Xiang Wang, Chongchong Zhu, Haoze Luo,Wuhua Li, and Xiangning He

(College of Electrical Engineering, Zhejiang University, Hangzhou, 430074, China)

Elimination of Collector Current Impact in TSEP-based Junction Temperature Extraction Method for High-Power IGBT Modules

Xiang Wang, Chongchong Zhu, Haoze Luo,Wuhua Li, and Xiangning He

(College of Electrical Engineering, Zhejiang University, Hangzhou, 430074, China)

Insulated gate bipolar transistor (IGBT) modules are widely employed in high-power conversion systems. Their junction temperature ranks as one of the most important factors in the reliability of power semiconductor devices. Thermo-sensitive electrical parameter (TSEP) is regarded as the promising solution to extract the junction temperature due to its non-invasion measurement, fast response and high accuracy. However, accurate collector current measurement is required if only the individual TSEP is adopted, which increases the complexity and cost. In this paper, the combined TSEP method is proposed to eliminate the influence of collector current (IC), where the turn-off delay time (tdoff) and maximum decrease rate of IC(max dIC/dt) are adopted and combined. The two TSEPs both have linear relationships with junction temperature and IC. When they are combined mathematically, the influence of ICis eliminated. Experiments have been implemented to validate the effectiveness of the proposed approach. The comparison between combined TSEP and two individual TSEP methods are illustrated and analyzed.

Junction temperature, thermo-sensitive electrical parameter, non-invasiveness, online application

1 Introduction

High-power IGBT modules are widely used in various power conversion fields, such as the high voltage direct current(HVDC) transmission and renewable energy system. The reliability of high-power IGBT modules is important for the safe operation in power electronic applications[1-2]. It is shown that the power electronic device ranks the most fragile component in an industrial survey[3]. In addition, the junction temperature is regarded as the first indicator of deterioration and nearly 60% of power device failures are thermally induced[4]. It is noted that the knowledge of junction temperature is essential for effective health management of power converters, enabling safe operation of the power semiconductors under complex operating conditions[5]. In general, the junction temperature is extremely useful, not only in the design of a converter system to ensure optimal reliability and operation, but also in the control of systems that present redundancies[6].

However, it is not easy to extract the thermal information of the chips from semiconductor devices. Several schemes have been developed to monitor the junction temperature, which include optical methods, physical contact methods and electrical methods[7]. However, the optical methods have the disadvantages of requiring professional equipment and being expensive. Moreover, the chip inside the device should be exposed, which means the package should be opened and makes it impossible to be applied in practicaloperating conditions. The physical contact method demands physical contact inside the package where an embedded thermistor is usually mounted on the substrate near the die to measure the junction temperature. Unfortunately, fast dynamic temperature changes can’t be followed.

Nowadays, the electrical methods are usually used in junction temperature extraction[7]. Academically, the electrical characteristics related to junction temperature are described as the thermo-sensitive electrical parameter (TESP). According to [6], the junction temperature measurement using TSEP is regarded to be the most promising way to conduct online temperature measurement on fully packaged devices. Complicated auxiliary measurement equipment is not required for the evaluation of junction temperature. Therefore, special focus has been given to the research of TSEP and many approaches have been presented to realize the online junction temperature monitoring[8-9].

Among the TSEPs, the turn-off delay time tdoffis considered to have the potential possibility to achieve online junction temperature extraction, because it’s of high linearity and fixed sensitivity[10]. With similar features, the maximum decrease rate of collector current (max dIC/dt) during turn-off transition is also a strong TSEP candidate[11]. In the aforementioned work, the temperature dependencies and performance of these two TSEPs are discussed and their advantages are carefully analyzed. However, accurate collector current measurement is needed, and the collector current is usually varies over the operating conditions.

In the high-power operating condition, expensive auxiliary equipment is needed for the accurate current measurement. To solve the problem, this paper proposes a combined TSEP method to evaluate thejunction temperature where tdoffand max dIC/dt are combined. After the introduction of the principle of these two TSEPs, section 2 presents the elimination impact of collector current with the combination

of tdoffand max dIC/dt. Then, the experiments are implemented to verify the effectiveness of the proposed solution. Finally, three mentioned TSEP methods are compared and their advantages and disadvantages are illustrated.

2 TSEP candidates in high-power IGBT turn-off transition

2.1 Features of High-power IGBT

The classic planar IGBT structure is illustrated in Fig.1. The structure of IGBT is dominated by a wide lightly doped n-drift region, sandwiched between a p+ emitter and p- well/MOS gate region. An n- channel is formed to conduct the carriers when the gate voltage is applied.

2.2 IGBT turn-off transition

In this paper, the turn-off delay time (tdoff) and the maximum rate of collector current during turn-off transition (max dIC/dt) are discussed as the TSEPs about the influence of collector current IC. These two TSEPs are both in the IGBT turn-off transition and the relative definitions are illustrated in Fig.2.

Three stages of the typical IGBT turn-off transition are described as follows:

Fig.1 Basic IGBT structure

Fig.2 Typical turn-off transition of IGBT

Stage 1 [t0～t1]: before t0, the inversion layer spreads across the p-based region, which enables the current carrier travelling across the emitter and collector. At t0, the electrons accumulated in inversion layer start to be extracted out. At this stage, the gate voltage stays at the Miller plateau to keep the collector current. More and more electrons are extracted from the inversion layer until it pinches off.

Stage 2 [t1～t2]: the pinch-off region extends, resulting in the rising of the collector-emitter voltage (VCE). As the electrons are still extracted from the inversion layer, the gate voltage stays at the Miller plateau and the collector current remains as before. This stage stays until the MOS channel is fully pinched off, resulting in the falling of the collector current IC.

Stage 3 [t2～t3]: the MOS channel is fully pinched off, and the collector current is totally sustained by the storage carrier extraction mechanism in the n-base region. This stage stays until the extraction stops and the collector current drops to zero.

As for the two chosen TSEPs, tdoffdefines the delay time from the starting of the change of the driver voltage to the decreasing of the collector current, while max dIC/dt defines the maximum decreasing rate of ICin the aforementioned stage 3. From the definition, the TSEP value can be measured when ICis monitored. However, the current measurement is usually a complicated task and extra equipment or sensors arerequired. Taking advantage of the special structure of package of high-power IGBT modules, the measurement of TSEP can be made easier.

2.3 Measurement of TSEP

Designed for the high-power operating condition, high-power IGBT modules have two kinds of terminals. One is designed to connect the power circuit, called the power terminal. Another is designed to connect the driver circuit, named the drive terminal. In this paper, the Infineon IGBT moduleFZ3600R17HP4_B2 rated at 1.7kV/3.6kA is taken as the example to show the measurements of TSEP. The package of the module is shown in Fig.3. “C” and “E” in Fig.3 are the power terminals and the driver terminals include“g”, “e”, “c”. In this paper, the power emitter refers to“E” while drive emitter refers to “e”.

The equivalent circuit of high-power IGBT module is shown as in Fig.4.

Fig.3 Infineon IGBT moduleFZ3600R17HP4-B2 terminals

Fig.4 Equivalent circuit of high-power IGBT module

It is illustrated that, there is parasitic inductor LeEbetween the power emitter and the driver emitter. LeEconsists of Lkeand LkE. Lkeis the parasitic inductor in the drive circuit part between the power emitter and the driver emitter while LkEis in the power circuit part. The voltage of the two terminals can perfectly reflect the two TSEPs as is shown in Fig.5.

In Fig.5, it is illustrated that the turn-off delay time (tdoff) is equal to the time between the two voltage wave crests, and the maximum decreasing rate of ICis equal to the negative peak value of LeEvoltage.

With the help of LeE, the maximum decreasing rate of IC(max dIC/dt) has a linear relationship with the negative peak value of LeEvoltage, as shown in (1).

From (1), it is suggested the max dIC/dt is linear to VeEPeak. Therefore in the following parts, VeEPeakis discussed instead of the max dIC/dt for the simple analysis.

3 Proposed combined TSEP method

Some work has been done to discuss the possibility of the junction temperature extraction via different TSEPs[6-7]. The effectiveness of TSEP is evaluated in the aspects of sensitivity, linearity, generalization, calibration, online implementation and so on.

Fig.5 TSEP’s reflection on LeEvoltage

In [10] the influential factors of the turn-off delay time are discussed. Table I summarizes the turn-off delay time dependence on three operating conditions. With increasing of Tjand VCE, tdoffincreases monotonically. Conversely, tdoffgets shorter if ICincreases. In addition, the turn-off delay time tdoffbased TSEP has a relatively fixed sensitivity compared with the static TSEPs like forward voltage (Vf) whose sensitivity ranges from 3.5mV/℃ at IC=1200A to 0.5mV/℃ at IC=200A in the tested IGBT module[10]. Owing to its advantages, the tdoff-based TSEP is a practical approach to monitor the junction temperature.

It can be summarized that the turn-off delay time tdoffcan be calculated by a fitting function of Tj, ICand VCE, which is shown by

According to the analysis in [10], the junction temperature can be derived with the monotonical relationship between tdoffand Tjas long as ICand VCEare known.

In [11], it is proposed that VeEPeakcan also be a potential sensor to indicate the junction temperature. Theoretically, the temperature dependence of the VeEPeakis discussed, as well as the influence of other operating conditions: ICand VCE.

VeEPeaktrends over different operation conditions are listed in Table 2. The value of VeEPeaktrends to increase when ICor VCEincreases, while it trends to decrease if the junction temperature rises. In conclusion, VeEPeakcan also be calculated by

From (3), the junction temperature can be calculated if the other three parameters are known and the calibration is done.

According to the analysis above, the influence of collector current (IC) and collector-emitter voltage (VCE) makes it impossible to use the methods without knowing the two parameters. If there is any way to eliminate the unwanted influence from ICor VCE, it would be more realistic to be applied in practice.

The influence of the collector-emitter voltage is not so important in the practical system compared with the collector current, because in practical applications, such as the voltage source converters, the collector- emitter voltage stays constant and equal to the DC bus-bar voltage. Consequently, VCEcan be considered as a constant parameter, and therefore (2) and (3) can be simplified as below:

Table 1 Turn-off delay time trends over operation conditions

Table 2 VeEPeaktrends over operation conditions

To take it a step further, from the discussion in [10] and [11], it is suggested that the junction temperature dependence of both turn-off delay time and the maximum collector current rate is of high linearity. To simplify the relationship between TSEPs and the influential factors, (4) and (5) can be described as the first-order equation:

To describe the relationship in matrix format, it can be summarized as below:

From (8), it can be calculated that:

Calibration is needed in the proposed combined TSEP method. During the calibration process, the value of Tj, IC, tdoffand VeEPeakare measured. The relations between TSEP (tdoffand VeEPeak) and Tj, ICare linearly fitted. After that, the matrix A and B can be extrapolated. In practical operation conditions, the junction temperature Tjcan be calculated by these TSEP when A and B are applied into (10).

4 Experiment verification

A double-pulse test platform is built to verify the effectiveness of the combined TSEP and its results are shown in Fig.6.

In the test platform, with a temperature-controllable heating board, the junction temperature can be considered the same as the stable base plate temperature. For a fixed junction temperature, the relationship between TSEPs and temperature can be investigated. Under a certain bus-bar voltage (800V), different collector current (IC) can be generated. In addition, an oscilloscope is used to get the waveform of VeE. From the waveform of VeE, the two chosen TSEPs, tdoffand VeEPeakcan be obtained. The experiment parameters are listed in Table 3.

Fig.6 Double-pulse test platform

Table 3 Experiment parameters

The experimental results of the TSEPs’ value under different junction temperature and collector current are shown in Fig.7 and Fig.8 and their relationship is linearly fitted.

Applying the experiment results in (8), the coefficient matrix can be reckoned as follows.

After the coefficient matrix A and B are calculated, the impact factor can be reckoned.

Fig.7 Linear fitting of relations between tdoffand temperature and IC

Fig.8 Linear fitting of relations between VeEPeakand temperature and IC

The measured junction temperature Tjcan be extracted from (12):

The accuracy of the proposed method is verified as illustrated in Fig.9, where the statistical comparison of measurement error of three kinds of methods is given.

In Fig.9, the accuracy comparison of each method is clearly displayed, verifying that the efficiency of the combined TSEP method is almost the same as tdoffbased method. No extra measurement error is introduced using the combined TSEP methods, while the drawback of accurate current measurement is avoided.

5 Comprehensive evaluation

In this section, the three methods were compared in the following five parts: application, accuracy, measure complexity, calibration complexity, current measure requirement. The comprehensive evaluation of the three methods was illustrated in three radar graphs, which were plotted in Fig.10.

Fig.9 Statistic comparison of measurement error of three methods

Fig.10 Comparison of three methods

Radar graph illustrates the advantages and disadvantages of each method clearly: in one aspect, the further it extends from the center, the better it performs over other methods. In the application aspect, the method which can be applied to wider conditions gets further from the center. As for accuracy, the method which is more accurate reaches further. Easy measured method extends longer in the aspect of measurement complexity. Similarly, easy the calibrated method gets further from the center. For current measure requirement, the method which requires less for current measuring reaches further from the center.

In the aspect of application, the three methods are the same because the TSEPs used in three methods are all extracted during the turn-off transition.

In terms of accuracy, from Fig.9, it is suggested that the tdoff-based method ranks as the most accurate method, followed by combined TSEP method. The VeEPeakbased method has the biggest measurement error.

As for measurement complexity, accurate time measurement is required in both tdoff-based method and the combined TSEP method, which adds the complexity of applying the method in operating condition. VeEPeakmeasurement is much easier compared with tdoff.

Calibration is necessary for all three methods, but combined TSEP method is the most complex method. More calculation is needed to get the measurement equation. VeEPeak-based method and tdoff-based method are the same in this respect.

The advantage of the combined TSEP method is that no current measurement is required in the method. As for the other two methods, VeEPeak-based method requires more accurate ICmeasurement than tdoffbased method. From (10), it can be reckoned that for a certain measurement of TSEP, 1A measurement error of collector current results in 0.048℃(|a12/a11|) error of junction temperature for the tdoffmethod, while the error is 0.415℃(|a22/a21|) for the VeEPeakmethod. In conclusion, TSEP based on VeEPeakrequires more accurate current measurement than the TSEP based on tdoff. This feature partly results in the low accuracy of VeEPeak-based method. For the combined TSEP method, when the sampling circuit is mounted to obtain tdoffand VeEPeak, there is no need to monitor the collector current.

6 Conclusion

This paper has proposed an advanced junction temperature extraction approach for the high-power IGBT modules. The tdoffand VeEPeakbased TSEPs are combined to extract the junction temperature without the knowledgeof collector current. Experiments have been implemented to verify the effectiveness of the proposed method. Two individual TSEP methods, VeEPeak-based method and tdoff-based method, are compared with the combined TSEP method to clearly show the detailed performance. Combined TSEP method has some advantages over individual TSEP method asthe current measurement is not necessary. By getting rid of current measurement, the combined TSEPmethod is easier and cheaper to be applied in the practical operation condition.

[1] H. Wang, M. Liserre, F. Blaabjerg, et al. "Transitioning to physics-of-failure as a reliability driver in power electronics," IEEE Journal of Emerging and Selected Topics in Power Electronics, vol.2, no.1, pp. 97-114, 2014.

[2] U. M. Choi, F. Blaabjerg, K. B. Lee, "Study and handling methods of power IGBT module failures in power electronic converter systems,"IEEE Transactions on Power Electronics, vol.30, no.5, pp. 2517-2533, 2015.

[3] S. Yang, A. Bryant, P. Mawby, "An industry-based survey of reliability in power electronic converters," IEEE Transactions on Industry Applications, vol.47, no.3, pp. 1441-1451, 2011.

[4] E. Wolfgang, "Examples for failures in power electronics systems,"ECPE Tutorial on Reliability of Power Electronic Systems, Nuremberg, Germany, 2007: 19-20.

[5] M. A. Eleffendi, C. M. Johnson, "Application of Kalman filter to estimate junction temperature in IGBT power modules," IEEE Transactions on Power Electronics, vol.31, no.2, pp. 1576-1587, 2016.

[6] N. Baker, M. Liserre, L. Dupont, "Improved reliability of power modules: a review of online junction temperature measurement methods," IEEE Industrial Electronics Magazine, vol.8, no.3, pp. 17-27, 2014.

[7] Y. Avenas, L. Dupont, Z. Khatir, "Temperature measurement of power semiconductor devices by thermo-sensitive electrical parameters—a review,"IEEE Transactions on Power Electronics, vol.27, no.6, pp. 3081-3092, 2012.

[8] H. Kuhn, A. Mertens,"On-line junction temperature measurement of IGBTs based on temperature sensitive electrical parameters,"IEEE 13th European Conference on Power Electronics and Applications, 2009,pp. 1-10.

[9] J. A. B. Ccoa, B. Strauss, G. Mitic, "Investigation of temperature sensitive electrical parameters for power semiconductors(IGBT) in real-time applications," Proceedings of International Exhibition and Conference for Power Electronics, Intelligent Motion, Renewable Energy and Energy Management, VDE, 2014 ,pp.1-9.

[10] H. Luo, Y. Chen, P. Sun, "Junction temperature extraction approach with turn-off delay time for high-voltage highpower IGBT modules," IEEE Transactions on Power Electronics, vol.31, no.7, pp. 5122-5132, 2016.

[11] Y. Chen, H. Luo, W. Li, "A thermo-sensitive electrical parameter with maximum dIC/dt during turn-off for high power Trench/Field-Stop IGBT modules," IEEE Applied Power Electronics Conference and Exposition (APEC), 2016, pp. 499-504.

Xiang Wangreceived the B.S. degree from the School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, Wuhan, China, in 2014. He is currently pursuing his M.S. degree in the College of Electrical Engineering, Zhejiang University, Hangzhou, China.

His research interests include IGBT modeling.

Chongchong Zhureceived the B.Sc degree from the Electrical Engineering, University of Jinan, Jinan, China, in 2012, and is currently pursuing the M.Sc degree in College of Electrical Engineering, Zhejiang University, Hangzhou, China.

His research interests include high power IGBT module test and IGBT junction temperature extraction.

Haoze Luoreceived the B.S. and M.S. degrees from the Department of Electrical Engineering, Hefei University of Technology, Hefei, China, in 2008 and 2011, respectively. He received the Ph.D. degree from Zhejiang University, Hangzhou, China in 2015. From January to April 2015, he was a visiting researcher at Newcastle University, Newcastle upon Tyne, U.K. He is currently working as a Postdoc at the Department of Energy Technology in Aalborg University, Denmark.

His research interests include high- power converters and reliability of high- power modules.

Wuhua Li(M’09) received the B.Sc. and Ph.D. degree in Applied Power Electronics and Electrical Engineering from Zhejiang University, Hangzhou, China, in 2002 and 2008, respectively.

From 2004 to 2005, he was a Research Intern, and from 2007 to 2008, a Research Assistant in GE Global Research Center, Shanghai, China. From 2008 to 2010, he joined the College of Electrical Engineering, Zhejiang University as a Post doctor. In 2010, he was promoted as an Associate Professor. Since 2013, he has been a Full Professor at Zhejiang University. From 2010 to 2011, he was a Ryerson University Postdoctoral Fellow with the Department of Electrical and Computer Engineering, Ryerson University, Toronto, ON, Canada. His research interests include high power devices, advanced power converters and operation optimization for renewable energy based power systems. Dr. Li has published more than 100 peerreviewed technical papers and holds over 30 issued/pending patents.

Due to his excellent teaching and research contributions, Dr. Li received the 2011 TOP TEN Excellent Young Staff Award and the 2012 Distinguished Young Scholar from Zhejiang University, the 2012 Outstanding Young Researcher Award from Zhejiang Province, the 2012 Delta Young Scholar from Delta Environmental & Educational Foundation and the 2012 National Outstanding Young Scholar. He received four Scientific and Technological Achievements Awards from Zhejiang Provincial Government and the State Educational Ministry of China in 2009, 2011 and 2014, respectively.

Xiangning He(M’95--SM’96--F’10) received the B.Sc. and M.Sc. degrees from Nanjing University of Aeronautical and Astronautical, Nanjing, China, in 1982 and 1985, respectively, and the Ph.D. degree from Zhejiang University, Hangzhou, China, in 1989.

From 1985 to 1986, he was an Assistant Engineer at the 608 Institute of Aeronautical Industrial General Company, Zhuzhou, China. From 1989 to 1991, he was a Lecturer at Zhejiang University. In 1991, he obtained a Fellowship from the Royal Society of U.K., and conducted research in the Department of Computing and Electrical Engineering, Heriot-Watt University, Edinburgh, U.K., as a Post-Doctoral Research Fellow for two years. In 1994, he joined Zhejiang University as an Associate Professor. Since 1996, he has been a Full Professor in the College of Electrical Engineering, Zhejiang University. He was the Director of the Power Electronics Research Institute and the Head of the Department of Applied Electronics, and he is currently the Vice Dean of the College of Electrical Engineering, Zhejiang University. His research interests are power electronics and their industrial applications. He is the author or co-author of more than 280 papers and one book Theory and Applications of Multi-level Converters (Beijing, China: China Machine Press, 2006). He holds 22 patents.

Dr. He received the 1989 Excellent Ph.D. Graduate Award, the 1995 Elite Prize Excellence Award, the 1996 Outstanding Young Staff Member Award and 2006 Excellent Staff Award from Zhejiang University for his teaching and research contributions. He received seven Scientific and Technological Achievements Awards from Zhejiang Provincial Government and the State Educational Ministry of China in 1998, 2002, 2009 and 2011 respectively, and six Excellent Paper Awards. Dr. He is a Fellow of The Institute of Electrical and Electronics Engineers (IEEE) and has been appointed as IEEE Distinguished Lecturer by the IEEE Power Electronics Society in 2011. He is also a Fellow of the Institution of Engineering and Technology (formerly IEE), U.K.

* Corresponding Author, E-mail:woohualee@zju.edu.cn.

Supported by the National Nature Science Foundations of China (51490682, 51677166).