Transphorm releases GaN FET reliability ratings
New ratings now segmented by power level
Transphorm has announced the latest reliability ratings for its GaN power FETs. Reliability is measured by Failures in Time (FIT), an analysis that considers the number of devices reported by customers to have failed in the field when used in applications.
To date, the company’s total product portfolio has achieved an average < 0.1 FIT rate based on more than 85 billion hours of field operation. According to the company, this rating stands as one of the industry’s best and only reported broad power spectrum reliability rating of any GaN power solution available today.
In 2019, Transphorm was the first GaN manufacturer to publish a complete validation data set backing its reliability claims. Since then, the company regularly shares its GaN reliability achievements to help potential customers make informed decisions when choosing semiconductor suppliers. Transphorm last reported its FIT rate to be < 0.3 in Q1 2022.
This year, Transphorm takes another step toward changing how customers assess GaN FET options. The company has taken its reliability data and segmented it into two categories: Low Power: GaN devices used in applications with power levels ≤ 500 W; and High Power: GaN devices used in applications > 500 W
When looking at device performance by power level type, Transphorm’s GaN FETs yield the following reliability ratings that are notably similar to those of Silicon-based power devices: Low Power: 0.06 FIT, and High Power: 0.19 FIT.
“Our high voltage GaN devices are designed into the broadest range of applications covering the widest power spectrum, from 45 W to 4 kW today with the potential to reach 10+ kW as GaN is adopted into new markets. This shows the immense versatility of our technology,” said Philip Zuk, SVP of Business Development and Marketing, Transphorm. “However, we realised that reporting just a singular reliability rating that lumps all application types together may not be as useful to customers. We felt it necessary to help them access more nuanced data that would apply to their specific design requirements. Hence, the breakdown between low and high power.”