Research Of Space Mapping Technology And Its Applications In LTCC Circuit Designs

Traditional microwave passive circuit optimization techniques directly utilize the simulated responses or possibly available derivatives to force the responses to satisfy the design specifications. Circuit-theory-based simulation and CAD tools using empirical device models are fast: analytical solutions or available exact derivatives may promote optimization convergence but its accuracy can not be guaranteed. EM simulators has long been used for design verification and also be used for optimize process recently. However, the higher accuracy of the simulation, the more expensive direct optimization cost is expected to be. Although the solutions of EM simulators are accurate, it's difficult to be directly applied to complex EM problems because of the large CPU intensive. So, alternative optimization schemes combining the speed of circuit simulators with the accuracy of EM solvers are desirable. The recent exploitation of space mapping (SM) methodologies addresses this issue. SM is a novel optimization concept, allowing expensive EM optimization to be performed efficiently with the help of fast and approximate models called as coarse models. Through the construction of a mapping between the parameter spaces of the coarse and fine models, a suitable surrogate is obtained. This surrogate is faster than the fine model and at least as accurate as the underlying coarse model. The SM approach continuously updates the surrogate to approximate the corresponding fine model better.Firstly, the SM technique and basic concept have been reviewed and some early reported SM algorithms have been expatiated especially. We present some new topologies of components and devices, embedded passive circuits and interconnections with good EM properties which can meet development of microwave and millimeter wave LTCC technology and domestic state of arts and crafts inland. These algorithms are applied to optimization of LTCC circuits by using relevant production and applicable techniques in SM study, i.e. optimizing and designing various LTCC capacitors and inductances by original space mapping (OSM), LTCC transitions by aggressive space mapping (ASM) and equivalent lumped-element LTCC filters by knowledge-based aggressive space mapping (ASM) and neural space mapping (NSM).A novel SM algorithm is proposed to overcome some drawbacks of the reported SM, called characteristic component space mapping (CCSM). This method applies dynamic coarse models, i.e. equivalent circuit models obtained by synthesis. It is a simple and convenient SM method which has not high requirements for start design parameters and has avoided the parameter extraction (PE) process which is very complex and easy to fail. Because coarse models can accurately track the response results of fine models after each iteratation and avoid the instable problem due to mismatch between coarse models and fine models. This method improves the optimization efficiency greatly, and the equivalent circuit obtained from CCSM can provide effective help for LTCC passive circuit modeling. In order to demonstrate the procedure of the CCSM, CCSM optimization of LTCC capacitor was performed.At present, SM is only confined to some theoretical researches and a few experiments while without the relevant applicable tools at present. SM method performed only by manual is complex because of too many steps, data processing and intercommunication. To promote the methodology widely applied to actual circuit designs, a space mapping optimization system (SMOS) software is developed. The SMOS attempts to automate SM optimization through linking EM simulator and circuit simulator, providing help of optimization of complex passive components with multi-dimensional parameters.A modeling methodology presented to develop space mapping based enhanced models library for LTCC embed passive components using ADS Design Kit Tool. The enhanced models namely new integrated elements can then be used in the same way as the ADS built-in elements. In order to simplify model library development process, only two space mapping relationship OSM and NSM are adopted to develop enhanced models. Modeling examples of LTCC capacitor and equivalent lumped-element filter are demonstrated for the modeling process, whose results show well feasibility and performance of the models methodology.Finally, this thesis is concluded with suggestions for further research on SM method. The above researches on optimization and modeling of multi-layer circuits with SM method above are pioneer work in China and they have indicated that exploitation of SM optimization promises significant efficiency in LTCC engineering design.