Stress analysis of lead-free solders with under bump metallurgy in a wafer level chip scale packageTseng, S.; Chen, R.; Lio, C.
doi: 10.1007/s00170-005-0165-zpmid: N/A
The wafer level chip scale assembly (WLCSP) has increasingly become popular due to its compact, wafer scale assembly. In a WLCSP assembly, the under bump metallurgy (UBM) connecting the solder joints and the chip is crucial for the assembly reliability. This study focuses on a WLCSP with 96.5Sn3.5Ag/95.5Sn3.8Ag0.7Cu solder joints and Ti/Cu/Ni UBM on a 2–layer microvia build-up electric board. Furthermore, the Garofalo-Arrhenius creep model in finite element analysis ANSYS 6.0 is used for simulations on the WLCSP assembly under thermal cycling to investigate the deformations of the assembly with different thickness of nickel layer, the maximum equivalent strain and maximum equivalent stress of microvias/joints. Finally, the Coffin-Manson equation is applied to predict the fatigue lives of four combinations of solder joints with different eutectic alloy and thickness of nickel layer.
Investigation of surface roughness in turning unidirectional GFRP composites by using RS methodology and ANNBagci, Eyup; Işık, Birhan
doi: 10.1007/s00170-005-0175-xpmid: N/A
Fibre reinforced plastics (FRP) contain two phases of materials with drastically distinguished mechanical and thermal properties, which brings in complicated interactions between the matrix and the reinforcement during machining. Surface quality and dimensional precision will greatly affect parts during their useful life especially in cases where the components will be in contact with other elements or materials during their useful life. Therefore, their study and characterisation is extremely important and, above all, those cases subjected to adverse environmental conditions and in contact with other elements or materials. Thus, measuring and characterising surface properties represent one of the most important aspects in manufacturing processes. In this paper, orthogonal cutting tests were carried out on unidirectional glassfibre reinforced plastics (GFRP), using cermet tools. During the tests, the depth of cut (a), feedrate (f), cutting speed (Vc) were varied, whereas the cutting direction was held parallel to the fibre orientation. Turning experiments were designed based on statistical three level full factorial experimental design technique. An artificial neural network (ANN) and response surface (RS) model were developed to predict surface roughness on the turned part surface. In the development of predictive models, cutting parameters of cutting speed, depth of cut and feed rate were considered as model variables. The required data for predictive models are obtained by conducting a series of turning test and measuring the surface roughness data. Good agreement is observed between the predictive models results and the experimental measurements. The ANN and RSM models for GFRPs turned part surfaces are compared with each other for accuracy and computational cost.
Ground surface roughness prediction based upon experimental design and neural network modelsFredj, Nabil; Amamou, Ridha
doi: 10.1007/s00170-005-0169-8pmid: N/A
The results presented in this paper are related to the prediction of the surface roughness generated by the grinding process. The main problems associated with the prediction capability of empirical models developed using the design of experiment (DoE) method are given. The first problem is a limited aptitude to calculate an accurate minimal output value as this optimal value was found to be absurdly negative in many cases. The second problem is that these models are not able to detect particular behaviour of the outputs for particular sets of inputs. This constitutes a serious limitation of the application of this method to ground surface roughness prediction as the surface generation mechanisms differ at low and high work speed. In this study an approach suggesting the combination of DoE method and artificial neural network (ANN) is developed. x-n-1 structures using the back-propagation algorithm were selected for the developed ANNs. Data of the DoE were used to train the ANNs and the inputs of the developed ANNs were selected among the factors and the interactions between factors of the DoE depending on their significance at different confidence levels, expressed by α%. The significance was tested using the ANOVA method. Results have shown particularly, the existence of a threshold value of α% to which correspond a critical set of inputs up to which increasing the inputs, improves the learning and the prediction capability of the constructed ANNs. The built ANNs using these critical sets of inputs have shown low deviation from the training data, low deviation from the testing data and high sensibility to the inputs levels. The high prediction accuracy of the developed ANNs was conformed by the good agreement with the results of empirical models developed by previous investigations. The obtained results were valid for three kinds of steels having different properties and different hardness.
Measuring wear of the grinding wheel using machine visionSu, J.; Tarng, Y.
doi: 10.1007/s00170-005-0172-0pmid: N/A
This paper proposes a new method for measuring grinding wheel contours using machine vision. The vision-aided measuring system comprises a CCD coupled with a telecentric lens, back lighting board and frame grabber. Measuring the image of the specimen with a grinded gap substitute directly captures the image of the actual grinding wheel. Using this method makes the 3 D of the topography of the grinding wheel into the 2 D of the contour of the grinding wheel. This method significantly simplifies grinding wheel wear measuring procedures compared with the traditional methods. Therefore, the presented paper provides a new method to improve efficiency and cost on measuring wear of the grinding wheel. The results show that this developed system achieves a repeatable accuracy of ±3 μm for the measurement of the grinding wheel contours.
Prediction of dimensional errors in 3D complex shapes due to press elasticityOu, H.
doi: 10.1007/s00170-005-0178-7pmid: N/A
This paper presents an approach to predict dimensional errors in 3D complex shapes due to press geometry errors and elasticity. Using a press stiffness matrix formulation for the press deflections in forging operation, a quantitative relationship between forging die deviations and the press geometry errors and elastic deflections is developed, which is a function of the forging force, press stiffness and the spatial relationship between the forging dies and the press table. The stiffness matrix of a screw press is obtained using finite element analysis. To evaluate the effect of the press elasticity on dimensional errors of 3D components, a case study of forging for aerofoil shapes is carried out based on the results from physical modelling experiments. With the representative information of the tool shape and forging force data, numerical results of the forging die deviations as a source of dimensional errors for the aerofoil shape are obtained and evaluated. It is demonstrated that this approach is applicable to forging and other metal forming processes for complex shapes.
An algorithm of NURBS surface fitting for reverse engineeringDan, Jiang; Lancheng, Wang
doi: 10.1007/s00170-005-0161-3pmid: N/A
Reverse engineering is an approach for constructing a CAD model from a physical part through dimensional measurement and a surface model. Different from conventional methods, this paper develops a new algorithm by which a desired fitted surface is obtained with less computation. Let selected m×n measured points be control points to construct B-spline or NURBS surface, then modify this constructed surface by using all the measured points and least squares minimization. A new algorithm for parameterization for measured points is also presented in this paper. The effectiveness and efficiency of these proposed algorithms are demonstrated.
Bayesian approach for measuring EEPROM process capability based on the one-sided indices CPU and CPLWu, Chien-Wei; Pearn, W.
doi: 10.1007/s00170-005-0144-4pmid: N/A
The purpose of process capability analysis is to provide numerical measures on whether a process is capable of reproducing items meeting the manufacturing specifications. Capability analyses have received considerable recent research attention and increased usage in process assessments and purchasing decisions. Most existing research works on capability analysis focus on estimating and testing process capability based on the traditional distribution frequency approach. In this paper, we propose a Bayesian approach based on the indices CPU and CPL to measure EEPROM process capability, in which the specifications are one-sided rather than two-sided. We obtain the credible intervals of CPU and CPL and develop a Bayesian procedure for capability testing. The posterior probability p, for which the process under investigation is capable, is derived. The credible interval is a Bayesian analog of the classical lower confidence interval. A process satisfies the manufacturing capability requirements if all the points in the credible interval are greater than the pre-specified capability level w. To make this Bayesian procedure practical for in-plant applications, a real example of an EEPROM manufacturing process is investigated, demonstrating how the Bayesian procedure can be applied to actual data collected in the factories.