Identifying the Source of BW Failures in High-Frequency Linear Analog Circuits Based on S-Parameter Measurements

摘要

High-frequency linear analog circuits are widely used in high-speed and wireless radio frequency communication circuits as front-end devices. The small-signal model of transistors is typically used to evaluate the bandwidth (BW) of such devices during design iterations. However, since such devices typically push the boundaries of the manufacturing processes, small-signal parameters are not entirely reliable, leading to systematic silicon failures due to inadequate BW. With increasing uncertainties in the modeling and processing of semiconductor devices, it is essential that the sources of BW failures be identified immediately once the devices are manufactured. This paper presents a methodology to diagnose the systematic BW failures in linear broadband analog circuits. The most important small-signal parameters of internal transistors are determined to enable the redesign process. An evolutionary algorithm specifically designed to mimic the expected errors is used to ensure fast convergence to the correct solution. Sensitivity analysis is used to determine the set of the most impactful small-signal parameters and to guide the evolutionary search. A weighed average approach is also used to improve the accuracy for large-scale systems. Experimental results indicate that the proposed algorithm determines the parameters accurately and scales well in terms of accuracy and computation time