Prediction of Operating Temperature of Electronic Components on Printed Circuit Boards

Abstract

The current study seeks to provide a perspective on the capabilities of the Laplace finite difference (LFD) method to predict operating temperatures of components of electronic systems. This paper presents a systematic assessment of the predictive accuracy of thermal analysis using LFD code for heat transfer of components mounted on a printed circuit board (PCB). The numerical algorithm is implemented in MATLAB that provides an estimate of the computational effort. The analysis methodology is presented using a case study. Using this methodology, a large calculation time reduction is achieved without losing accuracy. Results show that the predicted values were less than the given limit of 9 degrees Celsius. Accuracy of predicting component operating temperature depends on the component location on the PCB, air-flow velocity and flow model applied. Experimental measurement of component junction temperatures is recommended to enable a comparison between simulated values and actual measurements. The error estimated from the difference between measured temperature values and simulated values can help in strategic product design decisions and reliability predictions. The error estimate converges to zero after about 7 iterations as the predicted temperature converges to the theoretical temperature. Thus, the temperature profile is accurately predicted with the help of the proposed LFD-based algorithm.