymptotics on the event that a certain limiting random variable is non‐zero is also discussed.     

Projection methods in flexible multibody dynamics. Part I: Kinematics

In this paper a recursive projection method for the dynamic analysis of open-loop mechanical systems that consist of a set of interconnected deformable bodies is presented. The configuration of each body in the system is identified using a coupled set of reference and elastic co-ordinates. The absolute velocities and accelerations of leaf or child bodies in the open-loop system are expressed in terms of the absolute velocities and accelerations of the parent bodies and the time derivatives of the relative co-ordinates of the joints between the bodies. The dynamic differential equations of motion are developed for each link using the generalized Newton-Euler equations. The relationship between the actual joint reactions and the generalized forces combined with the kinematic relationships and the generalized Newton-Euler equations are used to develop a system of loosely coupled equations which has a sparse matrix structure. Using matrix partitioning and recursive projection techniques based on optimal block factorization an efficient solution for the system accelerations and joint reaction forces is obtained. This solution technique yields a much smaller operations count and can more effectively exploit vectorization and parallel processing. It also allows a systematic procedure for decoupling the joint and elastic accelerations.