A practical position-based visual servo design and implementation for automated fault insertion test

This paper presents the design and implementation of a practical visual servo. The visual servo has been designed for pose correction of an end-effector that is placed wrongly away from the test point on a printed circuit board (PCB). It is widely acknowledged that the error could be attributed to the connecting joints of robot arms, angular errors, geometric model, and/or camera projection model. In the automated fault insertion test (AFIT), the typical difficulty encountered by a robot is to servo and place the probe tip accurately on the targeted test point, i.e., the conductive pad on the PCB. Conventionally, touch and sense with a probe tip has been utilized. Upon detection of a failure, the visual servo is triggered and the robot literally looks through a camera and re-attempts with the use of visual information from the camera. Currently, no clearly defined specifications in carrying out feature extraction exist and thereby the test point detection as it is not cost-effective to designate part of PCB footprint as the test point separately from the main design to accommodate visual sensing. As a result, various kinds of test points have been implanted into PCB industrial design. This research work requires building of a custom knowledge base of the test point features to support image recognition. Furthermore, the operational factors in industrial manufacturing, e.g., head structure maintenance, the replacement of end-effectors and the changes of projection parameters caused by hardware adjustment, impact the accuracy of the probe placement. These factors cause the original geometric model to shift from the original configuration and thus errors. This research paper proposes the practical design of a closed-looped visual servo to address the issues of precision error, image feature extraction, and manufacturing factors. Besides, the paper details the findings from the design phase to the implementation phase.     

Dielectric properties of Pb[(1?x)(Zr1/2Ti1/2)?x(Zn1/3Ta2/3)]O3 ceramics prepared by columbite and wolframite methods

Polycrystalline samples of Pb[(1 − x)(Zr_1/2Ti_1/2) − x(Zn_1/3Ta_2/3)]O ₃ , where x = 0.1–0.5 were prepared by the columbite and wolframite methods. The crystal structure, microstructure, and dielectric properties of the sintered ceramics were investigated as a function of composition via X-ray diffraction (XRD), scanning electron microscopy (SEM), and dielectric spectroscopy. The results indicated that the presence of Pb(Zn_1/3Ta_2/3)O₃ (PZnTa) in the solid solution decreased the structural stability of overall perovskite phase. A transition from tetragonal to pseudo-cubic symmetry was observed as the PZnTa content increased and a co-existence of tetragonal and pseudo-cubic phases was observed at a composition close to x = 0.1. Examination of the dielectric spectra indicated that PZT–PZnTa exhibited an extremely high relative permittivity at the MPB composition. The permittivity showed a ferroelectric to paraelectric phase transition at 330 °C with a maximum value of 19,600 at 100 Hz at the MPB composition.     

Fully balanced current-tunable sinusoidal quadrature oscillator

An integrable current-tunable sinusoidal quadrature oscillator is presented. A current-tunable all-pass filter using a signal-differencing technique is realized as the frequency-selective network. The implementation is fully balanced so as to enable accurate quadrature signals with symmetry. The oscillation frequency is current-tunable over a wide-frequency sweep range of approximately three orders of magnitude. The quadrature signals possess the amplitude matching and the quadrature phase matching of better than 0.004dB and 0.15°, respectively. The maximum useful frequency of oscillation is in excess of 8MHz. Total harmonic distortions are less than 0.5%.     

A 10.7-MHz fully balanced,high-Q, 87-dB-dynamic-range current-tunable Gm-C bandpass filter

A 10.7-MHz fully balanced, high-Q, wide-dynamic-range current-tunable Gm-C bandpass filter is presented. The technique is relatively simple based on two fully balanced components, i.e. an adder and a low-Q-based bandpass filter. The Q factor is approximately equal to a typically high and constant value of a common-emitter current gain (β) and is, for the first time, independent of variables such as a center frequency. Possible solutions for good stability of the Q factor with temperature are suggested. Not only can the need for additional Q-tunable circuits be greatly reduced, the sensitivity of the Q factor can be greatly improved. Sensitivities of either the Q factor or the center frequency are constant between 1 and −1 and are no longer strongly affected by the Q factor or variables. As a simple example at 10.7 MHz, the paper demonstrates the high-Q factor of 121, the low total output noise of 5.303 μV_rms, the 3rd-order intermodulation-free dynamic range (IMFDR₃) of 74.45 dB and the wide dynamic range of 87.45 dB at 1% IM₃. The center frequency is current tunable over three orders of magnitude. Comparisons to other 10.7-MHz Gm-C approaches are also included.     

A minimum five‐component five‐term single‐nonlinearity chaotic jerk circuit based on a twin‐jerk single‐op‐amp technique

A minimum 5‐component 5‐term single‐nonlinearity chaotic jerk circuit is presented as the first simplest chaotic jerk circuit in a category that a single op‐amp is employed. Such a simplest circuit displays 5 simultaneous advantages of (1) 5 minimum basic electronic components, (2) 5 minimum algebraic terms in a set of 3 coupled first‐order ordinary differential equations (ODEs), (3) a single minimum term of nonlinearity in the ODEs, (4) a simple passive component for nonlinearity, and (5) a single op‐amp. The proposed 5‐term single‐nonlinearity chaotic jerk circuit and a slightly modified version of an existing 6‐term 2‐nonlinearity chaotic jerk circuit form mirrored images of each other. Although both mirrored circuits yield 2 different sets of the ODEs, both sets however can be recast into a pair of twin jerk equations. Both mirrored circuits are therefore algebraically twin 5‐component chaotic jerk circuits, leading to a twin‐jerk single‐op‐amp approach to the proposed minimum chaotic jerk circuit. Two cross verifications of trajectories of both circuits are illustrated through numerical and experimental results. Dynamical properties are also presented.