IGBT Module Rupture Categorization and Testing

Transcription

1 IEEE Industry Application Society Annual Meeting New Orleans, Louisiana, October 5 9, 1997 IGBT Module Rupture Categorization and Testing D. Braun D. Pixler P. LeMay Rockwell Automation - Allen Bradley 6400 W. Enterprise Drive Mequon, WI (414) Fax: (414) ; dhbraun@meq1.ra.rockwell.com Abstract - The isolated base, plastic molded, power semiconductor module assembly has been available for various devices for over twenty years. During this time the power bipolar junction transistor, the BJT, has given way to the Insulated Gate Bipolar Junction Transistor, the IGBT, as the self commutatible device of choice especially for voltage source inverter topologies. As part of the power structure design process, careful consideration should be given to possible power device failure. Obviously every power structure designer strives to remain well within the operating limits of the power device as specified by the IGBT manufacturer. However, there are component failures either of the power switching device or of supporting components that result in one or more IGBT power device failures. When the IGBT power devices fail, under certain circumstances, the failure can result in the rupture of the power module and extensive damage to the surrounding power compone nts. The topic under investigation in this paper is IGBT module rupture. The IGBT module rupture phenomenon will be characterized through testing and categorized into two separate ratings for each device investigated. A DC link inverter bus fuse is being implemented in most designs as a protective device to minimize extremely high overcurrent faults and power module rupture. The power semiconductor fuses available are optimized for lowest I 2 t let though energy, low forward drop and low insertion impe dance. As a foundation of the establishment of selected IGBT module rupture ratings the paper contains the testing of several selected fuses and their I 2 t values under specific conditions of 600 volts DC applied to an inverter grade capacitor bank. The selected fuses are tested together with the IGBT being evaluated for its rupture ratings. The power circuit is by design a low loss, low inductance structure which will result in the delivery of extremely high currents and energy transmission. The IGBT rupture testing is presented into two categories of overcurrent failure and of unclamped voltage failure. The two failure modes result in different values of rupture energy and I 2 t. In conclusion the rupture testing is compared with appropriately sized semiconductor protecting fuses. I. INTRODUCTION The design and development of IGBT module based power conversion concentrates on the proper interfacing of the total device. This work includes the very careful analysis of the device locus of operation for all conditions as the device is applied in the product. In order to assure that the product design is a robust and reliable entity, the drive normally is subjected to the most stringent product qualification testing that the product will ever experience throughout the product life. In addition, the product is also provided with numerous safety shutdown mechanisms in order to provide a coordinated method of limiting fault damage should a component fail. Thus for product developers to even consider catastrophic module rupture seems to be paradoxical to the main stream of product development. However there are unusual occurrences of IGBT module rupture that can lead to full product devastation. It is these IGBT module ruptures that this paper addresses and attempts to provide a guideline for others to consider and follow. Most power semiconductor IGBT module manufacturers will provide a calculated value of bond wire I 2 t for a given type of power module. These values however are not even close to the rupture I 2 t values which have been established through the testing and analysis presented in this paper. One of the goals of the paper was to fully evaluate and establish I 2 t ratings or minimum values for semiconductor fusing in order to protect the power structure should a power module fail in a mode that would lead to module rupture. In order to provide a proper fuse selection on the basis of a given value of melting I 2 t, it has been necessary to perform separate fuse testing in a similar topology to that of IGBT module. II. IGBT MODULE POWER CIRCUIT DESIGN AND PROTECTION CIRCUIT DISCUSSION As is well known, parasitic inductance[1] in an IGBT power inverter structure causes no end of problems in order to maintain reasonable power levels and yet stay within the power device safe operating area, SOA, especially when commutating an output short circuit current. A planar bus design methodology to minimize fixture stray inductance

2 (L stray ) is a critical design goal as device power levels increase. In addition to the traditional design goals of proper thermal design, gate resistance selection, and switching speed coordination through IGBT device characterization as in [2], component failure fault sensing and proper circuit shutdown also has to be addressed. Some power module device suppliers claim to have solved most of the added device protection needed through some degree of internal module intelligence. This higher degree of intelligence has to be thoroughly evaluated for the specific device design in the product to assure the power development engineer that indeed the fault protection is robust and repeatable. All this aside, the question of possible module rupture has to be addressed during the development cycle of the product. III. RUPTURE PHENOMENA The cause of IGBT rupture phenomena has been qualitatively theorized as follows: The fault current in the module either through a local failure at one of the IGBT or diode die has risen to the point of bond wire rupture. At this point the bond wire releases relatively large quantities of energy to the soft gel insulation medium. The energy transfer raises the soft gel temperature to the state change point and the gel quickly changes state first to the fluid state and next to the plasma state. At the same time, because of local bond wire failure, other bond wires are failing both on the same die and surrounding die due to excessive current. Next the soft gel in the plasma state produces relatively high internal pressure on the case of the module. If there is no external interruption of the fault current, such as with a fast acting, low I 2 t fuse, this energy transfer continues until either all of the bond wires have failed with no module rupture or until the pressure of the soft gel plasma reaches the point of module case failure. If the fault continues during this time, it is possible to transfer all of the power structure bus capacitance energy to the failure. Depending on the value of i 2 t rupture for the IGBT module, the value of bus capacitance, the value of the i 2 t melting of the protecting fuse, and the value of the DC bus voltage at the time of failure, the resulting occurrence can be widespread, to the point of causing product repair not economically justified. The theory is supported by recent work as report in [3] in which the authors removed the soft gel from one IGBT module. IV. FUSE I 2 t TESTING AT 600 VOLTS DC As a foundation for selection of the proper fuse for the prevention of IGBT module rupture, the fast acting fuses available from various sources have been tested in the same circuit at 600 volts DC with the same stray inductance in all cases except for the self inductance of each fuse. External Power Supply for Precharge V dc_bus [] 600 vdc External gate source I 2 C equiv 20,700 ufd SCR/ Diode L stray = 800 nh Fig. 1. Fuse Testing Circuit V fuse [Ch 2] I 1 [Ch 3] The test circuit for all of the fuse testing is given in figure 1. The test circuit was constructed with inverter grade electrolytic bus capacitors designed for high ripple current capability, and low parasitic resistance and inductance. The power connections were made using laminar bus design, in similar application of voltage source inverters presently in production. The stray inductance of 800 nh is actually very high for most high power inverter circuits but could not be avoided in order to insert the large current transducer to monitor the fault current. The value of 800 nh is the total circuit inductance including the typical fuse inductance. The current sensor selection will be further explained in the instrumentation section. The test circuit used exclusively for both the fuse testing and the IGBT module rupture testing uses the 20.7e-03 farad capacitor bank as the only source for fault current. The source inductance of the input supply whether AC or DC is at least 100 times greater than that of the stray inductance. The fault SCR used for fuse testing was selected as a 470 Amp RMS device in order to continue to use the device for as many multiple fuse tests as the device could withstand. As the fuse peak current and i 2 t increased each fuse test resulted in an SCR failure as well. The SCR gate source was selected to be as stiff as practical and provided a gate current of 1.25 amps typical with a di/dt of 2.5 amps per microsecond. During all of the fuse tests none of the SCR/diode modules used ever experience any evidence of rupture. The summary of fuse testing data is as listed in Table 1, located at the end of the paper. This data is a compilation of in most cases at least two tests of given fuse. The fuse suppliers were very helpful and provided the author s

3 support both with data that was available, sample devices and in some cases extra testing. The DC voltage testing to date is a good start to having a full catalog of fuse devices characterized in a common environment. As this data base continues to grow the correlation and application from an i 2 t basis will continue to provide value for power structure design of IGBT based voltage source inverters. FUSE TESTING Vbus Ch 3 Ifault Supplier X A70QS800 Math 1 I1 * I1 800A Fuse Ch #1 Vbuss Vdc 100V/div 59,600Amps Ch #2 Vfuse Ch2 Vfuse Vfuse 200V/div Ch #3 I1 I1 10,000A/div Math #1 I1 * I1 I1 *I1= 1KMA 2 /div A 2 s Fuse_3.PPT TDS 744A Scope # Unit T1-10A - Test #10 Vbus = 600VDC [SCR gated] Fig.2 DC Voltage Testing of Fuse # 10 In almost all cases the DC i 2 t for time periods less than 100 microseconds is not available from the fuse suppliers, thus this testing was a necessary first step in the process of determining the rupture i 2 t of a given IGBT module. Each test consisted of one shot of SCR gating and a coordinated single shot capture of the transient on 2-4 channel digital oscilloscopes. One event for unit # 10 referring to Table 1 is shown in figure 2. Figure 2 is one of at least 2 waveform pictures captured of each event. One waveform capture is taken for the melting i 2 t values and a second waveform capture is taken for the clearing i 2 t values. A. Overcurrent Testing It was important to observe the device performance during the fault, because even though the gate command from the driver circuit remained constant the device either due to a higher level of built in intelligence or due to the specific failure mode, actually attempted to commutate the fault current. If indeed the device attempted to commutate the fault current the specific test operating mode changed to that of an unclamped voltage fault. The summary of IGBT rupture testing is as listed in table 2. The table contains five basic IGBT module ratings of: (1) 300 amp; (2) 400 amp; (3) 600 amp; (4) 800 amp; and (5) 1200 amp ratings. These were all single switch IGBT modules. In addition most of the modules have actually more than one generation of IGBT module available and depending on the differences in the generation, can and do have different values of i 2 t rupture associated with them. An example of an overcurrent rupture without any attempt to commutate the fault current by the IGBT under test is shown as figures 4 and 5. Please note that each of the oscilloscope waveforms included in the paper also have an internal test # associated with them to better correlate the data. This example of a 600 ampere, 1200 volt IGBT module test resulted in a very slight rupture crack. Note that the fault current slope in figure 5 had very little slope change right from the onset of the fault. The current transducer used for the lower current tracking had a range of 6Kamperes, thus once above that range the waveform of figure 5 had no additional information. Also note that the cursor used for measuring the melting time i 2 t was placed at the point on the fuse voltage curve where the voltage slope changed dramatically marking the start of the fuse elements interrupting current. L stray = 800 nh The i 2 t values remained consistent in a given suppliers family type but varied dramatically from one type to the next. V. IGBT RUPTURE TESTING The IGBT rupture testing has been addressed into two classifications of failure, overcurrent and overvoltage. The overcurrent rupture [3, 4] is one in which the gate command signal is switched on and remains on throughout the failure of the module and interruption of the fault current. The overvoltage rupture is one in which the gate command signal is switched on only until the device under test reaches a precalculated value of current, which for these tests has been from 75% to 150% of module rating. External Power Supply for Precharge V dc_bus [-2] 600 vdc Ig [Ch 3-1] C equiv 20,700 ufd Vge [Ch2-1] IGBT Gate Driver Fig. 3 IGBT Rupture Testing Circuit V fuse [Ch 4-2] Ie [Ch4-1] Test IGBT Vce [Ch1-1]

4 Ch2 VgeQ1 Ch 4 Ie Q1 Ch3 IgateQ1 Math 1 Ie * Ie CM600HA-24H IGBT d/c: S37AB1 VceQ1 35,400Amps 29,183.6A 2 s TDS 744A Scope # Unit T2-55A - Vbus = 600VDC [+15 v / -15 v driver] Fig.4 IGBT Rupture Testing Unit #55 - Scope #1 Vdc_bus Ch3 Ie Q1 Ch4 Vfuse A70Q400A Fuse Ch #1 VceQ1 Vce 200V/div Ch #2 VgeQ1 Vge 5V/div Ch #3 Igate Q1 Ig 2A/div Ch #4 Iemitter Q1 Ie 7,500A/div Math #1 Ie * Ie Ie*Ie= 1e09A 2 /div Rgext = 2.1 T600_3.PPT A70Q400A Fuse Ch #1 Vdc_bus Vdc 100V/div Ch #3 Iemitter Q1 Ie 1000A/div Ch #4 Vfuse Vf 200V/div Rgext = 2.1 T600_3.PPT Ch2 VgeQ1 VceQ1 Ch 4 Ie Q1 Ch3 IgateQ1 Math 1 Ie * Ie A70QS300A Fuse Ch #1 VceQ1 Vce 200V/div Ch #2 VgeQ1 1,320Volts Vge 5V/div Ch #3 Igate Q1 31,800Amps Ig 2A/div Ch #4 Iemitter Q1 17,481.6A 2 Ie 7,500A/div s Math #1 Ie * Ie Ie*Ie= 1e09A 2 /div Rgext = 1.8 T400_4.PPT 1MBI400NN-120 IGBT d/c: 6N02 TDS 744A Scope # Unit T2-54A - Vbus = 600VDC [+15 v / -15 v driver] Fig.6 IGBT Rupture Overcurrent Testing Unit #54 - Scope #1 Ch3 Ie Q1 Vdc_bus Ch4 Vfuse A70QS300A Fuse Ch #1 Vdc_bus Vdc 100V/div Ch #3 Iemitter Q1 Ie 1000A/div Ch #4 Vfuse Vf 200V/div Rgext = 1.8 T400_4.PPT CM600HA-24H IGBT d/c: S37AB1 TDS 744A Scope # Unit T2-55B - Vbus = 600VDC [+15 v / -15 v driver] Fig.5 IGBT Rupture Testing Unit #55 - Scope #2 In many of the overcurrent IGBT rupture tests, the fault current did not remain continuous, but instead either reacted to a form of higher internal module intelligence as in the 300 and 400 ampere IGBT modules or to an internal failure mode for the specific device, the IGBT under test attempted to commutate and caused an unclamped overvoltage commutation to occur. The following figures 6 and 7 of a 400 ampere IGBT module rupture test in which no external evidence of rupture occurred, depicts an example of such an occurrence. Note that the device collector to emitter voltage reached 1320 volts in a very short time followed by a reduction of voltage to 600 volts for 40 microseconds after which the device completely failed and the familiar current versus time fuse clearing occurred. 1MBI400NN-120 IGBT d/c: 6N02 TDS 744A Scope # Unit T2-54B - Vbus = 600VDC [+15 v / -15 v driver] Fig.7 IGBT Rupture Testing Unit #54 - Scope #2 Also note that during the initial fault sequence the external gate supply did not decrease the gate voltage to normal device turn off threshold, but remained in the ON state. In this instance the internal gate intelligence attempted to commutate the device but was unsuccessful. As a result of the testing in the recent past, the classification of overcurrent and overvoltage rupture testing became less and less distinct. The authors have included a section on overvoltage rupture testing for completeness, however the results of the different testing modes have not proven to be clearly different in producing either distinctly different values of i 2 t rupture or different fault phenomenon. B. Overvoltage Testing The figures of 8 and 9 provide an example of an overvoltage IGBT rupture test. These tests were conducted in a similar test circuit of figure 3, however an external

5 OVERVOLTAGE FAILURE - IGBT MODULE TESTING Ch2 VgeQ1 Ch3 IgateQ1 Ch 4 Ie Q1 VceQ1 FZ1200R12KF1 IGBT s/n: ,464Volts TDS 744A Scope # Unit T2-32A - Vbus = 600VDC [+15 v / -15 v driver] A70Q600A Fuse Ch #1 VceQ1 Vce 200V/div Ch #2 VgeQ1 Vge 5V/div Ch #3 Igate Q1 Ig 2A/div Ch #4 Iemitter Q1 Ie 10,000A/div Math #1 Ie * Ie Ie*Ie= 1e09A 2 /div Rgext = 1.8 T1200_3.PPT Fig.8 IGBT Rupture Overvoltage Testing Unit #32 - Scope #1 OVERVOLTAGE FAILURE - IGBT MODULE TESTING Ch3 Ie Q1 1,080Amps A70Q600A Fuse Ch #1 Vdc_bus Vdc 100V/div Ch #2 Irev_fuse Ir_f 5A/div Ch #3 Iemitter Q1 Ie 250A/div Ch #4 Vfuse Vf 200V/div Rgext = 1.8 T1200_3.PPT FZ1200R12KF1 IGBT s/n: TDS 744A Scope # Unit T2-32B - Vbus = 600VDC [+15 v / -15 v driver] Fig.9 IGBT Rupture Overvoltage Testing Unit #32 - Scope #2 choke was selected to provide a reasonably constant inductance up the current value selected to switch off the external gate command and cause the unclamped voltage failure to occur. In this case a 1200 ampere module was tested and allowed to reach 1080 amperes just prior to the gate command switching point. The module began the commutation but failed when the collector to emitter voltage reached 1464 volts. C. No Fuse Rupture Testing One example of no fuse protection IGBT rupture testing is depicted in figure 10. This set of waveforms dramatically states the resulting current values for in this case a 300 ampere IGBT module. Other modules through 1200 amperes were also tested with the same test circuit and ruptured in a similar manner. Ch2 VgeQ1 Ch 4 Ie Q1 VceQ1 Math 1 Ie * Ie Ch3 IgateQ1 60,800Amps 260,090.A 2 s NO Fuse Ch #1 VceQ1 Vce 200V/div Ch #2 VgeQ1 Vge 5V/div Ch #3 Igate Q1 Ig 2A/div Ch #4 Iemitter Q1 Ie 10,000A/div Math #1 Ie * Ie Ie*Ie= 1e09A 2 /div Rgext = 2.7 T300_JNF.PPT 1MBI300JN-120 IGBT TDS 744A Scope # Unit T2-13C - Vbus = 600VDC [+15 v / -15 v driver] Fig.10 IGBT Rupture NO FUSE Overcurrent Testing Unit #13 VI INSTRUMENTATION Instrumentation range and quality proved essential for meaningful test data, especially for high current measurement. The experimental results of proper current and voltage measurements as well as proper selection of on board mathematical analysis for both fuse testing and for IGBT rupture testing are explained in the following sections. A. Current Measurement The current measurement of the fuse current in the fuse only testing as well as the emitter current of the IGBT power device during rupture testing, was a very demanding measurement for this paper. Current measurement during early testing of IGBT rupture caused the authors to evaluate several alternatives to main current measurement. The result of the evaluation was the procurement of a new current transducer, the Pearson Electronics Model # The device has a maximum current time product of 1200 amp-seconds and a peak current capability of 500Kamps. This device became the main measurement tool in combination with a Tektronix 744A digital oscilloscope that allowed on board mathematics of the square of current versus time. The main current was also monitored with a Rogowski coil device with a lower maximum range in order to observe the initial fault currents at levels up to 7Kamps. Current measurement of the lower levels of IGBT or SCR gate current for IGBT and fuse testing respectively was accomplished with Pearson Model 2878 current monitors. B. Voltage Measurement The voltage measurement of the IGBT power device during the rupture fault time is extremely difficult especially

6 at the IGBT gate to emitter levels. The common mode performance issues for low level signals, at the 15 volt levels, once the IGBT fault occurs are causes for very poor signal quality for any voltage transducer evaluated to date. C. Current Squared - Time Measurement With the solution of the main current measurement accomplished, the mathematic function of current squared can be accomplished directly on the digital oscilloscope. This current squared waveform of information can then be measured using an area measurement function which results in the current squared waveform integrated in time, to be displayed as a number on the oscilloscope screen. The oscilloscope measurement mode must be chosen to only perform measurements between the cursors on the screen. Thus multiple measurements of melting time i 2 t and clearing time i 2 t can be easily measured by moving only the ending time cursor. Since the oscilloscope chosen does not provide the option of converting current transducers signals from volts to amperes, the scales of the displayed waveforms including the mathematic waveform and the scales of the measurement values have to be converted to the proper units and recorded. In the measurement of i 2 t, it was suggested that the definition of the ending time for the melting time event be chosen at the inception of the fuse voltage dynamic slope change. This point is actually when the first of the fuse elements are interrupting their portion of the current in the fuse. Thus rather than choose the peak of the current waveform, the change in fuse voltage point was chosen which occurs slightly earlier in time than the peak current point. Referring to figure 11, the fuse selection process also involves the proper thermal design for maximum environmental conditions and maximum steady state load. This fuse selection flow chart is provided as a guide to the power structure development engineer to consider during the prototype development process. In order to ascertain that the fuses are properly thermally coordinated, the product must undergo extensive testing both for steady state conditions and for cyclic load conditions. Since the fuse devices are sensitive to absolute temperatures, especially for thermal fatigue and mechanical fatigue considerations, it is also necessary to perform the thermal testing in the worst case thermal environments that the product will be subjected to. In addition, caution must be applied to the interconnection terminations of the fuse or fuses. Normal current densities applied to normal copper bus sizing aren t necessarily the most complete answer. In general the fuse supplier is not supplying a fuse at a given labeled current rating to survive in the normal maximum industrial ambient temperature conditions. In many ways the fuse design coordination requires similar thermal design and attention as do the semiconductors of the power structure. VIII. CONCLUSION The IGBT module rupture testing has resulted in the establishment of a value of i 2 t limit for all of the ampere ratings of modules tested to date. All of the modules and their respective fuses, if applicable, have been cataloged and D. Energy Measurement For these experiments the capacitor bank of 20.7e-03 farads was consistently used for all testing. In addition, since all tests were conducted with only the capacitor bank as the source of energy, the change in voltage of the capacitor bank during the fuse test or the IGBT rupture event could be measured. Once the measurement was made the energy loss could be easily calculated according to the following equation: W e = 1/2 C equiv ( V 12 - V 22 ) (1) VII. FUSE SELECTION PROCESS Unfortunately the value of i 2 t and coordination of that value with a given IGBT power structure applied at a given maximum voltage is only one of the qualifying answers toward a robust solution for the proper selection of the device.

7 IGBT Fuse Selection Process whether the cause was one of overvoltage or overcurrent that began the failure sequence. Perform DC Fuse Testing of Selected Types and Rating for I 2 t Performance Only No Does Fuse Manufacturer provide DC I 2 t Characterization? Yes Validate DC Characterization in Application Initial testing of IGBT module rupture indicated that a distinct difference in i 2 t level between overcurrent and overvoltage rupture would be possible to identify. The results of the different testing modes have not proven to be clearly different in producing either distinctly different values of i 2 t rupture or different fault phenomenon. The test instrumentation employed brought clearer understanding of the test program, however work is still required to better instrument low level voltages in the 15 volt range that have high dynamic common mode voltages occurring simultaneously during the critical measurement window. Select fuse type and rating that is less than IGBT module rupture I 2 t rating The proper selection of fuse size based upon melting time i 2 t must be verified by testing in a similar circuit to that of the ultimate application. In addition the fuse that is selected has to be verified both thermally and from a fatigue failure standpoint, in order to assure that the fuse does not cause a nuisance failure to occur. Perform Overcurrent (and if possible), Overvoltage IGBT and Fuse Testing Additional work should be done to better understand the thermal and fatigue issues with the semiconductor type fuses, [5, 6, 7] and implement and / or develop industrial models for fuses compatible with current simulation tools. References: Determine Source of Error No No Does Selected Fuse Protect the IGBT Against Rupture? Yes Perform thermal testing and fuse fatigue testing Does selected Fuse Maintain Thermal Integrity? Yes Design Complete Fig. 11. Fuse Selection Flow Chart will be further examined in the future to ascertain whether one can determine, by physical examination, after a failure, [1] Skibinski, G., Divan, D., "Design methodology & modeling for low inductance planar bus structures", European Power Electronics ( EPE) Conf., 1993 [2] Braun, D., Lukaszewski, R., Pixler, D., Skibinski G., "Use of Co-axial CT and Planar Bus to Improve IGBT Device Characterization", IEEE - IAS Conf, Oct. 1996, pages [3] Duong, S., Schaeffer, C., "Investigation on Fuses Against IGBT Case Explosion", Proceedings of PCIM-Europe June 97, pp [4] Duong, S., Schaeffer, C., Rouve, L-L., De Palma, J.F., Mullert, C., "Fusses for Power IGBT - Converters", IEEE - IAS Conf, Oct. 1994, pages [5] Duong, S., Marechal, Y., Schaeffer, C., Mullert, C., Sarrus, F., Gelet, J.L., "Electrothermal Model of a Fuse", IEEE - IAS Conf, Oct. 1996, pages [6] Wilkins, R., Cline, H.C., "An Advanced Applications Tool to Enhance the Fuse Protection of Power Semiconductors, IEEE - IAS Conf, Oct. 1993, pages [7] Wilkins, R., Wilkinson, A., Cline, C., "Endurance of Semiconductor Fuses Under Cyclic Loading", Proceedings of Fourth International Conference on Electric Fuses and their Applications, Sept. 1991, pages

8 Table Volt DC Fuse Testing Summary Unit FUSE Supplier Mfr's Cat# I peak Melting I 2 t Arcing I 2 t Clearing I 2 t Energy # Current X, Y, Z Amps amp 2 sec amp 2 sec amp 2 sec Joules Rating X A70Q125 14,000 1,319 1,483 2, X A70Q150 15,600 1,845 1,998 3, X A70Q200 18,000 2,826 3,584 6, X A70Q250 23,200 6,436 6,280 12, X A70Q600 40,400 27,454 42,727 70,182 1, X A70QS200 24,200 5,968 8,020 13, X A70QS250 30,400 13,537 14,187 27, X A70QS300 32,900 17,080 16,107 33, X A70QS350 36,000 20,148 24,643 44,791 1, X A70QS400 39,000 27,879 29,348 57,227 1, X A70QS800 59, ,539 83, ,508 2, X A00-66C200D1 20,200 3,757 6,269 10, X A00-66C250D1 27,800 10,154 16,462 26, X A00-66C315D1 31,200 14,767 21,967 36, X A00-66C350D1 34,500 19,572 23,805 43,377 1, X A65C4504AB 41,100 33,130 46,403 79,533 1, X 2 * A65C450-4AB 60, , , ,186 2,313 20F 315 Y L ,400 8,539 13,310 21, F 350 Y R ,000 11,263 16,759 28, F 400 Y E ,500 19,198 28,798 47,996 1,034 23F 450 Y A ,000 27,453 38,283 65,735 1,139 Table 2 - IGBT Module Rupture Testing Summary Unit Module Module I peak Melting I2t Arcing I 2 t Clearing I 2 t Energy Fault Test # Rating Mfr's Cat# Amps amp2sec amp 2 sec amp 2 sec Joules Classification Comments MBI300NN ,700 11,804 14,727 26, Overcurrent no external damage MBI400NN ,000 16,977 17,023 34,000 1,034 Overcurrent no external damage CM600HA-24H 34,500 21,768 21,807 43,575 1,086 Overcurrent VERY Slight Rupture FZ800R12KF1 32,700 16,864 15,568 32, Overcurrent no external damage 59 1,200 FZ1200R12KF1 36,600 30,320 21,857 52,177 1,440 Overcurrent VERY Slight Rupture MBI300JN ,000 13,673 39,351 53, Overcurrent no external damage MBI300JN ,400 59,940 41, ,836 1,583 Overcurrent Ruptured MBI300JN , , , ,336 3,726 Overcurrent No Fuse Ruptured!! 31 1,200 FZ1200R12KF1 13,200 5,875 6,808 12, Overvoltage no external damage 32 1,200 FZ1200R12KF1 23,200 37,254 29,266 66,520 1,535 Overvoltage Ruptured