Fault Diagnosis in Advanced Analog Building Blocks Using Ensemble Stacking Technique

Detecting faults in large integrated electronic building blocks is challenging due to the intricate operating mechanisms. As electronics evolve, diagnosing faults in complex analog systems becomes harder. Hence, it is crucial to promptly and precisely detect problems in analogue circuits. This study introduces a new method, ensemble stacking, to improve fault diagnosis accuracy in advanced analog block i.e., cascode and differential amplifier stages. We also showed that CatBoost outperforms other classifiers such as decision trees, logistic regression, and SVM with 96% accuracy on a cascode amplifier and 97% accuracy on a differential amplifier circuit. However, ensemble stacking achieves 100% accuracy for the first test circuit and 99.95% for the second, enhancing precision and reliability in fault diagnosis for analog circuits.