Use this URL to cite or link to this record in EThOS:
Title: Geometric and frequency scalable transistor behavioural model for MMIC design
Author: Minghao, Koh
ISNI:       0000 0004 5916 0674
Awarding Body: Cardiff University
Current Institution: Cardiff University
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis presents research in developing and validating scaling in terms of geometry and frequency for Behavioural models in order to extend their functionality. Geometric and frequency scalability, once thought to be limited only to Physical and Compact models, greatly reduces the number of measurements for model generation. Besides saving precious time and effort, measurements do not need to be collected at high frequency or power levels, reducing the cost of purchasing measurement hardware. Scaling in terms of geometry is achieved by combining accurate measurement based non-linear look-up table models of a reference (smaller) transistor with the appropriate passive embedding networks. Experimental results show that the scalable model is successful in predicting the performance of devices up to 5 times larger in gate periphery on two separate Gallium Nitride wafers, one measured at 5 GHz and another at 9 GHz. This approach provides a robust utilization of Behavioural models by providing performance predictions at power levels beyond the limitations of high frequency measurement systems. The geometric scalable Behavioural model was also used in a CAD environment to help create a prototype single cell MMIC amplifier for operation at 5 GHz. Although the targeted performance was not achieved due to mismatch, the non-linear Behavioural model is still able to predict the performance of the actual fabricated circuit. The work in this thesis also introduces the first formulation and approach that enables Behavioural models to be frequency scalable. The experimental results T MINGHAO KOH ABSTRACT IV on HFETs from 2 different Gallium Nitride wafers measured from 2 GHz to 8 GHz (2 octaves), support theoretical analysis that frequency domain Behavioural models defined in the admittance domain have frequency scalable coefficients. Load-pull results show that the model can accurately predict nonlinear behaviour at frequencies that were not used during the model extraction process.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: TA Engineering (General). Civil engineering (General)