SHAP-1DCNN Based Detection of Analogue Operational Amplifiers

In the face of highly integrated and complex integrated circuits, traditional IC testing often relies on complex test designs and specialised test instruments, making it difficult to balance testing costs and efficiency. Deep learning algorithms can automatically learn and extract complex features from data, which can be adapted to various types of IC testing and help to realise the intelligence of the testing process. In this thesis, a SHAP-1DCNN based analogue operational amplifier testing technique (title) is proposed based on the actual test data of analogue integrated operational amplifiers (OPAs) using the SHAP-1DCNN deep learning algorithm.Firstly, the traditional CNN deep learning algorithm model is improved using SHAP analysis, and secondly, the general analogue integrated operational amplifier chip test circuit is drawn, and the precision components are patched and soldered to test the circuit board. Testing the key performance parameters of a large number of chips to obtain data sets. Train the optimised and improved SHAP-1DCNN deep learning model based on the collected data, and finally integrate this deep learning model in FPGA and accelerate its computation. The validation results show that the proposed test technique achieves a single OPA test time of 0.6s with an accuracy of 97.3%.