Method for measuring deterioration of thermal resistance between a power semiconductor and a heat sink, and control device for a power semiconductor
The method addresses offline limitations by using AC voltage on power semiconductor control terminals to monitor thermal resistance, providing accurate online monitoring with reduced complexity and parasitic effects.
Patent Information
- Application Number
- JP2023552404
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2021-03-01
- Filing Date
- 2021-07-15
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2041-07-15
AI Technical Summary
Existing methods for measuring thermal resistance between a power semiconductor and its heat sink are offline, require additional devices, and are not compatible with online monitoring or power semiconductor modules with parallel dies, lacking information on internal structural details and requiring complex calibration.
A method that measures thermal resistance by applying an AC voltage to the gate/source junction of the power semiconductor, monitoring internal gate resistance and junction temperature changes over time, without additional sensors or complex calibration, using available control terminals to generate heat and calculate thermal resistance deterioration.
Enables online thermal resistance monitoring with high accuracy, reducing complexity and parasitic effects, allowing for continuous degradation assessment without affecting switching behavior.
Smart Images

Figure 0007592185000023 
Figure 0007592185000024 
Figure 0007592185000025
Abstract
Description
[Technical field]
[0001] The present invention relates to monitoring the degradation of the thermal resistance between a power semiconductor, such as a switching semiconductor of a power converter, and its heat sink through measurement of the thermal resistance between such a power semiconductor and its heat sink, which degradation is an important parameter to be measured during operation of a power device, such as a power converter, in order to increase the robustness of the power device against overheating and to apply condition monitoring functions to the power modules of the power device. [Background technology]
[0002] Currently available methods for determining thermal resistance changes are primarily based on off-line measurements.
[0003] In such measurements, the thermal resistance between a power semiconductor module and its heat sink terminals can be measured using common thermal resistance measurement standards such as ASTM E1225-13, where a high temperature source is placed on a first terminal of the thermal resistor and a heat flux probe is placed on the second terminal of the thermal resistor. This approach is sufficient to measure the thermal resistance of a heterogeneous structure, but does not provide information about the internal structure details, such as the thermal resistance contributions of different components of the module. Furthermore, the need for a heat source and heat flux measurements requires additional devices on the power semiconductor module that are not suitable for measuring the thermal resistance under operation of the module.
[0004] The second approach consists in operating the power semiconductor under known drain / source or emitter / collector power dissipation losses and measuring the temperature difference between the power semiconductor and its base plate. Heat is generated using the main electrical path of the power semiconductor (drain-source or collector-emitter) and the dissipation losses on the power semiconductor are related to the power provided by the power supply. The power needs to be measured using accurate power measurement equipment and the operating conditions must be stable. This places constraints on when the measurements can be performed, on the operating conditions and on the circuits used to monitor the junction temperature, and requires accurate power and temperature measurement systems. Furthermore, this method is not transferable to online monitoring when the power semiconductor module is mounted on a power converter, or to power semiconductor modules with parallel dies. The dissipated power can be estimated using the electrical characteristics of the power semiconductor, but this requires accurate calibration of the temperature-dependent electrical characteristics of each power semiconductor in the production line due to significant chip-to-chip variations.
[0005] Another approach to measure the thermal resistance between a semiconductor device and its heat sink terminals consists in measuring the transient response of the junction temperature of the power semiconductor. The temperature of the semiconductor is raised to a certain value by means of the power dissipated in the semiconductor, and the transient temperature signal is recorded at a fast sampling rate after turning off the device. This method allows identifying details of the inhomogeneous structure of a power semiconductor module, but this measurement requires specific control procedures and measurement time scales that are not compatible with online thermal resistance estimation during module operation. Furthermore, the heat required to raise the semiconductor temperature requires high currents that are not always available, depending on the application.
[0006] According to the prior art, in order to measure the thermal resistance degradation online or when the power semiconductor module is already installed on the power converter, an additional accurate sensor is required to perform accurate power and junction temperature measurements, and / or a pre-calibrated thermal sensitive electrical parameter (TSEP) is required. In addition, the TSEP needs to be calibrated for each single device, considering the spread of electrical parameters such as Miller plateau or internal gate resistance between power semiconductor modules due to manufacturing variations.
[0007] Regarding measurements during operation, Patent Document 1 describes a method for estimating the temperature of a semiconductor chip housed in an operating power semiconductor device, such as an IGBT power module, which includes the steps of determining a voltage drop across the power semiconductor device for a value of an applied load current while the power semiconductor device is in operation, and estimating the temperature of the semiconductor chip by evaluating the relationship between the determined voltage drop and the value of the applied load current based on a semiconductor chip temperature model.
[0008] With regard to measuring ageing, US Pat. No. 5,399,433 shows a method in which an excitation signal is used to generate an excitation current which is passed through a power semiconductor, introducing a power loss in the power semiconductor and reading a temperature signal to determine the ageing of the power semiconductor.
[0009] In the above document, the excitation current flows between the collector-emitter junction or the source-drain junction of the power semiconductor. [Prior art documents] [Patent documents]
[0010] [Patent Document 1] U.S. Patent No. 9,689,754 [Patent Document 2] International Publication No. 2017 / 102262 Summary of the Invention [Problem to be solved by the invention]
[0011] In view of the prior art, it is an object of the present invention to provide an online measurement method for the degradation of the thermal resistance between a power semiconductor and its heat sink, which makes it possible to measure the evolution of the thermal resistance between said power semiconductor device and a heat sink during operation of the power semiconductor without prior calibration of the electrical characteristics of the power semiconductor and without the need for additional temperature sensors. [Means for solving the problem]
[0012] More specifically, the present invention relates to a power semiconductor in a power module assembly, the power semiconductor having an internal gate resistance, and a thermal resistance R between the power semiconductor and a heat sink of the power module assembly. Th A method for measuring degradation of Measuring a first initial parameter K0 related to a junction temperature of the power semiconductor at a stable operating temperature; An AC voltage V of an AC source with frequency f1 is applied to the gate / source junction of the power semiconductor. AC heating the internal gate resistor by applying a voltage Vcc to the internal gate resistor and measuring a second initial parameter K1 related to a junction temperature after the heating; storing the first initial parameter K0 and the second initial parameter K1 in a memory; Including, After at least one distinct time interval d ∈ {d1, ..., dn} during the life of the power module assembly, Stable operating temperature T0 d ∈{T0 d1 ,...T0 dn The first subsequent parameter K0 related to the junction temperature of the power semiconductor at d ∈{K0 d1 ,...,K0 dn}, and The internal gate resistor is heated by applying an AC voltage VAC having a frequency f1 to the gate / source junction of the power semiconductor, and the junction temperature after heating is T1d ∈{T1 d1 ,...T1 dn} The second subsequent parameter K1 d ∈{K1 d1 ,...,K1 dn}, and Including, The above method is The subsequent parameters K0d ∈ {K0d1, ..., K0dn} and K1 related to the junction temperature d ∈{K1 d1 ,...,K1 dn} and the initial parameters K0 and K1 read from the memory, the thermal resistance degradation ΔR Th Calculating
number
number
[0013] The proposed method provides an easy way to check the health of the power semiconductors with no or limited use of additional sensors and circuits besides those available in the power module. Moreover, the power dissipated on the power semiconductors by the gate resistors is the same throughout the life of the power semiconductors. Furthermore, since no power terminals are used to generate heat in the semiconductors, the complexity of the heating system is reduced considering that only the control pads of the semiconductor devices are used to generate heat on the semiconductor devices and any additional parasitic elements, thus no wired parasitic inductors are introduced that may degrade the switching behavior of the power semiconductor devices.
[0014] In one embodiment, the time interval d ∈ {d1,...,dn} between measurements of the thermal resistance is selected such that the first parameter K0d remains within a specified range K0+ / -k for K0.
[0015] The method comprises: determining an initial first parameter K0 in a specified initial part of the life of a power semiconductor for several stable junction temperatures T1...Tn of the power semiconductor during operation; (T) and the second parameter K1 (T) The table may include storing the initial value K0 d is the closest K0 to the measurement at the end of the interval d. (T) is selected within the prescribed range.
[0016] This provides a reference value for several temperatures of the power semiconductor during operation. Furthermore, the degradation of Rth can be measured at any operating temperature of the power semiconductor device, enabling sufficient in-operation thermal resistance degradation monitoring.
[0017] The specified initial portion of the life of a power semiconductor may be taken before the first 1000 to 1500 hours of operating time of the semiconductor.
[0018] This gives a large reference life.
[0019] The table may include a heating duration h and a first initial parameter K0 (T,h) and the second initial parameter K1 (T,h) can be measured for several of the heating durations h for layers with different time constants between the power semiconductor and the heat sink.
[0020] This provides a reference measurement that is useful for detecting the depth of degradation between the die and the heat sink.
[0021] The AC voltage may have a DC component to polarize the gate with respect to the source of the power semiconductor, and has a peak voltage below the flatband voltage of the power semiconductor.
[0022] This may allow limiting VGE or VGS to the drain / source or collector / emitter non-conducting states, making it independent of the load operating conditions of the power semiconductor device and avoiding pre-calibrating the method for the operating point of each operating power semiconductor device.
[0023] The frequency f1 may be lower than the inverse of the approximate gate input power semiconductor electrical time constant of the power semiconductor.
[0024] In such a case, F1 is approximated as f1=(2·π·Cin·R) -1 The frequency f1 may be selected as the input resonant frequency of the power semiconductor. In such a case, the approximate value of f1 is f1=(2·π·Lin·Cin) -2 may be also possible.
[0025] This makes it possible to calculate the thermal resistance degradation with low relative error, independent of the parasitic elements Cin and Lin.
[0026] To obtain a sufficient signal-to-noise ratio, the frequency f1 should be 0,1·(2·π·Cin·R) -1 It may be higher.
[0027] The parameter KXy related to the junction temperature of a power semiconductor may be the internal gate resistance RXy of said power semiconductor, where X is equal to 0 or 1 and y is null or d.
[0028] This allows the thermal resistance degradation to be determined using only the control terminals of the power semiconductor. Furthermore, since the thermal resistance degradation is calculated using differential values and the parameter KXy varies depending on the junction temperature, no pre-calibration of the TSEP-based method is necessary.
[0029] The method may include injecting a square DC current through an internal gate resistance of the power semiconductor without applying the AC voltage to obtain information of the internal gate resistance.
[0030] This provides a means of measuring the internal gate resistor without heating said resistor, increasing the accuracy of the measurement. Furthermore, the junction temperature measurement can be decoupled from the heating means and used independently for other features such as over-temperature protection.
[0031] In this case, a first AC voltage Vac1 with a first peak-to-peak level can be applied to obtain an R0 or R0d value, and a second AC voltage V with a second peak-to-peak value greater than the first peak-to-peak value can be applied to obtain an R0 or R0d value. AC is applied to heat the gate resistance of the power semiconductor, and the R1 value or the R1d value is obtained.
[0032] The method comprises providing a resistor with value Rmeas in a circuit including the gate / emitter or gate / source of the power semiconductor and an AC source, and applying the AC voltage V AC as a pulse of 5V to 40V to heat the internal gate resistor, and measure a first voltage Vmeas across the resistor as a first parameter Kv0 at a first peak level of the AC source to determine the R0 value as R0=Rmeas (V AC / K v0-1) and measure the second voltage Vmeas at the end of the resistor as the second parameter Kv1 at the last peak level of the AC source to determine the value of R1 as R1=Rmeas·(V AC / K v1 -1) and repeating the above measurements and calculations to obtain R0d and R1d values.
[0033] Said parameter KXy related to the junction temperature of the power semiconductor may be the junction temperature TXy of said power semiconductor, where X is equal to 0 or 1 and y is null or d.
[0034] Voltage V AC may be derived from a control signal of a power component modulated by the heating signal.
[0035] This provides a means to calculate the thermal resistance degradation without modifying the classical structure of the gate driver.
[0036] The present invention also relates to a control device for a power semiconductor, the control device comprising: a control circuit for controlling an AC voltage V with a frequency f1 which is higher than the frequency allowing rectification of said power semiconductor at the gate / source junction or gate / emitter junction of said power semiconductor; AC and means for measuring the temperature or a thermal sensitivity parameter of a power semiconductor at a selected point in the life of said semiconductor, and the resulting change in thermal resistance ΔR Th and computer means for calculating over the lifetime of said semiconductor according to the method of the invention.
[0037] In an advantageous mode of realization, the thermal sensitivity parameter is an internal gate resistance of the power semiconductor, and the control device is AC A voltage sensing means is included in the source branch of the injection circuit.
[0038] A detailed description of exemplary embodiments of the present invention is discussed below with reference to the accompanying drawings. [Brief description of the drawings]
[0039] [Figure 1] FIG. 1 is a diagram showing an example of a power module assembly with deteriorated thermal resistance. [Figure 2A] FIG. 2 is a schematic diagram of a heat measuring means. [Figure 2B] FIG. 2 is a schematic diagram of a heat measuring means. [Diagram 3] FIG. 1 is a simplified schematic diagram of a TSEP measurement circuit. [Figure 4] FIG. 1 is a simplified schematic diagram of a combined gate driver and thermal resistance measurement circuit. [Figure 5A] FIG. 5 is a diagram showing signal waveforms in the embodiment of FIG. 4. [Figure 5B] FIG. 5 is a diagram showing signal waveforms in the embodiment of FIG. 4. [Figure 5C] FIG. 13 illustrates waveforms of an alternative embodiment. [Figure 6A] 1A-1C show examples of waveforms used in the described process. [Figure 6B] 1A-1C show examples of waveforms used in the described process. [Figure 7A] FIG. 1 is a simplified schematic diagram of a TSEP measurement embodiment. [Figure 7B] FIG. 1 is a simplified schematic diagram of a TSEP measurement embodiment. [Figure 7C] FIG. 1 is a simplified schematic diagram of a TSEP measurement embodiment. [Figure 8] 1 is a flow chart of an embodiment of the disclosed method. [Figure 9] 1 is a flow chart of an embodiment of an initial measurement. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0040] The present invention relates to a method and a device for monitoring the thermal resistance between a power semiconductor and its heat sink in a power semiconductor assembly or power module assembly and for detecting degradation of said thermal resistance. Figure 1 shows a schematic diagram of a power semiconductor 1, such as a MOS transistor, an IGBT, a MOSFET, etc. The power semiconductor assembly is composed of a stack of the power semiconductor 1 and at least two of the following elements: a die attach layer 3a, a substrate composed of a metal structure 4, an electrical isolation structure 5 and another metal structure 6, a substrate attachment layer 3b, a base plate 7, a thermal interface 3c and a heat sink 8.
[0041] Such degradation of the thermal resistance of the stack may be primarily caused by cracks 3k in the die attach layer 3a, the attachment layer 3b, or the thermal interface 3c, which restrict the heat flow between the power semiconductors and the heat sink 8.
[0042] In this disclosure, a first embodiment of the method uses a direct temperature measurement at a first terminal of the thermal resistance at the power semiconductor junction with the first die attach layer 3a. The principle of this embodiment is illustrated in Figure 2A. In this principle, an AC voltage source V connected to a gate terminal G and a source or emitter terminal S / E is AC 11 and a temperature sensor 12 connected to a computerized calculation means 13 to measure the voltage V AC The temperature is measured before and after the voltage is applied, and the temperature difference T1-T0 during the life of the power module is calculated to determine the thermal resistance degradation ΔR Th Calculate.
[0043] The AC source is a low voltage or low current AC source, i.e., a peak-to-peak voltage amplitude V between 5 V and 40 V. AC, or a sinusoidal or square waveform characterized by a peak-to-peak current amplitude IAC between 2 A and 10 A, and a frequency comprised between 400 kHz and 8 MHz. This frequency is sufficient to inject current into the gate / source or gate / emitter junctions without causing or changing the conduction of the source / drain or emitter / collector junctions. The AC source may have a DC voltage component. The AC source may be created using a square waveform generated by a PWM signal. The AC source may be created using a classical gate drive circuit. The classical gate drive includes a DC power supply and a push-pull amplifier that amplifies the PWM signal.
[0044] A second embodiment of the method using indirect temperature measurement based on measuring the Thermally Sensitive Electrical Parameter (hereinafter TSEP) is shown in Figure 2B. In this embodiment, an AC source 11 is still connected to the gate and the source or emitter, but a measurement module 14, described below, is used to measure the Thermally Sensitive Electrical Parameter, together with a computerized module 15, to measure the degradation ΔR Th Measurements are performed.
[0045] 2A, 2B, the parameter K related to the junction temperature of the power semiconductor is obtained by the embedded temperature sensor 12 or the TSEP method block 14. The parameter K is directly the temperature T or the thermal sensitivity electrical parameter for calculating the thermal resistance.
[0046] Considering the junction temperatures T1 and T0, the initial thermal resistance can be calculated by the following formula: where P R is the power dissipated by the internal gate resistance under AC excitation.
[0047]
number
[0048] Measurements are taken at defined intervals d during the life of the power semiconductor and the thermal resistance degradation can be defined as follows:
[0049]
number
[0050] Therefore, throughout the measurement, P R By keeping constant, we obtain
[0051]
number
[0052] 6A and 6B show a temperature waveform 100 of the junction temperature Tj versus time, and a graph of the AC voltage V AC 1 shows waveforms 110, 120 of the AC source used to heat the internal gate resistor by applying a
[0053] With respect to temperature, in a first measurement M0 and subsequent measurements M1, ... Mn during the life of the power semiconductor, a direct measurement of Tj is made either before application of waveform signal 110 or 120 to obtain an initial temperature T0, or after or during application of the waveform signal to obtain a temperature T1.
[0054] In the case of the TSEP method, the principle of applying an AC voltage is the same, but the measurement relies on measuring the amplitude of the thermal sensitivity parameter K as a function of the temperature Tj.
[0055] In a first measurement M0, a first parameter K0 related to temperature is obtained before activating the AC source or during a first period of the AC source during which the junction temperature is in a plateau state 101 (thermal steady state) and a second parameter K1 is obtained after activating the AC source. Before activating the AC source, no power is dissipated in the power semiconductor or a certain power is dissipated in the power semiconductor by passing a current between the emitter and the collector.
[0056] In the TSEP method shown in Figure 6A, a reduced amplitude voltage pulse 111 generated by a current step using dc current source 142 and switch 141 in Figure 7A may be generated to obtain a first K0, which is taken as the TSEP voltage corresponding to the first instant of the pulse.
[0057] And the AC voltage V AC 110, the dissipated power starts to heat the internal gate resistance of the power semiconductor and the gate temperature 100 rises to a certain level 102. After a certain duration h, a parameter K1 related to the power semiconductor temperature T1 may be obtained in the same way as K0 with a pulse 112 to obtain the electrical parameter.
[0058] By choosing this duration h between 100 μs and 2 seconds, the degradation of a particular layer of the inhomogeneous structure can be identified. Importantly, the power semiconductor does not dissipate any power losses other than those caused by the internal gate resistance or dissipates a constant power loss between the emitter or source and the collector or drain during this first measurement phase.
[0059] In both the temperature measurement or TSEP method, the measurement can be repeated several times during the initial life of the power semiconductor to obtain several values of T0 and T1 or K0 and K1 for duration h in several operating conditions of the power semiconductor. After a time interval d1, which can be, for example, 100 hours or 2000 hours depending on the application, the measurement is repeated in a measurement phase Md1 to obtain T0d1 and T1d1 by direct temperature measurement or K0d1 and K1d1 by TSEP measurement. The junction temperature can be in a different plateau state than in the initial measurement. The measurement is repeated over the life of the power semiconductor. The thermal degradation measurement, which is the difference, does not depend on external factors such as the ambient temperature or the semiconductor power loss level. A first value of the overall thermal resistance or the thermal resistance of one of the substructures between the power semiconductor die and the heat sink, such as a solder layer, can be output at the end of this phase. This phase is repeated according to the application needs to refresh the thermal resistance degradation measurement.
[0060] The following description relates to using the internal gate resistance as the thermally sensitive electrical parameter TSEP. The internal gate resistance is a resistance having a first terminal connected to the gate pad of a power semiconductor and a second terminal connected to the gate electrode of a MOS semiconductor die, such as the internal gate resistance of an IGBT or MOSFET.
[0061] The internal gate resistance is a parameter that does not depend on the operating point (voltage, magnitude, frequency) of the power device, but uniquely depends on temperature. Moreover, the internal gate resistance can be conveniently located on the surface of the power semiconductor device or on the surface of the power module assembly, so that a controlled amount of heat can be generated through the power dissipated in the internal gate resistance. Heat flowing through the first terminal of the thermal resistance can be generated in separate time intervals to give a change in the resistance of the internal gate resistance according to the increase in temperature of the internal gate resistance when power is dissipated. The important point is that the internal gate resistance does not undergo degradation during the life of the power module, and therefore a controlled power dissipation and heat flow through the thermal resistance can be realized by applying the same AC voltage or current from the beginning of the semiconductor module product life to the end of the semiconductor module life, and measuring the resulting temperature increase through the measurement of the resistance value.
[0062] Advantageously, therefore, no further power or heat flux measurement probes are required, since the power dissipated by the internal gate resistor remains the same over the entire life of the power module assembly. As mentioned before, heat can be generated simply by using the available control terminals of the power semiconductor, e.g. the gate and source / emitter terminals of a power semiconductor transistor, to perform the measurement. Thus, the measurement system does not need to deal with high voltage or high current junctions (collector / emitter or drain / source junctions). Finally, R Th Since the degradation can be measured for any temperature operation of the power semiconductor, online thermal resistance degradation monitoring is possible.
[0063] Returning to FIG. 6A, the TSEP-based method may further be a combination of a pulsed current 110 through the gate of the semiconductor and a voltage measurement circuit that measures the voltage evolution that depends on the internal gate resistance and therefore on the junction temperature. In FIG. 6A, the voltage measurement is performed using trapezoidal pulses 111, 112 before and after applying the AC voltage 110. In this embodiment, the trapezoidal pulses 111, 112 may be generated using the current generator disclosed in FIG. 7A. The trapezoidal pulses are the result of a square current generator injecting a current step signal through the equivalent circuit of FIG. 7B.
[0064] In an alternative embodiment according to Figures 6B and 7C, the TSEP method can consist in having an AC voltage source in series with the measurement resistor and acquiring the voltage across the measurement resistor, which voltage depends on the internal gate resistance and therefore on the junction temperature. Measurements may then be taken at the first peak level 121 and the last peak level 123 of the signal 120. With this implementation, R Th The number of additional devices required for degradation measurements is reduced.
[0065] As will be shown later, it is not necessary to pre-calibrate the exact junction temperature in order to measure the thermal resistance degradation, as is typically the case for internal gate resistors, since the TSEP has a linear relationship to the junction temperature. Using this method, the need for a junction temperature sensor is eliminated.
[0066] In the TSEP method, the internal gate resistance is both a temperature sensitive electrical parameter and a heat source. The dependence of the internal gate resistance on the junction temperature Tj can be written as
[0067]
number
[0068] Here, δR g is the linear resistance change with temperature, typically between 0.5mΩ / ℃ and 50mΩ / ℃. R g0is the internal gate resistance at 0°C, for example, 0.5Ω to 50Ω. In claim 2, the parameter related to the junction temperature Tx (x=0, 1, 0d, 1d) of the power semiconductor is Rx. Therefore, Tj x is proportional to Rx, which is also proportional to Kx or Kvx, as measured by the TSEP-based method.
[0069] To further illustrate, an AC source may have an RMS value of V AC The power dissipated on the internal gate resistor P R is expressed by the following equation:
[0070]
number
[0071] As an example, R Th =(Tj1-Tj0) / P R Rx is proportional to Kx or Kvx. Therefore, R Th can be calculated as follows:
[0072]
number
[0073] Due to the deterioration of one or more of the substructures that make up the thermal resistance, the thermal resistance decreases to a value R Th From the value R Thd Therefore,
number
[0074]
number
[0075] The values of the elements Lin and Cin can be measured for each type of power semiconductor. They are usually listed in the datasheet of the power semiconductor power module. Note that this equation does not include the calibration parameter δR. This indicates that the TSEP calibration procedure is not necessary to measure the change in thermal resistance.
[0076] Concerning the creation of the AC source, Fig. 3 shows a possible embodiment in which the AC source is made using a square waveform generated by a PWM signal 21, a bipolar DC power supply constituted by two voltage sources 20a and 20b, and push-pull amplifiers 22, 23 that amplify the PWM signal. In this embodiment, the DC power supplies 20a, 20b can be referenced in order to keep the gate / source or gate / emitter voltages below 1 V to avoid conduction of the drain / source or emitter / collector junctions when measurements are made offline or during the non-conducting state of the power semiconductor.
[0077] Concerning the creation of the AC source, Fig. 4 shows a possible implementation where the AC source is made using a classical gate drive circuit with positive 37 and negative 38 polarization of the gate driver switches 22, 23. A heating PWM generator with a triangular signal generator 31 having a high frequency f1 is compared through a comparator 32 with a heating duty cycle 30, usually comprised between 0.05 and 0.2. The comparator output generates an AC signal 34 as a heating command which is combined with the user's control signal 33 through an OR gate 36.
[0078] The acquired signals are shown in Figures 5A and 5B. Figure 5A shows a waveform 51 of the user control signal 33, a waveform 52 of the Vheating signal 32 which is the output of the comparator 32, a waveform 53 of the driver output signal at the gate to source / emitter semiconductor terminals, a waveform 54 of the gate to source / emitter voltage Vgs or Vge, and a waveform 55 of PowerRg which is the power dissipated in the internal gate resistance.
[0079] Figure 5B shows the same signal in the expanded W window of figure 5A, showing that during a low logic of Vcontrol 56, a Vdriver pulse 58 is given by a Vheating pulse 57. The Vdriver pulse 58 causes a Vgs or Vge pulse 59 to be narrow enough, controlled by the heating duty cycle 30, to avoid conduction of the drain / source or emitter / collector junctions, so that only power dissipation in the internal gate resistance of the power semiconductor is allowed. Furthermore, the off-state of the power semiconductor required by the user's control signal 33 is taken into account. Advantageously, the heating phase is transparent to the user.
[0080] A design using XOR gates instead of OR gates is also possible, in which case the Vheating signal 52 is superimposed on the same user control 51 to obtain the Vdriver signal 53', as shown in FIG. 5C. A high frequency pulse train of f1 is present at positive and negative values of Vgs 54', resulting in a drain source conduction phase in the upper part of the pulse train 54'a and no drain / source conduction in the lower part of the pulse train 54'b. In waveform 55', the power dissipated in the gate resistor remains independent of the power dissipated in the drain / source junction, although it is dissipated over the entire segment. Furthermore, the semiconductor state required by the user control signal 33 is taken into account in the on and off states. Adjustments can be made to the heating duty cycle 30 during the on and off states in order to keep the Vgs voltage at a particular level. Advantageously, the heating phase is transparent to the user.
[0081] By means of an embedded temperature sensor or a TSEP method block, parameters related to the junction temperature of the power semiconductor are obtained.
[0082] Regarding the voltage measurement circuit that measures the voltage evolution that depends on the internal gate resistance and therefore on the junction temperature in the TSEP-based method, examples of the measurement circuit are shown in Figures 7A and 7C, while Figure 7B is an equivalent gate input circuit for the circuit of Figure 7A.
[0083] In FIG. 7A, a dc current source 142 is connected in parallel with a switch 141. The AC source 11 is deactivated via a heat command 30, while the switch 141 is opened via a measure command 143 to provide a peak-to-peak current amplitude I m , a current in the shape of a step current waveform is injected into the gate / source pad of the power semiconductor and up to the internal gate resistor. Since the AC source is deactivated, the voltages K0 and K1 measured across the current source are the internal gate resistance times the peak-to-peak amplitude of the injected current I m It is.
[0084] The equivalent input circuit of the MOS gate input impedance is shown in FIG. 7B, along with the parasitic inductance Lin, the internal gate resistance R, and the parasitic input capacitance Cin of the MOS.
[0085] In FIG. 7C, a resistor 144 having a value Rmeas is connected between the source / emitter of the power transistor and the AC source 11. A high level V of 5V to 40V as the pulse 120 in FIG. 6B appears on the gate-emitter junction. AC 6B, the first peak level 121 of the AC source, V AC First, we use Kv0 value and R0=Rmeas (V AC Kv0 takes the value Vmeas 122. Then, the last peak level 123 is used to calculate the Kv1 value and the associated gate resistance, R1=Rmeas·(V AC The same procedure is used in subsequent measurements to calculate the ΔR Th In this example, the resistances R0 and R1, which are thermal sensitivity parameters, can be obtained by the intermediate parameters Kv0 and Kv1. Th The number of additional devices required for degradation measurements is reduced.
[0086] Regarding the frequency of the AC source and the equivalent gate input circuit of FIG. 7B, advantages may be obtained according to some embodiments.
[0087] f1<0.33 (2 π Cin R) -1 By choosing a frequency such as , the thermal resistance degradation can be written as:
[0088]
number
[0089] Using this formula, the thermal resistance degradation of a power semiconductor can be calculated over the entire temperature range with an estimation error relatively less than 5% (with typical values such as δR / R0<1000 ppm / °C).
[0090] As shown in this equation, only the absolute value of the internal gate resistance needs to be measured to determine the thermal resistance degradation. Thus, neither measurement nor estimation of the junction temperature is required, and no TSEP calibration procedure is required.
[0091] In another embodiment, the frequency f1 is set to the input resonant frequency f1 of the power semiconductor, f1=(2·π·Lin·Cin) -2 Select as.
[0092] Thus, the measurements are not adversely affected by parasitic impedances associated with the voltage source, such as parasitic connection inductances, and the sensitivity of temperature related parameters is maximized.
[0093] And the thermal resistance degradation can be estimated by the following equation:
[0094]
number
[0095] (2*π*Lin*Cin) -2For a frequency equal to , the thermal resistance degradation can be expressed by the following equation:
[0096]
number
[0097] In another embodiment, as the frequency decreases, less power is dissipated in the internal gate resistor, so for a given V AC The temperature rise relative to Th The signal-to-noise ratio for degradation measurements decreases. To remain at a sufficient signal-to-noise ratio, the value f1 may be chosen such that the signal-to-noise ratio for the internal gate resistor is higher than the specified detection limit. In the example, f1 is chosen such that, for example, f1>M·(2·π·Cin·R) -1 where M is 0.1.
[0098] The thermal resistance degradation calculation may be performed in a module 13, 15 having a CPU or a specific FPGA with memory including program memory, calculation memory, and storage memory and communication means, and having sufficient processing power to obtain parameters related to the junction temperature of the power semiconductor, such as the Tj sensor value or the TSEP value, calculate the difference, compare the measurements, and output the thermal resistance change.
[0099] FIG. 8 shows a thermal resistance R between a power semiconductor in a power module assembly having an internal gate resistance and a heat sink of the power module assembly. Th This shows the principle of the method for measuring the deterioration of
[0100] The method comprises the steps of: Measuring (200) a first initial parameter K0 related to a junction temperature 101 of the power semiconductor at a stable operating temperature; An AC voltage V of an AC source with frequency f1 is applied to the gate / source junction of the power semiconductor. ACheating (210) the internal gate resistor through application of voltages 11, 110, 120 and measuring (220) a second initial parameter K1 related to the junction temperature 102 after the heating; and storing (230) the first initial parameter K0 and the second initial parameter K1 in a memory.
[0101] This allows obtaining a baseline value for an initial parameter, which may be a temperature or temperature sensitive electrical parameter. This may be done multiple times to obtain multiple baseline values as described below. After the initial stage described above, the method continues as a subsequent process after at least one distinct time interval d∈{d1,...,dn} during the life of the power module assembly 240: Stable operating temperature T0 d ∈{T0 d1 ,...T0 dn}(101d1,...,101dn) the first subsequent parameter K0 related to the junction temperature of the power semiconductor d ∈{K0 d1 ,...,K0 dn} (250), and An AC voltage V with frequency f1 is applied to the gate / source junction of the power semiconductor. AC The internal gate resistor is heated (260) through application of the d ∈{T1 d1 ,...T1 dn}(102d1,...,102dn) related second successor parameter K1 d ∈{K1 d1 ,...,K1 dn} (270), and may also include
[0102] This allows measurements of the same parameters during the life of the power semiconductor, which in turn allows the subsequent parameter K0, which is related to the junction temperature. d ∈{K0 d1 ,...,K0 dn} and K1 d ∈{K1 d1,...,K1 dn} and the initial parameters K0 and K1 read from the memory, the thermal resistance degradation ΔR Th Calculating
number
number
[0103] Once the calculations are done, the method Th Limit value ΔR Th max, and ΔR Th exceeds the limit and a fault flag 310 is set.
[0104] ΔR Th If d remains below the limit value, the method includes repeating the subsequent process with the same d time interval or with a new d time interval.
[0105] In the example, d is ΔR Th For example, ΔR Th If is less than +10%, the d interval may be increased, while ΔR Th If increases, for example, exceeds +10%, the d-spacing may be decreased.
[0106] If a fault is flagged, further tests may be performed, such as tests to detect the location of the fault in the layer between the die and the heat sink through measurements with different durations h for heating the gate / source or gate / collector junction.
[0107] As mentioned before, when the measurement is made through a temperature sensor, the parameter KX related to the junction temperature of the power semiconductor is y is the junction temperature TX of the power semiconductor y where X is equal to 0 or 1, y is null or d, and the parameter KX related to the junction temperature of the power semiconductor when the measurement is the TSEP method based on the internal gate resistance measurement. y is the internal gate resistance RX of the power semiconductor y where X is equal to 0 or 1 and y is null or d.
[0108] In a particular embodiment, subsequent measurements are performed with an internal gate resistance R0 equal to within a particular range + / - k of the initial internal gate resistance R0. d In this way, the thermal junction resistance degradation can be estimated with an error below 5% for any frequency f1 without prior knowledge of any of the power semiconductor parameters.
[0109] For example, k is 0.1 * It is R0.
[0110] In a further embodiment, in order to avoid waiting for a specific temperature to start the measurement, the temperature T (T) Or resistor R0 (T) The initial parameter K0 (T) and temperature T1 (T) or resistor R1 (T) The corresponding initial second parameter K1 (T) A table of T0d and R0d and T1d or R1d can be built during the initial part of the life of the power semiconductor. This allows the closest initial temperature T0 in the table to be found when subsequent T0d or R0d and T1d or R1d measurements are made. (T) Or initial resistance R0 (T) This is to calculate the ΔT or ΔR value and compare it with the initial value to determine the deterioration of the thermal resistance.
[0111] As an example of this embodiment, the following temperature table T0 (T) , T1 (T)The power module assembly was constructed when its usage time was less than 1000 hours and was operated under various loads and operating temperatures. (T) A number of stable junction temperatures T1...Tn of the power semiconductor are stored in memory.
[0112] Such a table can be expressed using temperatures as shown below.
[0113] [Table 1]
[0114] Using such a table, the temperature T1 after heating is calculated to calculate the thermal resistance deterioration. d Initial T1 (T) As a basis for comparison, the measured temperature T0 after a delay d d to the closest T0 (T) This simplifies measurements when testing power semiconductors under various loads. In an example, a table may be built using 10°C increments or more for temperatures from ambient to 150°C. The table may be built for the first 1000h of operation of the power semiconductor.
[0115] Such an embodiment is shown in Figure 9, where a generic parameter p is used instead of a measured temperature T. The process includes, for each p value 400, measuring K0(p) (410), heating with a voltage VAC (420), measuring K1(p) (430), and forming a table of K0 and K1 values for the list of parameters p1,...,pn at 440.
[0116] Once the table is formed, the process returns to step 240 of FIG. 8, and step 275 is performed to select values of p that give K0(p) and K1(p), such that the difference with K0d is below a minimum value, after the sequence of measurement 250, heating 260, and measurement 270, and ΔR Th This is introduced before the calculation 280.
[0117] In a similar manner, in order to improve the ability to detect which layer of the power module assembly between the die of the power semiconductor and the heat sink, several heating durations h for layers with different time constants between the power semiconductor and the heat sink are included, and an initial first parameter K0 measured for the heating duration h is calculated. (T,h) and the second parameter K1 (T,h) In such a case, a measurement after period d is made using the maximum h duration in the table to obtain ΔR Th When it is detected that is larger than the specified value, a test is performed with a smaller value of h to determine which layer is R Th Find out which shows the greatest increase.
[0118] With respect to the AC source, the AC voltage source preferably has a DC voltage component or polarization such that the maximum Vge or Vgs voltage value is not higher than the flat band voltage of the power semiconductor, e.g., −3V for Vgs in FIG. 5A.
[0119] The measurement therefore does not change the current switching state of the power semiconductor and the gate capacitance Cin is independent of the bus voltage. Advantageously, the measurement may further be performed while the bus voltage is applied across the source / drain or emitter / collector of the power semiconductor device or, more generally, while the bus voltage is applied on the power converter terminals to which the power semiconductor device is attached.
[0120] Regarding the positioning of the temperature sensor, it should be noted that critical layers in a semiconductor assembly rarely degrade in a non-uniform manner. For example, in some cases, cracks may propagate from the edge of a layer to the center of the layer, while in other cases the layer degrades preferably at the hottest location, usually the center of the layer under the semiconductor chip. These degradation mechanisms are related to the material selection and are usually inherent to the module design.
[0121] Furthermore, in order to improve the sensitivity of the measurement method and detect degradation at an early stage during the module's life, it is preferable to position the mounted temperature sensor so that the starting point of degradation is in the thermal path between the mounted temperature sensor and the heat sink. Thus, the type of degradation propagation can be known in advance by laboratory tests. In the laboratory tests, the optimal position of the mounted temperature sensor is selected so that the starting point of degradation is located in the thermal path between the mounted temperature sensor and the heat sink. For example, when degradation of the interconnect layer under the semiconductor chip is the most important layer to be monitored, this position can be the edge of the semiconductor chip when degradation starts at the hottest position, i.e., under the center of the semiconductor, when degradation is in a crack that propagates from a position under the edge of the semiconductor to a position under the center of the semiconductor or in a crack in the center of the semiconductor chip.
[0122] The invention is not limited to the described examples and in further embodiments the location of the series resistor in the embodiment of FIG. 7C may be varied relative to the series resistor on the gate side of the power semiconductor, i.e. on the boundary of the die or in the central part of the die, and may be formed of several series or parallel resistors arranged at various locations on the die.
Claims
1. A thermal resistance R between a power semiconductor in a power module assembly, the power semiconductor having an internal gate resistance, and a heat sink of the power module assembly. Th A method for measuring degradation of Measuring a first initial parameter K0 related to a junction temperature of the power semiconductor at a stable operating temperature; An AC voltage V of an AC source with a frequency f1 is applied to the gate / source junction of the power semiconductor. AC and measuring a second initial parameter K1 related to the junction temperature after the heating. storing the first initial parameter K0 and the second initial parameter K1 in a memory; Including, After at least one distinct time interval d ∈ {d1, . . . , dn} during the life of the power module assembly, as a subsequent process: Stable operating temperature T0 d ∈{T0 d1 , . . . T0 dn} a first subsequent parameter K0 related to the junction temperature of the power semiconductor d ∈{K0 d1 , . . . , K0 dn}; and The AC voltage V with frequency f1 is applied to the gate / source junction of the power semiconductor. AC The internal gate resistor is heated by applying a voltage to the internal gate resistor, and the junction temperature after heating is T1 d ∈{T1 d1 , . . . T1 dn} the second subsequent parameter K1 d ∈{K1 d1 , . . . , K1 dn}; and The subsequent parameter K0 related to the junction temperature d ∈{K0 d1 , . . . , K0 dn} and K1 d ∈{K1 d1, . . . , K1 dn} and the initial parameters K0 and K1 read from the memory, the thermal resistance deterioration ΔR Th Calculating [0010] and R Th = (K1 0 -K0 0 ) / P R and [0025] , d=d1, . . . , dn, and P R is the AC voltage V AC the power dissipated by the internal gate resistor under ΔR Th The limit value ΔR Th Compare with max, ΔR Th If ΔR exceeds the limit value, a fault flag is raised or Th repeating the subsequent process if remains below the limit value; Including, The frequency f1 is the reciprocal of the approximate gate input power semiconductor electrical time constant of the power semiconductor (2 π Cin R) -1 Way lower than that.
2. The time interval d ∈ {d1, . . . , dn} between measurements of thermal resistance is defined as the first subsequent parameter K0 d The method of claim 1 , wherein k is selected to remain within a specified range K0+ / -k relative to K0.
3. A first parameter K0 at an initial stage in a specified initial part of the life of the power semiconductor for several stable junction temperatures T1...Tn of the power semiconductor during operation. (T) and the second parameter K1 (T) The table is stored with the initial value K0 (T) and K1 (T) is the closest K0 to the measurement at the end of the interval d d(T)The method of claim 1 or 2, comprising selecting within the specified range:
4. The method of claim 3, wherein the specified initial portion of the life of the power semiconductor is prior to the first 1000 to 1500 hours of operating time of the semiconductor.
5. 5. The method of claim 1, wherein the AC voltage V AC has a DC component to polarize the gate with respect to the source or emitter of the power semiconductor and has a peak voltage less than or equal to a flatband voltage of the power semiconductor.
6. A thermal resistance R between a power semiconductor in a power module assembly, the power semiconductor having an internal gate resistance, and a heat sink of the power module assembly. Th A method for measuring degradation of Measuring a first initial parameter K0 related to a junction temperature of the power semiconductor at a stable operating temperature; An AC voltage V of an AC source with a frequency f1 is applied to the gate / source junction of the power semiconductor. AC and measuring a second initial parameter K1 related to the junction temperature after the heating. storing the first initial parameter K0 and the second initial parameter K1 in a memory; Including, After at least one distinct time interval d ∈ {d1, . . . , dn} during the life of the power module assembly, as a subsequent process: Stable operating temperature T0 d ∈{T0 d1 , . . . T0 dn} a first subsequent parameter K0 related to the junction temperature of the power semiconductor d ∈{K0 d1 , . . . , K0 dn}; and The AC voltage V with frequency f1 is applied to the gate / source junction of the power semiconductor. ACThe internal gate resistor is heated by applying a voltage to the internal gate resistor, and the junction temperature after heating is T1 d ∈{T1 d1 , . . . T1 dn} the second subsequent parameter K1 d ∈{K1 d1 , . . . , K1 dn}; and The subsequent parameter K0 related to the junction temperature d ∈{K0 d1 , . . . , K0 dn} and K1 d ∈{K1 d1 , . . . , K1 dn} and the initial parameters K0 and K1 read from the memory, the thermal resistance deterioration ΔR Th Calculating [0030] and R Th = (K1 0 -K0 0 ) / P R and [0045] , d=d1, . . . , dn, and P R is the AC voltage V AC the power dissipated by the internal gate resistor under ΔR Th The limit value ΔR Th Compare with max, ΔR Th If ΔR exceeds the limit value, a fault flag is raised or Th repeating the subsequent process if remains below the limit value; Including, The frequency f1 is the input resonance frequency f1 of the power semiconductor = (2 x π x Lin x Cin) -2 Select as,method.
7. A thermal resistance R between a power semiconductor in a power module assembly, the power semiconductor having an internal gate resistance, and a heat sink of the power module assembly. Th A method for measuring degradation of Measuring a first initial parameter K0 related to a junction temperature of the power semiconductor at a stable operating temperature; An AC voltage V of an AC source with a frequency f1 is applied to the gate / source junction of the power semiconductor. AC and measuring a second initial parameter K1 related to the junction temperature after the heating. storing the first initial parameter K0 and the second initial parameter K1 in a memory; Including, After at least one distinct time interval d ∈ {d1, . . . , dn} during the life of the power module assembly, as a subsequent process: Stable operating temperature T0 d ∈{T0 d1 , . . . T0 dn} a first subsequent parameter K0 related to the junction temperature of the power semiconductor d ∈{K0 d1 , . . . , K0 dn}; and The AC voltage V with frequency f1 is applied to the gate / source junction of the power semiconductor. AC The internal gate resistor is heated by applying a voltage to the internal gate resistor, and the junction temperature after heating is T1 d ∈{T1 d1 , . . . T1 dn} the second subsequent parameter K1 d ∈{K1 d1 , . . . , K1 dn}; and The subsequent parameter K0 related to the junction temperature d ∈{K0 d1 , . . . , K0 dn} and K1 d ∈{K1 d1 , . . . , K1 dn} and the initial parameters K0 and K1 read from the memory, the thermal resistance deterioration ΔR Th Calculating [0050] and R Th = (K1 0 -K0 0 ) / P R and [006] , d=d1, . . . , dn, and P R is the AC voltage V AC the power dissipated by the internal gate resistor under ΔR Th The limit value ΔR Th Compare with max, ΔR Th If ΔR exceeds the limit value, a fault flag is raised or Th repeating the subsequent process if remains below the limit value; Including, The frequency f1 is (1 / 10) (2 π Cin R) -1 Way higher than that.
8. A parameter KX related to the junction temperature of the power semiconductor y is the internal gate resistance RX of the power semiconductor y 8. The method according to any one of claims 1 to 7, wherein X is equal to 0 or 1 and y is null or d.
9. 9. The method of claim 8, comprising injecting a square DC current through the internal gate resistance of the power semiconductor without applying the AC voltage V AC to obtain information of the internal gate resistance.
10. Applying a first AC voltage Vac1 with a first peak-to-peak value to obtain an R0 value or an R0 d a second AC voltage V with a second peak-to-peak value greater than the first peak-to-peak value; AC to heat the internal gate resistor of the power semiconductor, and R1 value or R1 d The method of claim 9, further comprising obtaining a value.
11. providing a resistor with a value R in a circuit including the gate / emitter or gate / source of the power semiconductor and the AC source; and AC as a pulse of 5V to 40V to heat the internal gate resistor, and measuring a first voltage Vmeas across the resistor as a first parameter Kv0 at a first peak level of the AC source to obtain a value of R0=Rmeas·(V AC / K v0 -1), and measuring a second voltage Vmeas at the end of the resistor as a second parameter Kv1 at the last peak level of the AC source to obtain the value of R1 as R1=Rmeas·(V AC / K v1 -1), and repeat the measurement and calculation to obtain R0 d Value and R d and obtaining the value.
12. A parameter KX related to the junction temperature of the power semiconductor y is the junction temperature TX of the power semiconductor y 8. The method according to any one of claims 1 to 7, wherein X is equal to 0 or 1 and y is null or d.
13. The AC voltage V AC The method according to any one of claims 1 to 12, wherein the control signal for the power component is modulated by the heating signal.
14. A control device for a power semiconductor, the control device comprising: AC and means for measuring the temperature or a thermal sensitivity parameter of said power semiconductor at a selected point in the life of said semiconductor, and the resulting change in thermal resistance ΔR Th and computer means for calculating over the lifetime of the semiconductor by the method according to any one of claims 1 to 13.
15. The thermal sensitivity parameter is the internal gate resistance of the power semiconductor, and the control device is AC 15. A control device for power semiconductors as claimed in claim 14, comprising voltage sensing means in the source branch of the injection circuit.
Citation Information
Patent Citations
Deterioration determining method for electric power semiconductor element
JP2002005989A
Inverter
JP2003009541A
Power conversion apparatus and power cycle life prediction method
JP2008271703A
Semiconductor device and failure detection method
JP2017017822A
Power conversion device and heat resistance measurement method for power module
JP2017135868A
