7.3 Evaluating the ED60 results
7.3.2 On-state voltage measurements
The endurance test during which the measurement device was designed to be imple-mented in the inverter, follows the same procedure as the ED60-tests performed in section
6.4. In the following section, the junction temperatures of the DUT during aforemen-tioned tests are calculated utilizing the measured heat sink temperatures, approximation of power losses during power cycling and the measured thermal resistance of the power module from section 7.1.1. The calculated estimations of junction temperatures are then compared to the corrected calibration curve with the card formed in section 7.2.
Junction temperatures were calculated using the following equation:
Tj,n =Ts,DUT,n+PIGBT·Rth(j−s)+Tdiodes−igbt, (41)
wherenis the number of the cycle in ED60-test,PIGBTis the power loss of single IGBT, Rth(j−s)is thermal resistance from junction to heat sink andTdiodes−igbtis the temperature rise in IGBT caused by adjacent diodes. The resulting junction temperatures and their corresponding on-state voltages for the first and the second ED60-test are presented in table 11.
Table 11.Calculated junction temperatures and their corresponding on-state voltages from ED60-test 1 and ED60-ED60-test 2.
In Fig. 45 the results presented in table 11 are plotted against the corrected calibration curve with card. The linear trendlines of the both ED60-results are also presented.
Figure 45.The calculated junction temperatures based on results from ED60-test 1 (blue dots) and ED60-test 2 (red dots) with corresponding linear trend lines plotted against corrected calibration curve with card (green line).
It can be seen from Fig. 45 that the calculated junction temperatures are slightly lower compared to the calibration curve at corresponding VCE values. There are also minor differences in the slopes of calibration curve and trend lines of the junction temperatures.
Junction temperature point at (x = 106.9 °C, y = 205.0 mV) is significantly farther away from the trend line compared to other results. This temperature point is due to the same anomaly that was discussed in section 6.4 and can be regarded as a faulty measurement.
Excluding the temperature point at (x = 106.9 °C, y = 205.0 mV), the largest errors be-tween calculated junction temperature and the calibration curve are originated from cycle 1 and 2 of ED60 test 1 and cycle 1 and 2 of ED60 test 2. The errors are -4.1 °C, -4.0 °C, -3.8 °C and -3.9 °C, respectively. For both test results, the error between calculated junc-tion temperature and calibrajunc-tion curve seems to diminish as the temperature rises.
The error between calculated junction temperatures and the calibration curve could be due to several reasons, some of which are discussed below.
The calculations are partly based on approximations, which are prone to errors due to inaccuracies in the parameters used. The average error between the calibration curve and all of junction temperatures calculated from the ED60-test results is -3.3 °C. For example,
a rise of 0.0216 WK in the thermal resistance from junction to heat sink would decrease the average error to approximately zero.
Standard for thermocouple tolerances IEC-60584-2 specifies the tolerance class I for type K thermocouples as±2.5°C, or±0.75%, which ever is greater. All of the measurements performed with type K thermocouple were performed at the temperature of 125 °C, or below, which makes ±2.5 °C the tolerance to be used. All of the heat sink temperature measurements were performed utilizing type K thermocouples. Although it is unlikely that the errors in temperature measurements would have been as high as the standard tol-erance would suggest, some portion of the error between the calibration curve and calcu-lated junction temperatures could be explained due to the tolerance of the thermocouples.
To approximate the error outside of the range of the ED60-results, the trend lines of cal-culated junction temperatures were compared to the corrected calibration curve with card in temperature range from 90 °C to 100 °C and in temperature range from 120 °C to 130 °C. Trend lines and corrected calibration curve with card in aforementioned temper-ature ranges are plotted in Fig. 46 and Fig. 47, respectively.
Figure 46.Trend line of calculated junction temperatures based on ED60-test 1 (blue dotted line), trend line of calculated junction temperatures based on ED60-test 2 (red dotted line) and corrected calibration curve with card (green line) in temperature range from 90 °C to 100 °C.
As seen from Fig. 46, at the temperature of 95 °C, the difference between trend line of ED60-test 1 results and calibration curve is approximately 17 mV, and the difference between trend line of ED60-test 2 results and calibration curve is approximately 16 mV.
Using the sensitivity of corrected calibration curve with card, -2.58 mVK , these differences in on-state voltages result in 6.59 °C and 6.20 °C difference in temperature, respectively.
Figure 47.Trend line of calculated junction temperatures based on ED60-test 1 (blue dotted line), trend line of calculated junction temperatures based on ED60-test 2 (red dotted line) and corrected calibration curve with card (green line) in temperature range from 120 °C to 130 °C.
As seen from Fig. 47, at the temperature of 125 °C, the difference between trend line of ED60-test 1 results and calibration curve is approximately 5 mV, and the difference between trend line of ED60-test 2 results and calibration curve is approximately 1 mV.
Using the sensitivity of corrected calibration curve with card, -2.58 mVK , these differences in on-state voltages result in 1.94 °C and 0.39 °C difference in temperature, respectively.
As the approximated errors in temperatures of 95 °C and 125 °C are observed, it is obvious that the error diminishes as the temperature rises. Due to this observation, it could be suggested, that the measurement device is not of satisfactory accuracy when temperatures below 100 °C are measured. However, these approximations are based on trend lines of two sets of measurements with results in relatively narrow range of temperatures. Tests with lower and higher junction temperatures than in ED60-test performed in this work would be necessary to obtain more accurate approximations.
Despite of the errors between the calibration curve and the calculated junction tempera-tures based on ED60-tests, the measurement device satisfied all of the requirements which were set for it. The maximum error between the calibration curve and calculated junction temperature was -4.1 °C, which is within the required accuracy of 5 °C.
8 CONCLUSION
In this Master’s Thesis, a device for IGBT power module junction temperature measure-ment was designed, manufactured and tested. The measuremeasure-ment device was to be im-plemented in an inverter type welding power converter and the measurement was to be performed during an endurance test consisting of active switched-mode power cycling.
The basic concepts and characteristics of most common power semiconductors were dis-cussed. The most common methods for junction temperature measurement were reviewed and on-state voltage with low current as a temperature-sensitive electrical parameter was selected as a method to utilize in this work.
The welding power converter, the endurance test procedure and the subsequenting require-ments for the measurement device were discussed. The design process for the measure-ment device was described, including the basic functioning of the measuremeasure-ment device.
A PCB prototype for the device was manufactured using etching technology.
The prototype of the device underwent several tests to ensure proper functioning of the device and safe implementation to the power converter. Several measurements were per-formed with the prototype to evaluate the accuracy of the device.
By combining measurement results and calculations, it was demonstrated that the junction temperature can be measured utilizing the on-state voltage with low current, and that the prototype resulted from this work can be utilized to measure junction temperature measurements of IGBT power module within reasonable accuracy.
REFERENCES
[1] Song Y. & Wang B. Survey on reliability of power electronic systems. IEEE Trans-actions on Power Electronics, 28(1):591–604, 2013.
[2] Yang S. et al. An industry-based survey of reliability in power electronic converters.
IEEE Transactions on Industry Applications, 47(3):1441–1451, 2011.
[3] H. Niu. A review of power cycle driven fatigue, aging, and failure modes for semi-conductor power modules. InIEEE Conference Publications, 2017.
[4] Gopireddy L. R. et al. Power cycle testing of power switches: A literature survey.
IEEE Transactions on Power Electronics, 30(5):2465–2473, 2015.
[5] Infineon. https://www.infineon.com/cms/en/product/promopages/power2power/.
[6] B. J. Baliga. Fundamentals of Power Semiconductor Devices. Springer, second edition, 2019.
[7] Volke A. & Hornkamp M. IGBT Modules - Technologies, Driver and Application.
Infineon Technologies AG, second edition, 2012.
[8] J. Lutz, H. Schlangenotto, U. Scheuermann, and R. De Doncker. Semiconductor Power Devices - Physics, Charasteristics, Reliability. Springer, second edition, 2018.
[9] IEC 60747-9. Semiconductor devices - Discrete devices - Part 9: Insulated-gate bipolar transistors (IGBTs). IEC, 2.0 edition, 2007.
[10] Chen J. et al. Investigations on averaging mechanisms of virtual junction tempera-ture determined by vce (t) method for igbts.IEEE Transactions on Electron Devices, 67(3):1106–1112, 2020.
[11] Schmidt R. & Scheuermann U. Using the chip as a temperature sensor - the in-fluence of steep lateral temperature gradients on the vce(t)-measurement. InIEEE Conference Publications, 2009.
[12] L. Dupont, Y. Avenas, and P. Jeannin. Comparison of junction temperature evalua-tions in a power igbt module using an ir camera and three thermosensitive electrical parameters. IEEE Transactions on Industry Applications, 49(4):1599–1608, 2013.
[13] Dupont L. & Avenas Y. Evaluation of igbt thermo-sensitive electrical parameters under different dissipation conditions – comparison with infrared measurements.
Microelectronics Reliability, 52(11):2617–2626, 2012.
[14] Yong L. Power Electronic Packaging - Design, Assembly Process, Reliability and Modeling. Springer, first edition, 2012.
[15] M. Ciappa. Selected failure mechanisms of modern power modules. Microelectron-ics Reliability, 42(4):653–667, 2002.
[16] Y. Avenas, Dupont L., and Khatir Z. Temperature measurement of semiconductor devices by thermo-sensitive electrical parameters-a review. IEEE Transactions on Power Electronics, 27(6):3081–3092, 2012.
[17] Sathik M. H. M. et al. Comparison of igbt junction temperature measurement and estimation methods-a review. In 2017 Asian Conference on Energy, Power and Transportation Electrification (ACEPT), 2017.
[18] Qiu X. Z. et al. Review of igbt junction temperature extraction and estimation meth-ods. IOP Conference Series: Materials Science and Engineering, 774(1):012091, 2020.
[19] Baker N. et al. Igbt junction temperature measurement via peak gate current. IEEE Transactions on Power Electronics, 31(5):3784–3793, 2016.
[20] Khatir Z. et al. Investigations on junction temperature estimation based on junction voltage measurements. Microelectronics Reliability, 50(9):1506–1510, 2010.
[21] Dupont L. & Avenas Y. Evaluation of thermo-sensitive electrical parameters based on the forward voltage for on-line chip temperature measurements of igbt devices.
In2014 IEEE Energy Conversion Congress and Exposition (ECCE), 2014.
[22] Eleffendi M. A. & Johnson C. M. Evaluation of on-state voltage vce(on) and thresh-old voltage vth for real-time health monitoring of igbt power modules. In 2015 17th European Conference on Power Electronics and Applications (EPE’15 ECCE-Europe)), 2015.
[23] Bakun I. et al. Real-time measurement of igbt’s operating temperature. AU-TOMATIKA, 52(4):295–305, 2011.
[24] Ammous A. & Morel H. Transient temperature measurements and modeling of igbt’s under short circuit. IEEE Transactions on Power Electronics, 13(1):12–25, 1998.
[25] Sun P. et al. Online junction temperature extraction with turn-off delay time for high power igbts. In2014 IEEE Energy Conversion Congress and Exposition (ECCE), 2014.
[26] Kuhn H. et al. On-line junction temperature measurement of igbts based on temper-ature sensitive electrical parameters. InIEEE Conference Publications, 2009.
[27] Li L. et al. A turn-off delay time measurement and junction temperature estimation method for igbt. In2017 IEEE Applied Power Electronics Conference and Exposi-tion (APEC), 2017.
[28] Bryant A. et al. Investigation into igbt dv/dt during turn-off and its temperature dependence. IEEE Transactions on Power Electronics, 26(10):3019–3031, 2011.
[29] Infineon. F4-150r12ks4 datasheet.
[30] Vishay. Cny64/cny65/cny66 optocoupler datasheet.
[31] Nexperia. Hef4093b datasheet.
[32] Coto Technology. 9104 series/high voltage sip reed relays datasheet.