Search

It is crucially important to establish an accurate model to represent the relationship between the actuator forces and the lap surface changes when polishing a large and highly aspheric optical surface. To facilitate a computer-controlled optical polishing process, a neural network based stressed lap surface shape model was developed. The developed model reflects the dynamic deformation of a stressed lap. The original data from the microdisplacement sensor matrix were used to train the neural network model. The experimental results show that the proposed model can represent the surface shape of the stressed lap accurately and provide an analytical model to be used to polish the stressed lap control system and the active support system for a large mirror. OCIS codes: 240.0240, 240.5450.

In this paper, an adaptive controller is developed for the pressure tracking of the pressurized toolhead in order to maintain the constant contact stress for the polishing process. This is a new polishing control method, which combines the adaptive control theory and the constant stress theory of the contact model. By using an active pneumatic compliant toolhead, a recursive least-squares estimator is developed to estimate the pneumatic model and then a minimum-degree pole-placement method is applied to design a self-tuning controller. The simulation and experiment results of the proposed controller are presented and discussed. The main advantage of the constant contact stress control is high figuring accuracy. [DOI: 10.1115/1.4000959]

We report on magnetorheological finishing (MRF) spotting experiments performed on glasses and ceramics using a zirconia-coated carbonyl-iron (CI)-particle-based magnetorheological (MR) fluid. The zirconia-coated magnetic CI particles were prepared via sol-gel synthesis in kilogram quantities. The coating layer was ~50-100 nm thick, faceted in surface structure, and well adhered. Coated particles showed long-term stability against aqueous corrosion. 'Free' nanocrystalline zirconia polishing abrasives were cogenerated in the coating process, resulting in an abrasive-charged powder for MRF. A viable MR fluid was prepared simply by adding water. Spot polishing tests were performed on a variety of optical glasses and ceramics over a period of nearly three weeks with no signs of MR fluid degradation or corrosion. Stable material removal rates and smooth surfaces inside spots were obtained. OCIS codes: 160.2750, 220.4610, 240.5450, 120.6660.

In optical communication systems, 980 nm high-power laser modules are applied for long-distance communication amplifiers. Since the mode field of the 980 nm laser diodes is different from single-mode fibers, coupling losses may occur. In order to improve this problem, lensed fibers have been used widely recently, such as lensed fiber employing quadrangular-pyramid-shaped fiber endface (QPSFE). The QPSFE can easily yield any elliptical microlens aspect ratio to match the far field of the high-power laser diode but the fiber endface must be polished and then spliced during present manufacturing process. It is a time-consuming task and the melting depth of the fiber endface is controlled with difficulty. This study develops a novel fiber endface polishing method for manufacturing the lensed fiber employing QPSFE. Using this new method, the coupling efficiency exceeds 70%. [DOI: 10.1115/1.3168441] Keywords: polishing, lensed fiber, elliptical microlens

Although the Isoform[R] lengthwise-reciprocating grinding process is considered as one of the most accurate methods for generating the tooth profile geometry of a helical gear shaping cutter, the tooth profile accuracy produced by the Isoform[R] with a straight cone grinding wheel is not accurate enough for high precision requirement. That is why the shaper cutter is used as a rough cutting tool for most cases. A third-order profile correction to the cone grinding wheel is proposed to increase the accuracy of the work gear profile. A novel topography is developed to schematically show the work gear tooth profile accuracy cut by a resharpened shaping cutter. The profile errors corresponding to the varied resharpening depth are shown in the topography with information of true involute form diameter and semitopping depth. The usable resharpening depth of the shaping cutter can be determined by this topography. The numerical result indicates that third-order correction reduces the profile error of the major cutter enveloping gear to submicro and extends the resharpening depth. [DOI: 10.1115/1.3087553] Keywords: helical shaping cutter, Isoform[R] grinding process, cutter enveloping gear, profile correction, third-order correction, novel topography

During hypoid gears lapping process, a gear set is running at varying operating positions and under a light load in order to lap the complete tooth surface. The pinions and gears are lapped in pairs. Hence, their tooth surfaces are not only cutters but also workpieces. In the contact region, the contact pressure and sliding speed are different from point to point. This makes lapping to be a very complicated abrasive wear process. So far, knowledge about the relationship between the removed materials and the lapping time as well as how to optimize the lapping process is quite limited. An algorithm was presented (Jiang et al., 2008, 'Simulation of Hypoid Gear Lapping,' ASME J. Mech. Des., 130(11), p. 112601) to determine the wear coefficient k for the lapping process of hypoid gears. With the obtained wear coefficient k, a methodology for simulating the lapping process was proposed. Based on the wear coefficient obtained, this work presents a computer-aided machine setting procedure to optimize the lapping cycle in order to improve the lapping quality and efficiency. [DOI: 10.1115/1.3066548] Keywords: hypoid gears, lapping, machine setting, optimization

This paper is based on a microinteraction principle of fabricating a RB-SiC material with a fixed abrasive. The influence of the depth formed on a RB-SiC workpiece by a diamond abrasive on the material removal rate and the surface roughness of an optical component are quantitatively discussed. A mathematical model of the material removal rate and the simulation results of the surface roughness are achieved. In spite of some small difference between the experimental results and the theoretical anticipatiom which is predictable, the actual removal rate matches the theoretical prediction very well. The fixed abrasive technology's characteristic of easy prediction is of great significance in the optical fabrication industry, so this brand-new fixed abrasive technology has wide application possibilities. OCIS codes: 220.0220, 220.4000, 220.4610, 220.5450.

Most of the methods for generating regular surface texture (RST) consist of shaping a set of regular grooves (cavities) arranged in a regular way. This paper presents possibilities for regular surface texture generation by so-called 'pattern grinding' with the wheel prepared in a special way. The simple variant of the method involves grinding with the wheel having helical grooves. The grooves shaped on the work material are the result of specific wheel surface reproduction. The ratio between work-material feed and wheel speed is an important factor, determining the layout of the grooves generated on the work-material and the shape of the groove sides. Surface texture consists of two components: deterministic, resulting from the nominal wheel active surface, and random, resulting from the random shape and arrangement of abrasive grains. The limited contribution of the random component of surface texture is discussed based on the ratio between the undeformed chip thickness and the sizes of the grooves. Kinematical analysis of the wheel reproduction process is performed for description of nominal surface texture. Experimental results of flat and cylindrical surfaces, obtained with pattern grinding are also provided. Two critical values of the ratio between work-material feed and wheel speed were derived, and three ranges of this ratio are discussed. The kinematical approach provided relationships between input data of the process (wheel shape and grinding parameters) and nominal groove dimensions and groove layout. The geometrical characteristics of the work-material nominal surface texture are presented for each of the three types of surface texture. It is important to ensure that the work feeds are greater than the lower critical value. For achievable work feeds the shape of the sides of the grooves is cycloid. Experiments revealed the limited contribution of the random component of the surface structure of the work material. Random arrangement of abrasive grains is important only at local (micro-) level and affects the roughness of groove bottoms, while the dimensions and arrangement of the grooves are affected only to a minimal degree. [DOI: 10.1115/1.3070511] Keywords: grinding, regular surface texture

Chemical mechanical polishing (CMP) is a manufacturing process that is commonly used to planarize integrated circuits and other small-scale devices during fabrication. Although a number of models have been formulated, which focus on specific aspects of the CMP process, these models typically do not integrate all of the predominant mechanical aspects of CMP into a single framework. Additionally, the use of empirical fitting parameters decreases the generality of existing predictive CMP models. Therefore, the focus of this study is to develop an integrated computational modeling approach that incorporates the key physics behind CMP without using empirical fitting parameters. CMP consists of the interplay of four key tribological phenomena--fluid mechanics, particle dynamics, contact mechanics, and resulting wear. When these physical phenomena are all actively engaged in a sliding contact, the authors call this particle-augmented mixed lubrication (PAML). By considering all of the PAML phenomena in modeling particle-induced wear (or material removal), this model was able to predict wear-in silico from a measured surface topography during CMP. The predicted material removal rate (MRR) was compared with experimental measurements of copper CMP. A series of parametric studies were also conducted in order to predict the effects of varying slurry properties such as solid fraction and abrasive particle size. The results from the model are promising and suggest that a tribological framework is in place for developing a generalized first-principle PAML modeling approach for predicting CMP. Keywords: mixed lubrication, chemical mechanical polishing, contact mechanics, particle dynamics, computational fluid dynamics, abrasive wear

Based on the direct contact between the wafer and the pad, the pressure and velocity dependence of the material removal rate (MRR) in the fast polishing process (FPP) is investigated. There are three assumptions of the FPP material removal mechanism: the normal distribution of abrasive size, a periodic roughness of the pad surface, and the plastic contact between wafer--abrasive and pad--abrasive interfaces. Based on the particular FPP, a novel movement of the wafer is analyzed and a MRR equation is developed. The experiments with parameters of pressure and velocity are shown to verify the equation. Thus, a better understanding of the fundamental mechanism involved in FPP can be obtained. OCIS codes: 220.0220, 220.5450.

Microelectrochemical grinding (micro-ECG) is a variant of electrochemical grinding (ECG) process, in which material is removed by a combination of electrolytic dissolution and abrasive action that take place in a small interelectrode gap. This paper discusses analytical modeling of the material removal phenomenon in micro-ECG process to predict material removal rate. In the model, the phenomena, which contribute to the material removal in the process by electrolytic and abrasive actions, have been considered; these include streaming potential in the electrochemical action and shearing forces due to the flow of electrolyte through interelectrode gap and the abrasive action of grinding wheel. Two configurations of the process, viz., surface and cylindrical micro-ECG, have been modeled. The results have been validated by CFD simulation in the case of surface micro-ECG process, and specific experimentation in the case of cylindrical micro-ECG process. [DOI: 10.1115/1.2844587] Keywords: micromachining, micro-ECG, electrochemical grinding, material removal rate, CFD simulation, abrasion

This paper deals with the development of an online monitoring system based on feature-level sensor fusion and its application to OD plunge grinding. Different sensors are used to measure acoustic emission, spindle power, and workpiece vibration signals, which are used to monitor three of the most common faults in grinding--workpiece burn, chatter, and wheel wear Although a number of methods have been reported in recent literature for monitoring these faults, they have not found widespread application in industry, as no single method or feature has been shown to be successful for all setups and for all wheel-workpiece combinations. This paper proposes a systematic approach, which allows the development and deployment of process-monitoring systems via automated sensor and feature selection combined with parameter-free model training, both of which are especially crucial for implementation in industry. The proposed algorithm makes use of 'sparsity-promoting' penalty terms to encourage sensor and feature selection while the 'hyperparameters' of the algorithm are tuned using an approximation of the leave-one-out error Experimental results obtained for monitoring burn, chatter, and wheel wear from a plunge grinding test bed show the effectiveness of the proposed method. [DOI: 10.1115/1.2927439] Keywords: grinding, process monitoring, machine learning, feature selection, sensor fusion, leave-one-out error, least squares support vector machine

Chemical-mechanical polishing (CMP), active and via fills have become indispensable aspects of semiconductor manufacturing. CMP fills are used to reduce metal thickness variations due to chemical-mechanical polishing. Via fills are used to improve neighboring via printability and reliability of low-k and ultra low-k dielectrics. Active region fills are used for STI CMP uniformity and stress optimization. Although modern parasitic extraction tools accurately handle grounded fills and regular interconnects, such tools use only rough approximations to assess the capacitance impact of floating fills, such as assuming that floating fills are grounded or that each fill is merged with neighboring ones. To reduce such inaccuracies, we provide a design of experiments (DOE) which complements what is possible with existing extraction tools. Through the proposed DOE set, a design or mask house can generate normalized fill tables to correct for the inaccuracies of existing extraction tools when floating fills are present. Golden interconnect capacitance values can be updated using these normalized fill tables. Our proposed DOE enables extensive analyses of fill impacts on coupling capacitances. We show through 3-D field solver simulations that the assumptions used in extractors result in significant inaccuracies. We present analyses of fill impacts for an example technology and also provide analyses using the normalized fill tables to be used in the extraction flow for three different standard fill algorithms. We also extend our analyses and methodology to via fills and active region fills, which have more recently been introduced into semiconductor design-manufacturing methodologies and for which sufficient understanding is still lacking. Index Terms--Active region fill, chemical-mechanical polishing (CMP) fill, coupling capacitance, RC extraction, via fill.

We report a simple method for generating microaxicons at the extremity of commercial optical fibers. The proposed solution, based on a polishing technique, can readily produce any desired microaxicon cone angle and is independent of the nature of the fiber. An optical study of microaxicon performance, in terms of confinement ability and length of the generated Bessel-like beams, is presented as a function of the microaxicon angle. This study, made possible by the experimental acquisition of the 3D light distribution of the Bessel-like beams, reveals the relationship between the Bessel-like beam confinement zone and the beam length. Finally, the effect of diffraction of the Bessel-like beams, induced by the limited lateral extent of the incident fiber mode, is studied and discussed. OCIS codes: 120.4640, 130.1750, 220.3630, 220.1920, 230.1150, 350.3950.

In the chemical-mechanical polishing (CMP) process, the pad conditioning density distribution plays a crucial role in the pad wear. A precise model and detailed analysis of a conditioning density function are required. To this end, at first we construct the mathematical model of polishing trajectories on the pad, then under the assumptions that the diamond grains are uniformly distributed and a slow sweeping motion is applied during dressing, the conditioning density distribution for a pad in CMP process is determined. This conditioning density function is verified through numerous numerical examples. In the mean time, it was also observed that to have a flat distribution of pad wear rate we have to make the ratio of disk-radius to pad-radius small, and the effect of the pattern of grain distribution on conditioning density function is insignificant, which agrees with known results from literature. Index Terms--Chemical-mechanical polishing (CMP), conditioning density, diamond dresser, pad dressing, Preston equation.

In this paper, we analyze the nonuniformity of sliding distance on both the wafer and polishing pad from a kinematic point of view. Using the Fourier series expansion, it can be shown that in steady state the nonuniformity caused by contact relative velocity is determined by rotational speed ratio between platen and wafer carrier (m) and the ratio of wafer radius to the distance between the platen center and wafer center ([R.sub.c]/d). In general, the nonuniformity of wafer increases with |m| and ([R.sub.c]/d). An important observation for the polishing pad is that in two particular ranges of the ratio m, larger ([R.sub.c]/d) on the contrary yields smaller nonuniformity. Then, a ring-type polishing pad is proposed for the purpose of improving the nonuniformity of both wafer and pad. However, it turns out the result for the pad of large size is worse than the traditional shape, unless the rotational speed of the pad is much slower than that of the wafer. Index Terms--Chemical-mechanical polishing (CMP), nonuniformity, Preston equation, ring-type polishing pad, sliding distance.

In this paper, a multi-level fuzzy control (MLFC) technique is developed and implemented for a creep-feed grinding process. The grinding force is maintained at the maximum allowable level under varying depth of cut, so that the highest metal removal rate is achieved with a good workpiece surface quality. The control rules are generated heuristically without any analytical model of the grinding process. Based on the realtime force measurement, the control parameters are adapted automatically within a stable range. A National Instrument real-time control computer is implemented in an open architecture control system for the grinding machine. Experimental results show that the cycle time has been reduced by up to 25% over those without force control and by 10-20% compared with the conventional fuzzy logic controller, which indicates its effectiveness in improving the productivity of actual manufacturing processes. The effect of grinding wheel wear is also considered in the creep-feed grinding process, where the grinding force/power can be maintained around the specified value by the proposed MLFC controller as the wheel dulls gradually. [DOI: 10.1115/1.2718238]

The electrolytic, in-process dressing (ELID) grinding is a new and an efficient process for ultra-precision finishing of hard and brittle materials. Unlike conventional grinding processes, the ELID grinding is a hybrid process that consists of a mechanical and an electrochemical process, and the performance of the ELID grinding process is influenced by the parameters of the above said processes. Therefore, it is necessary to develop a new grinding model for the ELID grinding, which can be used to avoid the cumbersome and expensive experimental trials, In this paper, the authors proposed a new grinding model for ultra-precision ELID grinding. The main focus is to develop a force model for the ultra-precision ELID grinding where the material removal is significantly lower than the conventional grinding. When the material removal rate is very low, it is very important to estimate the real contact area between the wheel and work surfaces. The developed grinding model estimates the real contact area by considering the wheel and the work surface characterization and the effect of the electrolytic reaction at the grinding wheel edge. The effects of the microstructure changes on the wheel surface by the electrochemical reaction have been implemented in the model in order to improve the efficiency of the developed model. The grinding model has been simulated and the simulated results are substantiated by the experimental findings. [DOI: 10.1115/1.2515382] Keywords: ELID, nanogrinding, grinding model, superabrasive wheels

Surface defects generated by grinding deteriorate the flexural strength of the silicon nitride (Si3N4) ceramic. CO2 laser surface processing was applied to eliminate the grinding-induced defects and was concluded that laser surface processing had significant effects on fracture behavior of flexure silicon nitride samples.

This paper presents a time-domain dynamic model, which simulates cylindrical plunge grinding processes under general grinding conditions. The model focuses on the prediction of grinding chatter boundaries and growth rates. Critical issues are considered in the model including: the distributed nonlinear force along the contact length, the geometrical interaction between the wheel and workpiece based on their surface profiles, the structure dynamics with multiple degrees of freedom for both the wheel and workpiece, the response delay due to spindle nonlinearities and other effects, and the effect of the motion perpendicular to the normal direction. A simulation program has been developed using the model to predict regenerative forces, dynamic responses, surface profiles, stability regions, and chatter growth rates. The model is validated using existing numerical and experimental results. [DOI: 10.1115/1.2118748]

The fluid flow problem in a traditional electrolytic in-process dressing (ELID) system is analyzed and solved numerically. The predicted mean velocity profiles in the dressing zone show flow patterns that are in good agreement with the mean velocity distributions for plane laminar/turbulent Couette flows observed in the experiments. The computational results reveal that insufficient electrolyte supply rate is the cause of the failure of the traditional ELID system for high-speed grinding. Results also show that to obtain effective high-speed ELID grinding, a consistent high inlet electrolyte velocity or supply rate is required. For the foil ELID system, governing equations describing the fluid flow in the dressing zone and the foil elastic deformation are formulated. Analytical solution based on unidirectional flow model for the problem is presented and effects of wheel surface speed and foil tension on the performance of the dressing system are discussed. It is shown that the foil ELID system has the potential to be effective for high-speed grinding with low electrolyte supply rates. The results will be useful to the development of new machine systems and processes for high-speed grinding. [DOI: 10.111511.1765152]

This paper presents implementation results of surface grinding processes based on the model-based optimization scheme proposed by Lee and Shin (Lee, C. W., and Shin, Y. C., 2000 'Evolutionary Modeling and Optimization of Grinding Processes, 'Int. J. Prod. Res. 38(12), pp. 2787-2813). In order to accomplish this goal process models for grinding force, power, surface roughness, and residual stress are developed based on the generalized grinding model structures using experimental data. The time-varying characteristics due to wheel wear are also investigated in order to determine the optimal dressing interval Grinding optimization is considered as constrained nonlinear optimization problems with mixed-integer variables and time-varying characteristics in this study. Case studies are performed with various optimization objectives including minimization of grinding cost, minimization of cycle time, and process control. The optimal process conditions determined by the optimization scheme are validated by experimental results. [DOI: 10.1115/1.1537738]

This paper presents a scheme for extracting information from interferometry signals off patterned wafer polish during nonendpointed chemical-mechanical polishing (CMP). This enables one to obtain blanket removal rates immediately after a patterned wafer finishes polishing as well as estimate post-polish within lot thickness variation. A nonlinear regression algorithm is presented that enables one to estimate this information from less than a full interferometry trace cycle and with a lower sampling rate compared to peak-valley detection schemes. Index Terms--Chemical-mechanical polishing (CMP), interferometry, nonlinear regression, process control.

When a metal is surface ground the roughness generated is the summation of a function of the wheel roughness and the roughness due to wheel attrition and damage to the workpiece. We identify this function here as a maximum envelope profile, which is fractal within certain cut off wavelengths determined by the dressing conditions of the wheel. Estimating the global displacement of the binder-grit-workpiece system from the maximum envelope power spectra, we determine the plastic indentation of the workpiece at characteristic length scales using simple contact-mechanical calculation. The estimated roughness corresponds well with that recorded experimentally for hard steel, copper, titanium and aluminum.

Conditions during finishing of advanced ceramics by magnetic abrasive finishing (MAF) processes are found to be, by and large, transient. Consequently, the available quasi-steady-state solutions for the moving heat sources are not directly applicable for this case. Hence, the general solution for a moving disk heat source, developed in Part I of this three-part series, is applied to determine the minimum flash temperatures and flash times generated at the contact points between the workmaterial ([Si.sub.3][N.sub.4] roller) and the abrasive ([Cr.sub.2][O.sub.3]). Since chemo-mechanical action between the abrasive - the workmaterial - the environment depends on both the thermodynamics and kinetics of the process, it is important to determine the flash temperatures as well as flash times during polishing. These were determined as a function of the polishing pressure and the rotational speed of the work material in this investigation. Thermodynamic considerations (not covered in this paper) indicate that even the minimum flash temperatures generated under the conditions of lower pressure, lower sliding velocity, and transient state would be adequate to initiate chemo-mechanical action, and experimental results confirmed the formation of chemo-mechanical reaction products during polishing (Bhagavatula and Komanduri, 1996).

The thermal model developed in Part I of this three-part series is applied in this paper to magnetic float polishing (MFP) of ceramic ([Si.sub.3][N.sub.4]) balls. Using this method, the flash temperatures, flash times, and temperature distribution at the interface between the balls and the shaft of the MFP apparatus are calculated. Examination of the polished surfaces (scratch lengths) of the balls showed that the length of most scratches during the final stage of polishing is

A thermal model for magnetic field assisted polishing of ceramic balls/rollers is presented. The heat source at the area of contact between the balls and the abrasives where material removal takes place is approximated to a disk. The disk heat source is considered as a combination of a series of concentric circular ring heat sources with different radii. Each ring in turn is considered as a combination of a series of infinitely small arc segments and each arc segment as a point heat source. Jaeger's classical moving heat source theory (Jaeger, 1942; Carslaw and Jaeger, 1959) is used in the development of the model, starting from an instantaneous point heat source, to obtain the general solution (transient and steady-state) of the moving circular ring heat source problem and finally the moving disc heat source problem. Due to the formation of fine scratches during polishing (on the order of a few micrometers long), the conditions are found to be largely transient in nature, Calculation of the minimum flash temperatures and minimum flash times during polishing enables the determination if adequate temperatures can be generated for chemomechanical polishing or not. This model is applied in Part II for magnetic float polishing (MFP) of ceramic balls and in Part III for magnetic abrasive finishing (MAP) of ceramic rollers.

Magnetic float polishing is a new technique developed for efficient finishing of advanced ceramics. The equipment for this application thus far has been developed on the basis of empirical designs and trial and error experimental approach which can be costly and time consuming. In the work reported here both analytical (FEM simulation) and experimental approaches were taken to determine such design parameters as the variation of magnetic float stiffness with buoyant force and variation of buoyant force with the gap between the magnet and the float. The agreement between the experimental and analytical results is found to be excellent. Consequently, it is possible to simulate different designs without actually building different equipment for the performance evaluation. Also, the effect of ball circulation speed in magnetic float polishing is discussed qualitatively for assessing the conditions for high removal rates and/or best finish. Such an approach can facilitate the development of operating conditions maps for magnetic float polishing.

Four algorithms for removing shape and waviness errors in sculptured surface production processes are described in the paper. One of the algorithms employs an open-loop strategy without inspection, error analysis, and error compensation. The other three algorithms employ closed-loop inspection error analysis and error compensation strategies to manipulate control surfaces used in sculptured surface production. Coordinate measurements made on the surface being produced are compared with a designed surface and the results are used to modify related control surfaces that are used to guide processing equipment. Two of the closed-loop algorithms also use intermediate planned surfaces to improve error compensation and production control. Experiments are described in the paper in which the algorithms were tested on an experimental surface finishing system that included an optical probe, grinding spindle, and computer control system integrated with a CNC machining center. The results obtained using open-loop and closed-loop algorithms are compared, and it is shown that surface inspection, surface error analysis, surface compensation, and surface grinding can be iteratively applied to converge rough-machined test surfaces to their designed shape. The closed-loop algorithms are shown to be capable of compensating for disturbances in the finishing process that went undetected when the open-loop algorithm was used. The closed-loop algorithms have significant potential for application in automated finishing systems for molds and dies.

We have investigated the surface finishing of lithium niobate by using the single-point diamond turning technique. Surface finishes of better than 5 nm rms on z-oriented samples have been achieved. However, tool wear and spalling are much more significant with lithium niobate than with materials such as the crystals KDP and LAP. We present preliminary results comparing the optical damage thresholds of polished and diamond-turned samples.

The growing demand for a wide variety of data communication and telecomunication service requires interconnection devices, such as fiber optic connectors, cable assemblies, and adapters. The strong demands for high performance and cost reduction of these components are thus arising. To ensure connector performance in actual use, several process parameters related to geometric and optical characteristics of the connector must be thoroughly understood during the manufacturing stage. This paper describes an improved polishing technique, which greatly saves production cost and polishing time. A suggested grinding profile for each step is also presented to extend the lifetime cycle on lapping films. The less the number of polishing process involve, the less the geometric imperfection it can produce. By a proper choice of rubber pad hardness and processing conditions, we are able to obtain fiber connectors with 90% overall yield using two-step process. [DOI: 10.1115/1.2927438]