EMI investigation and modeling of a flat panel display

It is often important to carry out EMI analysis in the design phase of an electronic product to predict the radiated emissions. An EMI analysis is important to predict if the product complies with the FCC regulations as well as to gain an understanding of the noise coupling and radiation mechanisms. EMI analysis and prediction of radiated emissions in electronic products that have an electrically large chassis, pose a challenge due to the presence of multiple resonant structures and noise-coupling mechanisms. The study focusses on the investigation of the main noise coupling mechanisms, the approach and methods used for the modeling of a flat panel display. Full-wave simulation models are a powerful tool for the prediction of radiated emissions and the visualization of coupling paths within the product. The first part deals with the measurement of radiated emissions from the display under standard test conditions and the identification of the main noise sources using near-field scanning. The contribution of the chassis components - frame, back cover and the back panel, to the radiated emission is analyzed using shielding measurements. Noise coupling from the main board, flex cables, display driver boards and the display is analyzed from measurements. The second part deals with the full-wave modeling of the components - main board, flex cables, chassis and the display driver boards. The modeling approach is demonstrated by highlighting some of the challenges in modeling larger structures having many details. The simulation model contains the main components of the TV that contribute to far-field radiation. The full-wave modeling is done using the CST Microwave Studio. Two sets of simulation models are described - the common mode models and the complete models. The use of the common mode models for the identification of the resonant structures is demonstrated. The far-field radiated emissions along with the coupling mechanism within the flat panel display can be predicted using the simulation model. --Abstract, pag