To enhance the reliability of analog circuits in electrical systems, this letter proposes a novel incipient fault diagnosis method by integrating wavelet transform(WT) and improved convolutional neural network. Different from traditional methods, where feature extraction and classification are separately designed and performed, this letter aims to automatically learn fault features and classify the type of faults simultaneously. An improved convolutional neural network named multi-channel compactness convolutional neural network (MC-CNN) is proposed, which can obtain complementary and rich diagnosis information from multi-scale components extracted by wavelet transform. Moreover, we adopt center loss as an auxiliary loss function to maximize the interclass separability and intraclass compactness of samples. The proposed method is fully evaluated with the Sallen-Key bandpass filter circuit and the four-opamp biquad high-pass filter circuit. The experimental results demonstrate that the proposed method is very effective in feature extraction for fault diagnosis, and has higher diagnosis accuracy than other typical fault diagnosis methods.
As a high-performance step-up/step-down converter, the synchronous switching Z-source overcomes the theoretical obstacles and limitations of traditional voltage source and current source converters. However, due to the existence of switching devices, the system has strong nonlinear characteristics. At the same time, it belongs to a high-order system, which is easily affected by changes in environment and structural parameters, causing local system instability, namely bifurcation and chaos. To solve the above problems, the discrete iterative model is established on the basis of the state equation of the system, the steady-state operating point of the system was analyzed by nonlinear theory, and the mechanism of the nonlinear behavior was revealed, so as to determine the parameter conditions of the system stability region. Compared with solving the differential equation directly, this paper maps the discrete iterative model to the complex frequency domain for solution, which can greatly reduce the amount of calculation while ensuring the accuracy of the model. The analysis provides the parameter basis for the stable operation of the system and provides theoretical support for further optimization of the system design.
The current aggregation mechanism created by the gate structure is proposed for electrostatic discharging (ESD). Through device simulation, the size-expanded gate structure in gate-control dual-direction silicon controlled rectifier (GC-DDSCR) is found to aggregate the surface parasitic current path and the main SCR current path. The SCR current path is consequently twisted and extended to increase the holding voltage (Vh). Two GC-DDSCRs are fabricated in a 0.5µm CMOS technology and tested by transmission line pulse (TLP). The Vh increases from 13.84V to 16.44V as the gate size expands from 2.5µm to 4.5µm. The mechanism of current aggregation is verified.
This letter presents a wideband and high conversion gain mixer based on Gilbert-cell in InP DHBT process. Capacitive degeneration is introduced to increase the bandwidth of the mixer with no voltage drop and few cost of area. Current bleeding technique is employed to improve conversion gain. Utilization of device ft is analyzed to achieve 0.919. The measurement demonstrates a conversion gain of 14.5 dB at 26 GHz with -3 dB bandwidth of 49 GHz. BW/ft of 0.306 and GBP of 260 GHz are obtained, which are believed to be among the best compared to mixers fabricated with the same size process.
We demonstrate the modeling and experimental results on the modulation bandwidth and frequency chirp of the transverse coupled cavity vertical-cavity surface-emitting lasers (TCC-VCSELs). Both a 3-dB-modulation bandwidth enhancement and chirp reduction are shown in the results. These improvements could be explained by an increase in differential net gain and photon-photon resonance introduced by strong coupling in coupled cavities. Our TCC VCSEL could be useful for higher data rates and longer link lengths of single-mode fiber transmissions.