EPC has launched its 11th report on GaN reliability with the strong claim that its transistors are more rugged than silicon devices. This comes from the Phase Eleven Reliability Report that covers field experience of 123 billion device hours over the last ten years.
The field test strategy used in the report relied upon tests forcing devices to fail under a variety of conditions. This failure testing approach to GaN reliability creates an understanding of the amount of margin between data sheet limits and products in application, and more importantly, identifies intrinsic failure mechanisms. This knowledge of the intrinsic failure mechanisms is used to determine the root cause of failures. Knowledge of the behaviour of a device over time, temperature, electrical or mechanical stress can provide users with an accurate representation of the safe operating life of a product over a more general set of operating conditions.
All of this leads the report to claim that EPC’s GaN transistors are more robust than silicon. Reliability is key for applications such as lidar for autonomous vehicles, LTE base stations, vehicle headlamps and satellites.
“eGaN devices have been in volume production for over ten years and have demonstrated very high reliability in both laboratory testing and high-volume customer applications,” said Dr. Alex Lidow, CEO and co-founder of EPC. “The release of EPC’s 11th reliability report represents the cumulative experience of millions of devices over a ten-year period and five generations of technology. These reliability tests have been undertaken to continue our understanding the behaviour of GaN devices over a wide range of stress conditions.
“The results of our reliability studies show that GaN is an extremely robust technology that continues to improve at a rapid pace. EPC is committed to subjecting GaN devices to rigid reliability standards and sharing the results with the power conversion industry,” he said.
The report showed that the key mechanisms impacting dynamic RDS(on) have been identified and are used to create more robust designs. The result