Harnessing the power of GaN and machine learning – News – Compound Semiconductor

Military installations, especially on ships and aircraft, require robust power electronics systems to operate radar and other equipment, but there is limited space onboard. Researchers from the University of Houston will use a $2.5 million grant from the US Department of Defense to develop compact electronic power systems to address the issue.

Harish Krishnamoorthy, assistant professor of electrical and computer engineering and principal investigator for the project, said he will focus on developing power converters using GaN (GaN) devices, capable of quickly storing and discharging energy to operate the radar systems.

He is working with co-PI Kaushik Rajashekara, professor of electrical and computer engineering, and Tagore Technology, a semiconductor company based in Arlington Heights, Ill. The work has potential commercial applications, in addition to military use, he said.

Currently, radar systems require large capacitors, which store energy and provide bursts of power to operate the systems. The electrolytic capacitors also have relatively short lifespans, Krishnamoorthy said.

GaN devices can be turned on and off far more quickly - over ten times as quickly as silicon devices. The resulting higher operating frequency allows passive components in the circuit - including capacitors and inductors - to be designed at much smaller dimensions.

But there are still drawbacks to GaN devices. Noise - electromagnetic interference, or EMI - can affect the precision of radar systems, since the devices work at such high speeds. Part of Krishnamoorthy's project involves designing a system where converters can contain the noise, allowing the radar system to operate unimpeded.

He also will use machine learning to predict the lifespan of GaN devices, as well as of circuits employing these devices. The use of GaN technology in power applications is relatively new, and assessing how long they will continue to operate in a circuit remains a challenge.

"We don't know how long these GaN devices will last in practical applications, because they've only been used for a few years," Krishnamoorthy said. "That's a concern for industry."

The health and well-being of AngelTech speakers, partners, employees and the overall community is our top priority. Due to the growing concern around the coronavirus (COVID-19), and in alignment with the best practices laid out by the CDC, WHO and other relevant entities, AngelTech decided to postpone the live Brussels event to 16th - 18th November 2020.

In the interim, we believe it is still important to connect the community and we want to do this via an online summit, taking place live on Tuesday May 19th at 12:00 GMT and content available for 12 months on demand. This will not replace the live event (we believe live face to face interaction, learning and networking can never be fully replaced by a virtual summit), it will supplement the event, add value for key players and bring the community together digitally.

The event will involve 4 breakout sessions for CS International, PIC International, Sensors International and PIC Pilot Lines.

Key elements of the online summit:

Register to attend

Read the original here:
Harnessing the power of GaN and machine learning - News - Compound Semiconductor