Design Of GaN Servo System Based On FPGA

The diversification of application scenarios puts forward higher demands on the control performance and response of the servo system.It is difficult to increase the switching frequency because traditional silicon-based devices suffer problems such as heat dissipation and loss issues,which gravely restricts the promotion of response of the servo system.The breakthrough development of wide bandgap semiconductors makes it possible to design servo drives based on GaN power devices,which can further improve the dynamic response of the control loop while achieving high switching frequency.However,the switching cycle of the servo system under high switching frequency is extremely short,and the traditional digital control chip cannot complete all complex control algorithms in a single cycle.FPGA can meet the complex calculation requirements of servo system under high switching frequency due to its high main frequency and parallel execution characteristics.At the same time,it has the advantages of semicustom hardware circuits,which can provide an experimental basis for the realization of motor control ASIC chips.Therefore,it is momentous to research the design of FPGA-based GaN servo system.The analysis of servo system control theory is completed and a PMSM mathematical model is given.Theoretical derivation results regarding to the parameter selection of the loop controller under vector control strategy points that inertia online identification is crucial to implement the parameter tuning of the speed loop controller.Furthermore,an inertia identification algorithm is proposed derived from the least squares of the reduced-order forgetting factor.The algorithm requires limited calculation resources,which is simple for digital implementation.Fast and accurate identification of parameters is achieved in both offline and online modes.Secondly,a servo system based on GaN power devices is designed.The dv/dt and di/dt problems of GaN drive at high turn-off speed are analyzed.It is pointed out that reducing the gate loop impedance and optimizing the parasitic parameters of the hardware layout are useful methods to decrease the crosstalk.On this foundation,the software design of the servo system was accomplished in FPGA,which effectually reduced the consumption of hardware resources through time-sharing multiplexing,and applied timing optimization technique to improve the main frequency of the control algorithm to 203.05 MHz.Lastly,the proposed scheme was validated in specific experiments.Based on double pulse test,the validation of driving circuit is carried out,where the switching characteristics of GaN devices under extreme operation conditions is tested.The current loop bandwidth of servo system using GaN power devices as switching elements can be increased to 5KHz at 100 KHz switching frequency,which is verified in the closed-loop and open-loop experiments.In addition,the experimental results of inertia identification algorithm indicate that the identification convergence time in online mode is only 2.2s,and the relative error of identification in steady state is only 1.43%.