晶闸管的瞬态热阻抗及其结温温升的研究

系统地阐述了晶闸管的瞬态热阻抗,针对晶闸管承受不同的功率冲,利用不同的计算方法计算瞬态热阻抗值,从而确定结温温升,并比较了它们的原理、精度及其应用范围。特别是利用了一种新的瞬态热阻抗计算模型,当晶闸管承受持续时间极短且周期、占比均变化的任意波形功率脉冲时,它能够比较精确地分析晶闸管的热特性,计算半导体结的最大温升。     

基于IGBT模块饱和压降温度特性的结温探测研究

实验室条件下,IGBT模块的结温探测是瞬态热阻抗测试的关键。首先分别在热稳态和热瞬态下证明了饱和压降温度特性只与芯片有关,然后建立了IGBT模块结温探测模型,利用饱和压降值和集电极电流值来计算结温值,并将用模型计算出的结温与光纤实测的结温相比较,吻合性良好,证明了模型计算法能够准确探测结温。该方法可以用于恒流加热过程中瞬态热阻抗的测量,比起热敏参数法中冷却过程测量瞬态热阻抗相比,更具有实际意义。     

结型LED的热特性分析与工程设计

文章从半导体结型发光器件热特性的概念出发，导出了ＬＥＤ各种瞬态工作状态下输出功率与热特性的概念出发，导出了ＬＥＤ各种瞬态工作状态下输出功率与热特性参数的关系。分析讨论了ＬＥＤ输出光功率对不同电流脉冲的响应特性以及ＬＥＤ热特性的设计方法。     

SOI-LIGBT器件正偏安全工作区的研究

研究了高压SOI—LIGBT器件的单脉冲正偏安全工作区（FBSOA）的绘制方法。首先,分别提取器件的热阻与热容,并借助热阻抗公式计算器件的瞬态热阻抗,进而根据热阻抗与最大允许功率之间的关系,获得直流与脉冲条件下的正偏安全工作区。同时,还对其二次击穿边界与安全工作区的相互关系做了进一步分析,为更加准确地确定高压LIGBT器件的单脉冲安全工作区提供理论指导。     

不同脉冲工作条件对GaN HEMT器件结温的影响

研究了不同脉冲工作条件对GaN高电子迁移率晶体管(HEMT)结温的影响,通过改变脉冲信号发生器的输出频率和占空比来改变器件的工作条件,利用具备高空间分辨率的显微红外热像仪进行瞬态结温测试。结果表明:器件工作在给定的平均功率下,可以通过提高脉冲信号占空比和频率来改善器件的寿命和性能可靠性;工作在给定的峰值功率下,可以通过降低脉冲占空比和提高脉冲频率来改善器件的寿命和性能可靠性。沟道温度影响着半导体器件的寿命,因此,可以在器件能够承受的范围内通过改变脉冲占空比和脉冲频率来改善器件的寿命和性能可靠性。     

结温在线控制系统的IGBT功率模块热耦合模型

应用有限元法，对IGBT功率模块的三维热分布进行了仿真研究，得到了器件的稳态热阻及瞬态热阻抗。研究了功率模块各芯片之间的相互热影响，提出了热耦合效应热模型的统一结构，基于对瞬态热阻抗曲线的拟合，得到热模型的相关参数，从而建立了热耦合效应热模型。以一个降压变换器为例，阐述了结温在线控制系统的工作原理，并将热模型应用于该系统中，计算结果与测量结果非常一致。     

电磁脉冲加热Debye媒质的快速计算

本文提出了一种计算电磁脉冲加热Debye媒质的快速算法。从麦克斯韦方程组和Debye方程出发,重新定义了Debye媒质中的瞬态耗散功率。该定义能有效避免传统定义中瞬态耗散功率为负的情况。通过时间尺度变换,把电磁场的计算和热场的计算统一到同一时间尺度下,这样就可以节省大量的计算时间。另外,研究了长时间脉冲加热的效率和均匀性。结果表明,在同样的重复频率下,脉冲加热的效率与均匀性均低于连续微波。     

脉冲半导体激光器端面抽运Nd:YAG晶体瞬态热分析

为了研究脉冲半导体激光器端面抽运激光晶体产生的热效应,对激光晶体瞬态温度场以及热形变场进行解析分析与计算。考虑到脉冲LD出射光具有超高斯分布,且Nd:YAG晶体热传导各向同性的特点,利用热传导Poission方程得到了超高斯分布脉冲LD端面抽运Nd:YAG晶体瞬态温度场以及热形变场的一般解析表达式,定量分析了单脉冲抽运过程中超高斯抽运光光斑半径及超高斯阶次、脉冲宽度对Nd:YAG晶体瞬态温场的影响以及准热平衡状态温度场的时变特性。结果表明,当脉冲LD端面抽运光具有3阶超高斯分布、抽运功率为80W、脉冲频率为100Hz、脉宽为200 s、钕离子掺杂质量分数为0.01的Nd:YAG晶体瞬态温度场随抽运脉冲呈现出周期性分布,准热平衡状态的温度在25.5℃到29.2℃之间成锯齿形周期分布;晶体抽运面的热形变量在0.13m和0.19m之间也呈现出周期性变化。该研究对于脉冲LD端面抽运全固态激光器热不敏谐振腔设计具有理论指导意义。     

采用动态差分模型表征脉冲驱动LED电热特性

为了研究不同驱动方式下LED结温的差异性,依据LED热结构模型,从功率传递角度构建了差分型LED电-热动态模型,利用相同平均功率、不同频率的注入电功率,计算脉冲驱动条件下的LED结温.同时,采用正向电压法测量实际结温,对模型预测结温的准确性进行了分析.研究表明,差分模型能够表征脉冲驱动方式下LED电热特性,其理论预测值与实测值误差约为4℃,基本能够满足工程应用要求.更有意义的是,采用差分模型,能够分析在特定驱动条件下,LED热容、热阻等参数对结温的影响,为不同驱动模式下LED个性化结构设计提供了理论基础.     

一种用于功率模块热分布特性研究的精确模型

应用有限元法 ,对一个 IGBT功率模块的三维热分布进行了仿真研究 ,提出了通过 ANSYS仿真建立热模型的基本方法 ,进而探讨了功率模块上各芯片之间的热耦合关系 ,提出了考虑热耦合效应在内的功率模块热模型的统一结构 ,基于对瞬态热阻抗曲线的拟合 ,获得了热模型的相关参数 ,从而建立了热耦合模型 .该模型可方便地应用于电路仿真软件如 PSPICE中 ,仿真结果与有限元计算结果一致 ,并与实际测量值相符 .     

The current pulse ratings of thyristors

This paper describes how the current pulse ratings of thyristors during turn-on spreading can be obtained from the maximum allowable junction temperature based on time-to-failure using a combination of analytical and experimental methods. The instantaneous junction temperature rise of a thyristor is analyzed with the aid of a digital computer. In the calculation of temperature rise, the transient thermal impedance during turn-on spreading is considered. Analyzed results agree with the values obtained directly by an infrared detector. In order to estimate the maximum allowable junction temperature based on time-to-failure, "step-stress capability tests" were conducted. In many cases, it was found that there were modes of both catastrophic and degradation failure. The maximum allowable junction temperature was estimated for the nonrepetitive and repetitive current pulse ratings.     

Transient thermal response of power semiconductors to short powerpulses

Thermal response curves used to calculate the peak junction temperature of power semiconductors are normally derived by experimental identification of the parameters of a known model. Unfortunately the model, developed many years ago, is inappropriate for large surges of short time duration, as they are encountered in present day power conditioning systems. An alternative model is derived, the limits of its accuracy are estimated, and a correction factor is described. A verification of the accuracy of the two methods is also presented. For pulse widths shorter than the thermal transit time, which is in the order of 300 μs, the peak junction temperature can be more accurately calculated with an expression derived in the present work, which takes into consideration the active volume in which the heat is generated, than with the transient thermal response curve. A correction factor, a function of the width of the pulse, inserted in this equation, further improves its accuracy     

Modulation method for measuring thermal impedance components of semiconductor devices

The paper describes the modulation method of measuring the thermal impedance of semiconductor devices as well as its implementation. In contrast to the standard method (JESD51-1 standard) which requires heating the device under test by the stepped power, the modulation method uses heating power modulated harmonically. A pulse sequence of heating current, with the pulse length varying harmonically, is passed through the device under test. The p-n junction temperature is measured through a temperature-sensitive parameter, namely a forward voltage drop on the p-n junction between heating pulses at low measuring current. First harmonic of the p-n junction temperature oscillation is determined by the discrete Fourier transform, which allows to determine thermal impedance absolute value and phase at modulation frequency of heating power. An analysis of the dependence of thermal impedance on modulation frequency allows to determine thermal impedance components corresponding to the structural elements of the device under test. Numerical simulation shows that the thermal resistance components on the Foster's network may be determined at the modulation frequencies corresponding to the first derivative minima of the thermal impedance of the real part of frequency dependence. The main characteristics of the device that implements the method are described.     

The effect of electro-thermal parameters on IGBT junction temperature with the aging of module

This paper presents the effect of the change of electro-thermal parameters on IGBT junction temperature with module aging. Five IGBT modules are subjected to advance power thermal cycling, and IGBT I–V characterization, switching loss, and transient thermal impedance curve are measured every 1000 power thermal cycles. Then, electro-thermal models of IGBT module under power thermal cycles were built by change electro-thermal parameters, and the influence of various parameters of the electro-thermal model on the junction temperature was researched respectively. Experimental results demonstrate that IGBT collector-emitter voltage, switching loss and thermal resistance increase more quickly with the aging process of module. Simulation results indicate that the variations of electro-thermal parameters have crucial influences on the IGBT junction temperature. After 6000 power thermal cycles, the IGBT steady state junction temperature mean and variation are increased 1.97 K and 0.1656 K over its initial value, respectively. The relative temperature rise is 38.10% and relative temperature variation is 15.08% after 6000 power thermal cycles. The rise in switching loss increases both the steady state junction temperature mean and variation. The change of thermal impedance has great influence on the steady state junction temperature mean, but has little effect on steady state junction temperature variation.     

Novel techniques and procedures for the assessment of fault currentwithstand capability of power thyristors

In applications using high-power thyristors, the designer has to make sure that the selected thyristor will withstand stresses caused by overloads and fault currents. If the surge current characteristics found in the thyristor data sheet do not provide sufficient information, he has to find the transient excursions in junction temperature that will be caused by the worst expected fault current and then make a judgment on whether or not they can be tolerated. The standard way of predicting changes in junction temperature due to a known current waveshape is to determine the corresponding power loss using the on-state (conduction) characteristic and then find the time trace of the junction temperature using the curve of transient thermal impedance. The calculation procedure based on a superposition method has been in use for some 40 years. An improvement based on current state-of-the-art computer software is overdue. However, the main problem facing the designer is that the information found in contemporary data sheets is often neither sufficient for a meaningful calculation nor for deciding whether or not the calculated temperature excursions can be tolerated. This paper deals with three subjects. First, it shows the application engineer how to use off-the-shelf computer software for more accurate and much easier prediction of junction temperature excursions. Second, it advises what to do with the results. Finally, it points to the pieces of information which are needed in the process and which should therefore be found in data sheets of all high-power thyristors. The proposed method for calculation of temperature excursions in high-power thyristors is also applicable to other electrical apparatus such as ZnO arresters, transformers, electric machines, etc     

Analysis of the temperature rise in PIN diodes caused by microwave pulse dissipation

With the increasing use of PIN diodes in high power, microwave switching components, knowledge of the diode transient thermal response to such pulse dissipation becomes of increasing importance. Thermal models for the microwave dissipation within the diode are proposed for both the forward and reverse biased states. Based on these models, both the maximum and junction temperature step responses and pulse waveforms are determined for both bias states. For the forward bias state, the maximum temperature occurs near the center of the dissipation or I-region and for a step increase in dissipation, initially increases linearly with time then levels off to a constant difference with respect to the junctions. Under the same conditions, the junction temperature (relative to an appropriate heat sink) initially increases linearly also followed by a square root dependence to the steady-state value. The total response is therefore characterized by two fundamental thermal time constants and thermal resistances corresponding to the dissipative and conductive regions. Of major importance is the substantial calculated temperature differences which can occur between the center of the I-region and the junction where the temperature is normally experimentally monitored. For the reverse bias state, the maximum temperature occurs at the junctions because of the more localized dissipation and follows the square root dependence to the steady-state value. With knowledge of the diode thermal parameters, the relations given, and quantitative substantiation of the model proposed, the transient temperature response, together with its imposed component design limitations, are now obtainable.     

Thermal modeling and experimentation to determine maximum powercapability of SCR's and thyristors

This paper develops and explores a new thyristor thermal model that accounts for the temperature-dependent nature of the device material and construction. The model iteratively calculates temperature rise of a thyristor under arbitrary pulse conditions. The model is then correlated to an experiment that places a silicon-controlled rectifier (SCR) in a controlled test circuit at room and cryogenic temperatures. The knowledge gained from the thermal model and correlative experiment will allow the circuit designer to maximize thyristor capability in pulsed power applications. The increased power capability of operating a thyristor at cryogenic temperature is also examined     

Step-like shifts in the voltage—Current characteristics of power thyristors and their effect on transient thermal impedance determinations

In the course of determining the transient thermal impedance of power thyristors, step-like shifts are often observed in the temperature-sensitive voltage characteristic used to estimate the virtual junction temperature of the device. This correspondence describes investigations which have shown that these step-like shifts are in fact spontaneous voltage transients which are highly sensitive functions of current and temperature, and which cause small hysteresis effects in the voltage-current characteristics of the device. A simple method for adjusting the temperature sensitive voltage data to correct for such shifts is described; the method yields valid transient thermal impedance determinations in far less time than could be realized without the correction method.