Integrating a large number of embedded memories in System-on-Chips (SoC's) occupies up to more than 70% of the die size, thus requiring Built-In Self-Test (BIST) with the smallest possible area overhead. This paper analyzes MATS++(6N), March C-(10N), March SR(14N), and March CL(12N) test algorithms and shows that they cannot detect either Write Disturb Faults (WDFs) or Deceptive Read Destructive Faults (DRDFs) or both. Therefore to improve fault detection, an automation program is developed based on sequence operation (SQ) generation rules. However after solving the undetected fault, the outcome in term of its detection result of Static Double Cell Faults using the specified test algorithm especially Transition Coupling Faults (CFtrs), Write Destructive Coupling Faults (CFwds), Read Destructive Coupling Faults (CFrds) and Deceptive Read Destructive Faults (CFdrds) are observed.
In this study, a six‐track standard cell library with a multi‐finger layout structure is proposed to improve the delay, energy, and area of the digital circuit design for near‐threshold operation. The proposed library is optimised by using the parasitic effects of the technology and optimising the layout. To enhance the performance and energy efficiency, inverse narrow width effect has been considered in the design, whereby the minimum width of the process was used as the based width unit. To minimise the design area, the standard cell was designed in the lowest possible height with a multi‐finger layout structure. The proposed library with a few basic cells was developed and characterised in 130 nm technology, which is available for synthesis and automatic place‐and‐route (P&R). The proposed library was analysed and compared with two eight‐track multiplier layout libraries in the cell and block design level. Based on the place‐and‐route results of ISCAS'85 benchmark circuits, the proposed six‐track library could achieve up to 27% of delay improvement, 29% energy reduction and 44% area reduction as compared to the multiplier structure library at the minimum critical path delay.
Carbon nanomaterials have been magnificently high demand in the development of nano-electronic technologies since the past few years because of the splendid performance in electrical, mechanical and physical properties. This research is to prepare the dispersion of functionalized carbon nanotube in six different types of organic solvent as the pre-preparation of conductive carbon nanotube nano-ink. The focus of this research is in the study of the morphology of dispersion sample by Raman Spectroscopy data analysis, discussing on conductivity and physical properties of dried samples. From this study, it shown that dispersion with methanol and propanol are stable compared to other samples. The morphology of dispersed functionalized carbon nanotube plays an important role in the quality and performance of the nano-ink in the production of sensing material of flexible biosensors and electronics circuit for biomedical and other potential applications. Therefore, it shows that the organic solvents are suitable in dispersion of MWCNT as preparation of conductive nano-ink.
Design-for-test (DFT) in an integrated circuit is one of essential parts in System-on-Chip. DFT enables testing and debugging of an integrated circuit before it is being produced in high volume. Due to increasing of functionality in advanced nodes of integrated circuit designs, DFT is imperative in reducing defect counts and improving performance of the integrated circuits before reaching the customers. Thus, many research have been done in DFT area in achieving an optimum performance of integrated circuits. Scan test is one of the DFT techniques that enable the integrated circuit design to be tested and debugged. However, due to additional components are being inserted to improve the controllability and observability, high power consumption and dissipation is expected. In this paper, an evaluation of optimum scan chain parameter with respect to its power performance of an integrated circuit will be performed. A block of a microcontroller unit, CORTEXM0DS will be used to where by the scan cells are inserted using DFT Compiler and TetraMAX ATPG from Synopsys. Scan chain from 2 to 20 chains with increment of 2 is simulated and the test power is obtained. The simulation result shows that scan chain with 16 chains count in the design contributed the highest test power of 15.5409 mW while scan chain with 10 chains count result in the lowest test power of 15.2842 mW. The optimum scan chain will also consider the number of test coverage and test pattern.
Imbalance and out-of-range input signals can cause inaccuracy in fully differential successive approximation (SAR) analog to digital converter (ADC). Therefore, implementation of an ADC driver can solve the problem since the input can be properly adjusted to suit with an ADC input. AD8139 single to differential amplifier was chosen as an ADC driver in this design and placed on a printed circuit board (PCB) to drive differential input signal of SAR ADC. The result shows each of output amplitude of the amplifier remains equal and is 180° out of phase for DC and AC input signal. The fabricated 10 bit SAR ADC is capable to digitize full code from analog input produced by the ADC driver.
9 September 2026 in Journal of Intelligent Media Computing
Object detection models trained on static images may perform poorly when deployed on continuous video, creating challenges for automated industrial inspection. This study investigates whether supplementing static training data with consecutive video frames can improve real-time halal compliance detection in poultry processing. Two YOLOv11l models were trained for 100 epochs under identical hyperparameters. A total of 1,835 video frames were extracted and annotated, with 1,479 allocated to training, 182 to validation and 174 to testing. The baseline model used 1,400 static images, while the proposed model used 2,868 images comprising 1,389 static images and 1,479 video frames. Both models were evaluated on an independent video containing 1,820 frames and 2,439 annotated instances. Although validation performance was similar, video evaluation showed a clear difference. The proposed model increased recall from 49.77% to 60.07% and AP@0.5 from 45.74% to 58.37%, while precision remained comparable at 83.14% and 83.21%, respectively. The proposed model recovered 251 additional structures at the cost of 52 additional false positives and achieved a mean inference time of 9.2 ms per frame. Error analysis showed that most remaining missed detections occurred under severe motion blur, blood coverage, partial visibility or unfavourable orientation. These findings indicate that video-supplemented training improves deployment performance without architectural modification, while human verification remains necessary for final certification.