A Study On The Parameter Extraction Of Bsim3v3Model Based On A High-voltage-compatible CMOS Process

The Smart Power Integrated Circuit (SPIC) is widely used in fields such asautomotive, household appliances and industry control, et al. It’s development couldsignificantly increase the reliability and intelligentialize of the system or reduce theweight and cost of the machine.In the subject as this thesis refer, a novel SPIC has been designed. The high voltagedevices and the low voltage circuits should be integrated into one chip when SPIC isbeing fabricated, therefore the isolation technology should be carefully considered.According to the patent technology of Optimized Variational Lateral Doping made byProf. Xingbi Chen, it is possible to design a new CMOS-compatible process, which haslower cost of fabricate, less difficulty to use, and could make the SPIC with betterperformance. Based on this process, this thesis make a study of device model and modelparameter extraction of low voltage circuits in SPIC.In the beging of this thesis, the development and the research achievement ofpower electric and SPIC is reviewed, the importance of the isolation technology is beingexplained briefly. After that, a discuss is made on the device model, pointing out thatthe device model is the bridge of IC design and fabricate, and that the model parameteris closely linked to the fabricate process or technology.Then, the new CMOS-compatible process which is used in this subject is beingintroduced. After a discuss of the theory of the Bsim3v3MOSFET model, the flow pathof the model parameter extraction is made based on the test data of the fabricated MOSdevice. Then make a work on the model parameter extraction using the tool namedUTMOST IV.In the end, this thesis make a discuss on the algorithm of the extraction, pointingout that with the increasing of the complexity of device structure, using the traditionalway to extract model parameter is being more and more difficult, and genetic algorithmis good enough to be a alternative algorithm. After make a review of the theory of thegenetic algorithm, this thesis discussed how to achieve the extraction using the geneticalgorithm. At last, give a try on the Bsim3v3model parameter extraction using genetic algorithm based on the same batch of devices.