Experimental study on the influence of junction temperature on the relationship between IGBT switching energy loss and device current

doi:10.1016/j.microrel.2017.11.023

Microelectronics Reliability

Volume 80, January 2018, Pages 134-143

https://doi.org/10.1016/j.microrel.2017.11.023 Get rights and content

Highlights

•
Extensive experimental study on the influence of junction temperature on turn-on and turn-off IGBT switching energy losses.
•
Experimental data show that both turn-on and turn-off switching energy losses are quadratic functions of junction temperature.
•
The coefficients of the above quadratic polynomials are impacted by junction temperature.
•
The IGBT switching loss in an inverter can be calculated easily based on the above coefficients at any operating condition.
•
The work enables precise estimation of device loss over wide ranges of link voltage, load current and junction temperatures.

Abstract

Accurate determination of power losses in semiconductor devices is important for optimal design and reliable operation of a power converter. The switching loss is an important component of the total device loss in an insulated-gate bipolar transistor (IGBT) in a voltage source inverter. The objective here is to study experimentally the influence of junction temperature on the turn-on switching energy loss E_on and turn-off switching energy loss E_off. More specifically E_on and E_off are both related to device current I_c; the influence of junction temperature on the relationship between E_on and I_c and that between E_off and I_c is studied. As the operating environmental conditions and load conditions of power converter vary widely, a wide range of junction temperatures between − 35 °C and + 125 °C is considered here. The experimental data enable precise determination of the switching loss in each device in a high-power converter at any practical operating condition. This leads to precise estimation of total device loss and optimal thermal design of the converter. This further helps off-line and/or on-line estimation of device junction temperatures required for study of thermal cycles and reliability.

Introduction

Power electronic converters play an important role in high-power and high-efficiency power conversion applications such as electric traction, renewable energy systems and energy storage systems [1], [2], [3], [4]. Silicon carbide metal-oxide semiconductor field-effect transistors (SiC MOSFETs) are increasingly used in power converters due to low switching losses and their ability to operate at high junction temperatures [5], [6], [7], [8], [9]. But the SiC device technology is still emerging; SiC devices of adequate current ratings are not yet available for power converters rated hundreds of kW or a few MW. Though integrated gate commutated thyristors (IGCTs) of suitable ratings are available for such power levels [10], [11], these devices require complex gate drive circuit. IGBTs are widely used in two-level converters at such power levels, since devices of adequate ratings are available [4], [5], [10], [11], [12], and also the gate drive requirement is simple [4], [12], [13].

As reliability of IGBT depends on maximum junction temperature and temperature cycling [1], [3], [14], [15], [16], [17], [18], reliability studies require precise estimation of device junction temperatures. This needs accurate thermal modeling of the complete thermal path, starting from the active regions of the device to the ambient [3], [12], [14], [15], [18], and the total amount of power loss in each device. A good estimate of the device power loss is also required for optimal thermal design of the converter.

To calculate the power loss in each device module, the normal practice is to calculate or estimate the individual loss components, namely, IGBT conduction loss, IGBT switching loss, diode conduction loss and diode switching loss. It is fairly straightforward to measure and/or estimate the IGBT and diode conduction losses based on the forward drop and device current [16]. The diode switching loss can be estimated based on the reverse recovery data available in the datasheet. While the reverse recovery loss is not very significant, the IGBT switching loss is a significant component of the total device loss [12], [16], and is also the most challenging one to measure or estimate. The device voltage and current change fast during switching transitions. Their variations with time depend on several factors.

There are numerous approaches for calculating the IGBT switching losses. The switching loss factor based approach [19], [20], [21], [22] is simple but not quite accurate [19] due to assumptions, such as, invariance of switching transition times with magnitude of current. Switching energy loss can be estimated with reasonable accuracy by dividing the total switching transition time into different sub-intervals, and calculating the energy loss in each sub-interval [5], [23]. However, this requires knowledge of the various sub-intervals (e.g. delay time, rise time, fall time) which themselves vary significantly with DC link voltage V_dc, device current I_c and device junction temperature T_j [24], [25]; but these transition times are available only for a few specific values of V_dc and T_j from the datasheet. Also, while switching energy loss at any junction temperature can possibly be evaluated based on a detailed physics-based model of the device, [14], [26], the calculation is complex due to a large number of parameters involved [26] and non-availability of these parameter values to application engineers. Further, external circuit parameters such as stray inductances and capacitances make the calculations more complex.

A widely employed method to calculate turn-on switching energy loss E_on and turn-off switching energy loss E_off is to consider their respective values (i.e. E_{on_test} and E_{off_test}) under test conditions (i.e. V_dc = V_test, I_c = I_test and T_j = 125 °C, say), and to scale these linearly with off-state blocking voltage as well as on-state current, [27], [28], as indicated by Eqs. (1), (2): $E_{on} = (E_{on_test}) \frac{V_{dc} I_{c}}{V_{test} I_{test}}$ $E_{off} = (E_{off_test}) \frac{V_{dc} I_{c}}{V_{test} I_{test}}$

Note that the above expressions in Eqs. (1), (2) assume that both E_on and E_off scale linearly with V_dc as well as I_c. A recent experimental study has focused on the relation between E_on and V_dc and that between E_off and V_dc [29]. According to the study, E_on and E_off are not exactly proportional to V_dc, but these vary in an incrementally linear fashion with V_dc. Further the relationship between E_on and I_c and that between E_off and I_c (at a given T_j) are very much non-linear [14], [30], [31]. Hence, calculation of losses using Eqs. (1), (2) could lead to a significant error. Further, the experimental study in [30] has shown that these relationships, at a given T_j, could be represented by quadratic functions with reasonable accuracy. A main objective of this paper is to study experimentally how these nonlinear relationships between switching energy losses and device current vary with T_j.

Recent works in [29], [30] have experimentally studied the scalability of E_on and E_off with V_dc and I_c, respectively. This work aims at studying the scalability of E_on and E_off with T_j based on experimental data. The device datasheets usually provide the switching energy loss, at most, for two different values of T_j, i.e. worst case T_j (e.g. at 125 °C or 150 °C) and room temperature (e.g. T_j = 25 °C) [31], [32]. Until recently, not many experimental studies were available in literature with regard to variations in switching characteristics with T_j. Extensive experimental data, namely, measured switching characteristics at different T_j, have been reported recently [24], [25], [29], [30], [33]. This paper utilizes the measured switching characteristics of a particular IGBT (termed device-1) reported in [25]. Further, the switching characteristics of another IGBT (termed device-2) are newly reported here. These experimental data on device-1 and device-2 are used to understand the relationships between E_on and T_j and between E_off and T_j. Both device-1 and device-2 are of punch through planar technology and both devices have the same voltage and current ratings.

Section snippets

Measurement of switching energy loss

The test circuit to measure the IGBT switching characteristics is a standard double pulse circuit [7], [24], [27], [28]. Fig. 1(a) shows the circuit diagram of the test set up which consists of one leg of the actual inverter. The same busbar arrangement is used for double pulse testing of both device-1 and device-2. This ensures that the stray circuit parameters are the same in both cases. During double pulse tests, the gate turn-on pulses are provided to the top device of the inverter leg. The

Measured turn-on and turn-off switching energy losses variation with junction temperature

The experimental procedure and setup explained in the previous section are used to study the influence of junction temperature on the relationships IGBT turn-on energy loss (E_on) versus junction temperature and turn-off energy loss (E_off) versus junction temperature.

The measured values of device-1 IGBT turn-on switching loss E_on at junction temperatures of − 35 °C, − 20 °C, − 10 °C, 0 °C, 25 °C, 50 °C, 75 °C, 100 °C and 125 °C are shown plotted in Fig. 3(a). Each line in the plot corresponds to a specific

Turn-on switching energy loss

The measured values of IGBT turn-on switching loss E_on against device current are shown plotted in Fig. 5(a) at various junction temperatures of − 35 °C, − 20 °C, − 10 °C, 0 °C, 25 °C, 50 °C, 75 °C, 100 °C and 125 °C, considering V_dc = 0.625 p.u. Fig. 5(b) and (c) present similar measurements at DC link voltages of 1.071 p.u. and 1.214 p.u., respectively. In each figure, the dots are the actual losses measured, which are connected by solid lines.

As seen from Fig. 5(a), E_on increases with I_c, but not

Variation of switching energy loss over a line cycle of an IGBT in a two-level VSI

For the purpose of calculation of switching energy loss, the line current in a VSI may be assumed to vary sinusoidally as indicated by Eq. (9). $I_{c} = I_{m} sin θ .$ where, θ is the fundamental angle. Using Eq. (9) and the relationships E_on versus I_c and E_off versus I_c, derived in the previous section, E_on and E_off can expressed as function of I_c for a particular value of V_dc and T_j.

Considering peak current I_m = 1 p.u., the variations in E_on with respect to θ at different junction temperatures, namely, − 35 °C,

IGBT switching power loss in a three phase two-level VSI

Assuming steady and balanced operating condition, the switching power loss (P_sw) in an IGBT in a three-phase two-level VSI leg can be expressed as, $P_{sw} = \frac{1}{2 π} \cdot f_{c} \int_{0}^{π} (E_{on} + E_{off}) \cdot dθ$ where, f_c is the carrier frequency. Substituting Eqs. (8), (9) into Eq. (10), P_sw can approximately be expressed as, $P_{sw} = \frac{f_{c}}{2} \cdot ((\frac{a + p}{2}) I_{m}^{2} + \frac{2}{π} (b + q) I_{m} + (c + r)) \cdot$

Total P_sw is evaluated for the range of I_m between 0.1 p.u. and 1.8 p.u. at V_dc = 0.625 p.u. for different junction temperatures, which is shown in solid lines in color (other than

Conclusion

A quantitative and extensive experimental study on the influence of T_j on turn-on switching energy loss (E_on) and turn-off switching energy loss (E_off) of an IGBT is reported. The study has been carried out for the device junction temperature in the range of − 35 °C to + 125 °C. The device current range is considered from 0.103 p.u. to 1.78 p.u., where 1.0 p.u. stands for the device rated current. Thus, the scope of the study includes overloaded conditions and extreme ambient temperatures.

The

References (36)

X. Ding et al.
Comprehensive comparison between silicon carbide MOSFETs and silicon IGBTs based traction systems for electric vehicles
Appl. Energy
(2017)
H. Luo et al.
Modern IGBT gate driving methods for enhancing reliability of high-power converters — an overview
Microelectron. Reliab.
(2016)
X. Du et al.
Lifetime estimation for IGBT modules in wind turbine power converter system considering ambient temperature
Microelectron. Reliab.
(2016)
S.C. Das et al.
Variation of Switching Sub-intervals of IGBT with Device Current and Junction Temperature: An Experimental Study
(Dec 2016)
S.C. Das et al.
Experimental Study on IGBT Voltage and Current Stresses During Switching Transitions
(Nov 2013)
Y. Kabalyk
Determination of energy loss in power voltage inverters for power supply of locomotive traction motors
Procedia Engineering
(2016)
N. Baker et al.
Improved reliability of power modules a review of online junction temperature measurement methods
IEEE Trans. Ind. Electron. Mag.
(2014)
S. Kouro et al.
Recent advances and industrial applications of multilevel converters
IEEE Trans. Ind. Electron.
(2010)
K. Ma et al.
Complete loss and thermal model of power semiconductors including device rating information
IEEE Trans. Power Electron.
(2015)
H. Abu-Rub et al.
Medium-voltage multilevel converters—state of the art, challenges, and requirements in industrial applications
IEEE Trans. Ind. Electron.
(2010)

N.L. Oswald et al.

An experimental investigation of the trade-off between switching losses and EMI generation with hard-switched all-Si, Si-SiC, and all-SiC device combinations

IEEE Trans. Power Electron.

(2014)

M. Rahimo et al.

Characterization of a silicon IGBT and silicon carbide MOSFET cross-switch hybrid

IEEE Trans. Power Electron.

(2015)

S. Hazra et al.

High switching performance of 1700V, 50A SiC power MOSFET over Si IGBT/BiMOSFET for advanced power conversion applications

IEEE Trans. Power Electron.

(2016)

J. Fabre et al.

Characterization and implementation of dual-SiC MOSFET modules for future use in traction converters

IEEE Trans. Power Electron.

(2015)

F. Filsecker et al.

Comparison of 4.5-kV press-pack IGBTs and IGCTs for medium-voltage converters

IEEE Trans. Ind. Electron.

(2013)

K. Ma et al.

The impact of power switching devices on the thermal performance of a 10 MW wind powerNPC converter

Energies

(2012)

A.D. Rajapakse et al.

Electromagnetic transients simulation models for accurate representation of switching losses and thermal performance in power electronic systems

IEEE Trans. Power Delivery

(2005)

D.A. Murdock et al.

Active thermal control of power electronic modules

IEEE Trans. Ind. Appl.

(2006)

Cited by (5)

Thermo-sensitive electrical parameters of high-power IGBTs based on gate-emitter voltage measurement during switching delay intervals
2023, Microelectronics Reliability
High-power insulated gate bipolar transistors (IGBTs) are widely used for wind power conversion and electric traction. Accurate estimation of IGBT junction temperature and active thermal control improve the reliability of power converters in such applications. Among the various available methods for junction temperature estimation, thermo-sensitive electrical parameter (TSEP) based estimation is simple, fast and amenable to real-time implementation. Evaluation of many TSEPs (e.g. forward voltage, threshold voltage, peak dv/dt, peak di/dt etc.) require measurement of multiple electrical quantities in the power circuit and/or in the gate-drive circuit. The measurement systems need to have high precision and/or high bandwidth as well. Moreover, the quantities in the power circuit require isolated measurement systems. This paper explores possible new TSEPs, that could be obtained through a single measurement of low-voltage quantity, namely, the gate-emitter voltage. Since the gate-emitter voltage varies primarily during turn-on and turn-off delay intervals, these switching intervals and their parameters are studied over a wide range of operating conditions. Since many practical converters are expected to operate at an ambient temperature as low as −35 °C, the experimental study is carried out over a wide temperature range from −35 °C to 125 °C. Further, to ensure wider applicability of the findings of the study, four IGBTs of different makes are considered for the experimental investigations. The experimental results bring out four different parameters pertaining to turn-on delay and another four parameters pertaining to turn-off delay, which are potential TSEPs. Of these eight TSEPs, five of them can be evaluated based on measurement of gate-emitter voltage only, and four of these have not been reported in the literature previously. In addition, the experimental data are useful for development of temperature-sensitive device models for simulation, performance assessment and design of gate-drive circuit for different temperatures, and for temperature-sensitive compensation of inverter dead-time effect.
BP neural network for non-invasive IGBT junction temperature online detection
2023, Microelectronics Reliability
Citation Excerpt :
Subhas Chandra Das also explored the relationship between switching energy and load current Ic. The experimental results show that the relationship between Eon, Eoff and Ic can be described as a quadratic polynomial [15]. Hui Chen Yang presents a method for estimating the junction temperature.
A number of studies have focused on Insulated Gate Bipolar Transistor (IGBT) junction temperature prediction methods. Some methods introduce extra circuits or sensors for an invasive prediction. However, the most of current methods hardly consider the situation that the junction temperature prediction may be affected by the IGBT degradation after long term operation. In this paper, the neural network using DC link voltage, switching frequency, load current and Negative Thermal Coefficient thermistor (NTC) temperature as inputs is proposed for junction temperature online prediction. The neural network training is completed through the IGBT module tests. The accelerated aging tests are then performed to validate the neural network results over lifetime. The static and transient working conditions are respectively validated at different aging degrees. Finally, a neural network with good accuracy could be achieved by a series of experiments considering IGBT device degradation.
Mine-used Flameproof Inverter Heat Dissipation Design and Its Performance Analysis
2023, E3S Web of Conferences
Strategy for Diagnosing the Aging of an IGBT Module by ON-State Voltage Separation
2019, IEEE Transactions on Electron Devices
A model of the on-state voltage across IGBT modules based on physical structure and conduction mechanisms
2019, Energies

View full text

Experimental study on the influence of junction temperature on the relationship between IGBT switching energy loss and device current

Highlights

Abstract

Introduction

Section snippets

Measurement of switching energy loss

Measured turn-on and turn-off switching energy losses variation with junction temperature

Turn-on switching energy loss

Variation of switching energy loss over a line cycle of an IGBT in a two-level VSI

IGBT switching power loss in a three phase two-level VSI

Conclusion

Appl. Energy

Microelectron. Reliab.

Microelectron. Reliab.

Procedia Engineering

Improved reliability of power modules a review of online junction temperature measurement methods

IEEE Trans. Ind. Electron. Mag.

Recent advances and industrial applications of multilevel converters

IEEE Trans. Ind. Electron.

Complete loss and thermal model of power semiconductors including device rating information

IEEE Trans. Power Electron.

Medium-voltage multilevel converters—state of the art, challenges, and requirements in industrial applications

IEEE Trans. Ind. Electron.

An experimental investigation of the trade-off between switching losses and EMI generation with hard-switched all-Si, Si-SiC, and all-SiC device combinations

IEEE Trans. Power Electron.

Characterization of a silicon IGBT and silicon carbide MOSFET cross-switch hybrid

IEEE Trans. Power Electron.

High switching performance of 1700V, 50A SiC power MOSFET over Si IGBT/BiMOSFET for advanced power conversion applications

IEEE Trans. Power Electron.

Characterization and implementation of dual-SiC MOSFET modules for future use in traction converters

IEEE Trans. Power Electron.

Comparison of 4.5-kV press-pack IGBTs and IGCTs for medium-voltage converters

IEEE Trans. Ind. Electron.

The impact of power switching devices on the thermal performance of a 10 MW wind powerNPC converter

Energies

Electromagnetic transients simulation models for accurate representation of switching losses and thermal performance in power electronic systems

IEEE Trans. Power Delivery

Active thermal control of power electronic modules

IEEE Trans. Ind. Appl.