FSRFT—Fast Simplified Real Frequency Technique via Selective Target Data Approach for Broadband Double Matching

摘要

This brief introduces a broadband double-matching (DM) solver called fast simplified real frequency technique (FSRFT). FSRFT is essentially a greatly accelerated variant of the well-known classical simplified real frequency technique (SRFT). The basic idea that turns the classical SRFT into a “fast” SRFT relies on two main approaches: the selective target data approach (STDA) and the constraint optimization approach (COA). STDA constructs an optimization target data set formed of only critically selected target data whose element number is equal to or slightly greater than the order of the system unknowns n plus 1, n+1. In order to exhibit speed performance comparison between SRFT and FSRFT, an example design is considered. An exemplary DM problem, dealing with an n=6th order low-pass Chebyshev-type equalizer design to match the given generator and load impedances, has been solved by SRFT within 29 s using 90 target data in a typical computer—e.g., Intel 2.20-GHz i7 CPU with 8-GB RAM. On the other hand, the same problem has been solved by the newly proposed FSRFT within only 0.6 s using only n+1=7 critically selected target data in the same computer. FSRFT introduced herein works in any domain, i.e., lumped, distributed, and mixed.