Intel awards Chiu grant to study data converter energy efficiency

2/11/2010 Megan Kelly, Coordinated Science Lab

ECE Assistant Professor [profile:chiuyun] would like to push data converter energy efficiency as close to its fundamental limit as possible. And this goal may become a reality, thanks to a competitive three-year grant that Intel recently awarded Chiu.

Written by Megan Kelly, Coordinated Science Lab

Yun Chiu
Yun Chiu

ECE Assistant Professor Yun Chiu would like to push data converter energy efficiency as close to its fundamental limit as possible. And this goal may become a reality, thanks to a competitive three-year grant totaling $478,850 that Intel recently awarded Chiu.

The Intel "High Performance CMOS ADC with Unprecedented Energy Efficiency" funding will support Chiu's research in the data converter area, where he and his team of students are attempting to make analog-to-digital conversion more energy efficient. Chiu is a researcher in the Coordinated Science Lab.

According to the Intel grant proposal, during analog-to-digital data conversion "physical quantities such as audio, video, radio-frequency signals, sensory outputs, etc. are converted back and forth between their original analog forms and computer-compatible, ready-to-be-processed and -stored digital bits."

The performance of analog-digital electronic interface affects the signal qualities people perceive and experience, no matter how accurately they are processed internally within a digital computer.

"It is fair to state that data conversion interface is the gatekeeper of the ever-growing digital world," the proposal continues.

Currently, Chiu and his students make converters that consume about ten femtojoules (fJ), or 10-15 joules, of energy for each conversion step. His goal is to reduce the consumption to one femtojoule per conversion step after three years.

"Our prototype chip has achieved the best energy efficiency ever reported," Chiu said. "But if we can reduce it to one femtojoule, we will be getting close to the fundamental (thermal) limit that can ever be achieved. It's pretty incredible."

It's especially incredible if one realizes that just 10 years ago, 1,000 fJ were required per conversion step.

Chiu believes Intel awarded him and his team the grant because of their impressive track record at the annual International Solid-State Circuits Conference (ISSCC), also known as the "Olympics of Circuit Design". His research has been published at both the 2009 and 2010 conferences in the data-conversion technology sessions. Only a handful of papers are published in this area each year.

"It's very difficult to be published at ISSCC, due to the intense international competition that involves both academia and the semiconductor industry," he said. "A lot of people are working in the same area. If your paper gets in one year, that's pretty significant, and we got selected two years in a row."

Chiu believes that from Intel's perspective, his research may be a major investment for the microprocessor giant and that they may be betting their converter technology on his research for the next few years.

"It's sort of a competition [among the peer institutions and research groups] to reach our goal," Chiu said. "They see a major contributor to the semiconductor industry in Illinois people."

In addition to its monetary benefits, the Intel grant includes other allowances.

"Intel will provide additional perks, including internship positions for students, hardware and software for research, and experimental support," Chiu said. "Also, interns on site [at Intel] will have an opportunity to use the costly Intel processes for free."

In addition to Chiu, University of Illinois engineering students Wenbo Liu, Charles Lee, Dae Hyun Kwon, Richard Tseng, Pingli Huang, Brian Elies, Bei Peng, former students Paritosh Bhoraskar, Victor Lu, and former visiting students Hao Li (USTC) and Peiyuan Wan (BJUT) all contributed to the research.


Share this story

This story was published February 11, 2010.