High-Precision GaN-Based-SenseFET Design Based on a Lumped Parameter Electro-Thermal Network Model

The lossless and accurate current sensing technology is highly desirable for feedback control, fast over-current protection, and diagnostics-prognostics development for high-frequency and high-efficiency power systems. The SenseFET technology, where a current sensor is monolithically integrated with a power transistor, has been widely used in power ICs due to its high precision and low cost. However, for a gallium nitride (GaN) lateral power device in multi-finger configurations, the non-uniform temperature distribution hinders its application in high-precision scenarios. This paper aims to address this issue through a design method of SenseFETs based on a lumped parameter electro-thermal network (LPETN) model. Based on the proposed model, the time-dependent temperature and conduction current distribution are obtained, and the optimized finger selection for the accurate current sense is performed. The thermal network part of the model is validated by the finite element method (FEM) results, and the electrical part is validated through LTSPICE simulation. Finally, taking a 50-finger GaN high electron mobility transistor (HEMT) device as an example, this model is used to select the fingers of a SenseFET for current sensing. Compared with the traditional method, the proposed approach significantly improves the accuracy of the SenseFET, which demonstrates its effectiveness.