Research On The Method Of Modeling The Nonlinear Behavior Model Of GaN HEMT Power Device

In terms of high frequency,high temperature,and high power,Ga N high electron mobility transistors have advantages over other semiconductor devices and are therefore widely used in microwave circuit designs.If the circuit design needs to achieve better performance,the model of the selected device is required to be more accurate.Therefore,the small-signal and large-signal modeling of Ga N HEMT devices have become the current research hotspot.The traditional 14-element small-signal model(that is,an equivalent circuit model consisting of 14 circuit elements)has a large error at high frequencies,and the 20-element or 22-element model is more difficult to extraction parameters because of the large number of components.Therefore,this article considers comprehensively,based on the 14-element model,adds a parasitic capacitance to establish a 16-element small-signal model.For Ga N HEMTs model CGH40010 F,this paper first performs small signal modeling.By comparing the advantages and disadvantages of the two parameter extraction methods,the step-by-step extraction method is used to obtain the model parameters,and the parasitic parameters that are not related to the bias voltage and the intrinsic parameters related to the magnitude of the bias voltage are extracted.The obtained component parameters were substituted into the 16-element model circuit diagram of ADS for simulation optimization,and the final small signal model was obtained.Finally,the accuracy of the model is verified by comparing the error between the test data and the model simulation data.It shows that the established 16-element model can accurately reflect the small-signal characteristics of the device.The establishment of the small-signal model is based on S-parameter test data,but as the research progresses,the researchers found that when describing the large-signal characteristics of the device,the S-parameter cannot correctly reflect the circuit characteristics of the device or the network,thus proposing X Parameters and X parameter behavior model.Domestically,the modeling work for X-parameters has been carried out relatively late.More researches on the X-parameter behavioral model theory and improvement of the algorithm have been made.Therefore,it is worthwhile to establish an X-parameter behavioral model that can characterize large-signal operating characteristics based on the actual operating characteristics of the device.In this thesis,the mathematical formula of the X-parameter behavior model is deduced,and the physical meaning of the specific X-parameter item in the model is specifically described.There are two main methods for X-parameter extraction: nonlinear vector network testing and X-parameter production in ADS software.According to the existing experimental conditions,we use ADS to obtain the X parameters of the device.Because of the large amount of data,in order to improve the accuracy of the model,the theory of neural network modeling is used in this paper.The input signal power,frequency and bias voltage are used as the network input,and the X parameter item is the network output,A microwave device X-parameter model based on neural network was established using MATLAB.Finally,the established X-parameter model was imported into Simulink to verify the model.The accuracy of the model was proved by comparing the error of model output and test data.In this thesis,a 16-element model is built for Ga N HEMT devices.the parametric capacitance method is optimized.It is proved that the model can correctly reflect the small-signal characteristics of the device.Second,combined with neural network technology,an X-parameter model that can reflect the nonlinear characteristics of the device was established.The correctness of the model was proved by MATLAB.