How To More Accurately Predict The Field Reliability Of Automotive Power Electronics

Extrapolating test data from the lab to the field doesn’t work because thermal models can’t predict junction temperatures with enough accuracy.


If your company makes planes, trains, automobiles, medical devices, computers, and communication systems, or you are a large electronic device supplier, the reliability of your products in the field is crucial to your business success. The growing market for electric and hybrid vehicles is increasing the pressure on life-time performance of the devices that power them. Estimating the actual field reliability before the product ships is difficult for two reasons. First, test data on the reliability of the individual parts is limited, and second, the experiments needed to subject a sufficient number of parts to accelerated testing are time-consuming and costly. This situation has hampered the introduction of new electronic devices, power units, and PCBs, and is slowing the adoption of new technologies such as wide-band gap devices, for example, silicon carbide (SiC).

On top of this, we have discovered an inability to extrapolate lab-based test data to the field because thermal models are unable to predict junction temperature rise during operation across a drive cycle with good enough accuracy. In this paper, we take a closer look at these problems and how two recent innovations from Mentor Graphics can help solve them.

To read more, click here.