Diagnosing Thermal-Interface Aging of Power Devices Using Self-Sensing

In this article, we propose a unique and simple method that diagnoses multiple aging effects in power electronic devices minimally invasively and without the necessity of expensive sensors. Different degradation modes, e.g., fatigue of solder and thermal-interface layers, influence the phase of the thermal impedance frequency response function <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${\boldsymbol{\angle} \underline{\boldsymbol Z}_{\mathbf{th}}(j\omega)}$</tex-math></inline-formula> at specific bandwidths. Thus, tracking changes of the thermal impedance's phase, i.e., the phase shift between periodic device-loss excitation at specific frequencies and the resulting junction temperature response allows identifying these degradation modes. This is exploited by the proposed method for degradation diagnosis, which takes advantage of the temperature dependency of the drain-source voltage. The method excites periodic conduction losses at selected frequencies via small-signal manipulation of the gate-source voltage and measures the phase delay between gate-source and drain-source voltage. The measurable phase delay results partially from the dynamic response of the thermal impedance, because the phase-delayed junction temperature impacts the <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">on</small>-state resistance and therefore the drain-source voltage. Consequently, changes of the measurable phase delay allow identifying changes of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${\boldsymbol{\angle} \underline{\boldsymbol Z}_{\mathbf{th}}(j\omega)}$</tex-math></inline-formula> and thus diagnosing the above mentioned degradation modes. This article features a detailed analysis of the proposed method using a general sensitivity analysis. A detailed analysis shows that the discussed method is applicable for silicon (Si) <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">mosfet</small>s, gallium nitride (GaN) high-electron-mobility transistor (HEMTs) as well as silicon carbide (SiC) <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">mosfet</small>s. Experiments with SiC <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">mosfet</small>s, being the most emerging technology in industry, demonstrate that the implemented method can effectively diagnose changes of the thermal path between the device and the heat sink that result from different degradation modes.