Advertisement

Characterization and Variable Temperature Modeling of SiC MOSFET

  • Mengzhu Wang
  • Yujia Guo
  • Lei Wang
  • Guofu Chen
  • Ruichang Qiu
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 482)

Abstract

Silicon power semiconductor device is difficult to meet the requirements of high temperature, high pressure and high frequency. Among them, the MOSFET which has the fast switch speed and the simple driving circuit, become the most popular object in SiC power electronic devices. In this paper, we choose the C2M0160120D chip of CREE company, establishing a complete model. And the static characteristics of SiC MOSFET under different temperature points are simulated. The switching characteristics of SiC MOSFET under different driving resistances are analyzed and compared with the experimental results, and the accuracy of the model is verified in this paper.

Keywords

SiC MOSFET Simulation model Pspice Characteristic analysis 

1 Introduction

Si and GaAs, as the representative of the traditional semiconductor devices, can only work under 200 °C, and they can’t meet the new requirements of the development of modern electronic technology [1]. Since 1990s, with the outstanding performance advantages of band gap, breakdown field strength, thermal conductivity and saturation electron drift rate, the third generation wide band gap semiconductor material, represented by SiC and GaN, have become the research focus. At present, SiC MOSFETs have a very good application in the civil power substation and transmission field, the aerospace field, and the new energy field, such as PV inverter, hybrid/electric vehicles, rail vehicles, wind power [2].

With the wide use of SiC MOSFET, it is very important to use a model of the device to evaluate performance. Therefore, obtaining a precise and concise model is the key to simulate exactly of the device characteristics. The MOSFET device model can be divided into physical model and equivalent circuit model [3]. According to the structure diagram of the SiC MOSFET, the circuit schematic diagram and the physical equations of the model, a SiC MOSFET model based on the physical level was established in Ref. [4]. However, this method has a large amount of calculation. It is suitable for the research at the physical level and the analysis of the intrinsic characteristics of the device, but industrial applications. Taking SiC MOSFET CMF20120D chip of CREE company as an example, a variable temperature parameter PSpice model was built in Ref. [5]. The model was tested under different voltage, current and temperature conditions. It was turned out that the model was accurate and had a high reference value [6, 7, 8]. But the way, modeling of temperature controlled voltage source, is very complex and easy to make mistakes. In this paper, we use ABM (Analog Behavioral Modeling) simulation behavior model to improve the modeling method. Gate-drain capacitance CGD is an important factor to affect the switching characteristic of SiC MOSFET, and the method in Ref. [6] has a poor accuracy. In this paper, based on the establishment of the SiC MOSFET model of MOS3, we make a thorough inquiry of modeling CGD.

2 Variable Temperature Modeling of SiC MOSFET

Figure 1 shows the SiC MOSFET PSpice model that needs to be established in this paper. M and DBODY are used to describe the basic characteristics of N channel MOSFET with Model Editor in PSpice software. The temperature dependent voltage source ETEMP, is employed to describe the static characteristics of SiC MOSFET. CGD and CGS are used to describe the dynamic characteristics.
Fig. 1

PSpice model of SiC MOSFET

2.1 Temperature Compensation Modeling of on-State Resistance

Basic unit M only sets up a model at a single temperature (25 °C). In order to improve the accuracy of the model, we will build a model of temperature compensation of on-state resistance Rds. From the datasheet of C2M0160120D, we can know that the value of Rds increases with temperature. To simplify the modeling process, We combine the drain and source resistance into Rds(on), and we use the second-order fit method for its mathematical treatment
$$ R_{\text{ds(on)}} (T) = R_{\text{ds(on)}} (T_{25} ) \cdot \left[ {1 + TC_{1} \cdot (T - T_{25} ) + TC_{2} \cdot (T - T_{25} )^{2} } \right] $$
(1)
Through the data points extraction and curve fitting, we can get the formula of on-state resistance
$$ R_{\text{ds(on)}} (T) = 0.0024 \cdot \left[ {1 + 0.1309 \cdot (T - T_{25} ) + 0.0016 \cdot (T - T_{25} )^{2} } \right] $$
(2)
where Rds(on)(T25) is the typical Rds(on) value at 25 °C, we take 160 mΩ, T is the temperature point in simulations, TC1 and TC2 are fitting coefficients.

Place the temperature compensation resistor in the circuit, and test the Rds(on) of SiC MOSFET with the condition of UGS = 20 V,ID = 10A.

As is shown in Fig. 2b, the blue line that Model Editor default is far away from the trend of On-Resistance versus Temperature in datasheet. The red line can describe the on-state resistance with temperature changes better.
Fig. 2

a Test circuit of drain-source on-state resistance, b before and after the compensation of on-resistance versus temperature

2.2 Modeling of Temperature Control Voltage Source

In this paper, ABM (Analog Behavioral Modeling) is used to model the temperature control voltage source. ABM is an extension of the controlled source. It calls mathematical formulas or look-up tables to describe the device, without the need to design specific circuit [9].

In PSpice, the default threshold voltage change rate is −1Mv/°C, which isn’t match with the actual device threshold voltage variation. ETEMP is used to compensate for the change of threshold voltage caused by temperature change. It proposed in Ref. [10]. Simple linear fitting can’t perfectly represent the curve of Threshold Voltage vs. Temperature. So we use the three order function and three order fitting to the ETEMP modeling,
$$ E_{\text{TEMP}} = VT_{3} \cdot (T - T_{25} )^{3} + VT_{2} \cdot (T - T_{25} )^{2} + VT_{1} \cdot (T - T_{25} ) $$
(3)
where T is the temperature point in simulations, VT1, VT2 and VT3 are fitting coefficients. The ABM model used in this paper is simpler, and the improved formula is
$$ E_{\text{TEMP}} = VT_{3} \cdot T^{3} + VT_{2} \cdot T^{2} + VT_{1} \cdot T $$
(4)

2.3 Modeling of Gate-Source Capacitance

The capacitance between several poles affect the switching characteristics of SiC MOSFET. It is almost independent of temperature, but sensitive to voltage parameters. So, in this paper, temperature factors will not be considered, and mainly discuss the modeling of CGD which has a significant influence on the switching characteristics of the device. Two modeling methods of CGD will be explored in this paper.

2.3.1 Sub Circuit Modeling Method

Figure 3a shows the sub circuit of nonlinear capacitance, it uses two diodes in series to describe the nonlinear capacitance [11]. Diode has PN junction capacitance effect. In the reverse bias state, the capacitance decreases as the voltage increases, which conform to the changing trend of CGD. We can reasonably configure the parameters of two diodes to accurately simulate the changes in capacitance.
Fig. 3

a The sub circuit of nonlinear capacitance. b The test simulation curve of diode’s capacitance

When the device is in off-state, UGD < 0, DGD1 and DGD2, in series, are used to describe the changes in CGD. When the device is in the on-state, UGD > 0, fixed capacitance CGDMAX = CGD. In this way, the sub circuit can accurately describe the nonlinear variable capacitance CGD.

The waveform of the capacitance with voltage is shown in Fig. 3b. In the simulation process, the running speed is slow and the simulation curve is not easy to converge, which is prone to error. The capacitance value increases with the UGD value, which is far from the actual one. Diode parameters can only be set by trial and error, which has poor accuracy.

2.3.2 Descriptive Statement Modeling Method

CGD is nonlinear before the device is fully turn on, and it is a fixed value after the opening. It needs to describe the nonlinear variation of the capacitance value, which is expressed in Cg. Because measure capacitance directly is hard, referring to the formula (6), when dUGD/dt is linear to 1, the ig − t change can be used to replace the description of Cg − UGD change.
$$ i_{\text{g}} = C_{\text{g}} \cdot \frac{{{\text{d}}U_{\text{GD}} }}{{{\text{d}}t}} $$
(6)
Use the diode charge formula to derive the junction capacitance formula (7),
$$ C_{\text{JO}} = \frac{{Q_{\text{D}} \left( {m - 1} \right)}}{{\varphi_{\text{D}} \left[ {\left( {1 - \frac{{U_{\text{GD}} }}{{\varphi_{\text{D}} }}} \right)^{{ 1 {\text{ - m}}}} - 1} \right]}},m \ne 1 $$
(7)
Define \( a = \frac{{Q_{\text{D}} \left( {m - 1} \right)}}{{\varphi_{\text{D}} }} \), \( b = \varphi_{\text{D}} \), \( c = 1 - m \). Optimize the formula (7) so that the variable capacitor Cg only works when the CGD is less than zero. We can get the fitting function (8).
$$ \left\{ {\begin{array}{*{20}l} {C_{\text{GD}} = \frac{a}{{\left( {1 + \frac{{\left| {U_{\text{GD}} } \right|}}{b}} \right)^{c} }},} \hfill & {U{}_{\text{GD}} < 0} \hfill \\ {C_{\text{GD}} = a\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,,} \hfill & {U{}_{\text{GD}} > 0} \hfill \\ \end{array} } \right. $$
(8)

According to the fitting formula and using statement modeling method of Model Editor, we can build CGD sub circuit module.

Figure 4 shows the simulation results.
Fig. 4

Test simulation curve of sub circuit of CGD

As is shown,when UGD < 0, the capacitance value decreases with the increase of voltage. When UGD > 0, the capacitance value keeps constant. Therefore, this model can accurately simulate the nonlinear capacitance CGD in SiC MOSFET.

The above two models respectively use diode and voltage control current source to complete the modeling of CGD. There is a great deal of uncertainty in parameter setting by using diode modeling. So we select the second method by using VCCS to the modeling in this paper.

3 Characteristics Verification

3.1 Static Characteristics Verification

Compare the simulation results with the characteristic curves (solid lines) provided in datasheet as shown in Fig. 5.
Fig. 5

Compare of transfer characteristic between simulation and datasheet on 25 °C and 150 °C

Figure 6 shows the output characteristic of SiC MOSFET. The proposed model fits the datasheet well.
Fig. 6

Compare of output characteristic between simulation and datasheet

3.2 Dynamic Characteristics Verification

In this paper, the dynamic characteristics are verified by double pulse test (Fig. 7).
Fig. 7

The material object of circuit

The test circuit uses a double pulse gate drive mode. The drive voltage is −5/19 V. The drain-source voltage is 600 V. The load inductance is 5 Ω, 10 Ω, 20 Ω, 30 Ω. Observe the on-state and off-state simulation waveform of SiC MOSFET on different drive resistances, and compare to the experimental result.

Due to the fast switching speed of the SiC MOSFET, it requires wider band gap of current and voltage probes. We use the 100 MHz bandwidth voltage probe and 120 MHz current probe in this paper.

The test results are as follows,
  1. (1)
    The driving resistance is set to 5 Ω. The moment the device turn on (Fig. 8)
    Fig. 8

    The comparison between simulation and experiment when the drive turn-on and resistance is 5 Ω

     
The driving resistance is set to 5Ω. The moment the device turn off (Fig. 9).
Fig. 9

The comparison between simulation and experiment when the drive turn-off and resistance is 5 Ω

Above the pictures, the prior is simulation waveform. The latter is the experimental waveform. The blue line represents the change in drain-source voltage with time, and the red line represents the change in drain current with time.

From the simulation and experimentation results can be seen, the model built in this paper can fit the waveforms of current and voltage in the turn-on and turn-off time of SiC MOSFET C2M0160120D.

4 Conclusion

This paper focuses on the establishment of the simulation model based on SiC MOSFET C2M0160120D in Pspice. When modeling temperature-controlled voltage sources, we use the ABM module to simplify the modeling process. On the basis of the previous modeling methods, the model is further improved in this paper. Test the static characteristics of the model at different temperatures, the simulation results are in good agreement with the data provided in datasheet. Two models respectively use diode and voltage control current source to complete the modeling of CGD. After the comparison, we select the second method by using VCCS to the modeling in this paper. Build the test circuit by the dynamic model and test switching characteristics under different driving resistance. Compare the simulation result with experiment. It verifies the correctness of the dynamic model.

Notes

Acknowledgements

This work was supported by the Fundamental Research Funds for the Central Universities of China (No. E16JB00160/2016JBM062/2016JBM058) and The National Key Research and Development Program of China (2016YFB1200504-C-02).

References

  1. 1.
    Jiang W (2015) Development and application of silicon carbide power electronic devices. China High Tech Enterp 36:1–3 (in Chinese)Google Scholar
  2. 2.
    Wang L, Zhu P (2014) Overview of application of SiC power devices in power electronics. J Nanjing Univ Aeronaut Astronaut 46(4):524–532 (in Chinese)MathSciNetGoogle Scholar
  3. 3.
    Cheng S (2001) MOSFET modeling in electronic device simulation software. Hunan University (in Chinese)Google Scholar
  4. 4.
    Kraus R, Castellazzi A (2015) A physics-based compact model of SiC power MOSFETs. IEEE Trans Power Electr PP(99):0885–8993Google Scholar
  5. 5.
    Sun K, Wu H, Lu J et al (2013) Modeling of SiC MOSFET with temperature dependent parameters. Proc CSEE 33(3):37–43 (in Chinese)Google Scholar
  6. 6.
    Kumar K, Bertoluzzo M, Buja G (2015) Impact of SiC MOSFET traction inverters on compact-class electric car range. In: Power electronics, drives and energy systems (PEDES), Mumbai, India, pp 1–6Google Scholar
  7. 7.
    Ozdemir S, Acar F, Selamogullari US (2015) Comparison of silicon carbide MOSFET and IGBT based electric vehicle traction inverters. In: International conference on electrical engineering and informatics, Bali, IndonesiaGoogle Scholar
  8. 8.
    Cui Y, Chinthavali M, Tolbert LM (2012) Temperature dependent Pspice model of silicon carbide power MOSFET. In: Applied power electronics conference and exposition, Orlando, Florida, USA, pp 1698–1704Google Scholar
  9. 9.
    orctn101 Analog behavioral model using PSpice. www.cadence.com
  10. 10.
    Lu J, Sun K, Wu H et al (2013) Modeling of SiC MOSFET with temperature dependent parameters and its applications. In: IEEE applied power electronics conference and exposition—Apec, Long Beach, California, USA, pp 540–544Google Scholar
  11. 11.
    Zhao B, Zhou Z, Xu Y et al (2015) Study on the PSpice model of SiC MOSFETs applied in the electric vehicle (in Chinese)Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  • Mengzhu Wang
    • 1
  • Yujia Guo
    • 1
  • Lei Wang
    • 1
  • Guofu Chen
    • 2
  • Ruichang Qiu
    • 1
  1. 1.School of Electrical EngineeringBeijing Jiaotong UniversityHaidian District, BeijingChina
  2. 2.State Key Laboratory of Advanced Power Transmission TechnologyGlobal Energy Interconnection Research InstituteBeijingChina

Personalised recommendations