As a graduate student, the best way to learn — and be challenged – is by tackling a real-world problem or research project. But it’s rare for a student to be pulled onto an assignment with a major company, working side-by-side with experts, taking on a start-to-finish challenge that makes a true difference – and let alone, receive a grant to support full-time study.
Yet that’s exactly what happened to Texas A&M University graduate student Parijat Mukherjee, who collaborated with Texas Instruments and his professor, Peng Li, to improve a software tool that saves engineers time and resources.
The necessary tool was a circuit stability checker, a software algorithm that allows designers to identify potential glitches such as noise, which can cause oscillations, in analog design. This prevents malfunctions in the manufacturing process. If instability is found, the cause must be identified and communicated to analog designers, who can fix the issue.
G. Peter Fang, a TI numerical simulation engineer, knew the prototype could be enhanced to be scalable — while benefiting a student — by partnering with Li, an associate professor in the Texas A&M University Department of Electrical and Computer Engineering. “This was a real problem that had real data that could help students learn along the way,” Fang says.
Li had analog and mixed-signal circuit research and design experience, and was excited to give one of his students the chance to work on an assignment from beginning to end. “When dealing with stability, the complexities are high,” Li says. He knew he “needed to have a bright student,” and found that in Mukherjee, who financed the final year of his masters with the TI research grant. “The timing was perfect,” says Mukherjee.
For 15 months, Mukherjee completed research, wrote equations, and tested numerous search methods to search for instability-causing problems, giving real-time feedback to designers during monthly calls. Over a two-week period, he transferred the tool into TI’s production flow for designers. “We consider the problem solved,” Fang says.
The project team published a joint paper on the topic at the 2011 IEEE/ACM Design Automation Conference (DAC), the leading electronic design automation conference worldwide. The paper earned a best paper award — recognition at the highest level for research in electronic design automation.
“We know that whatever we do is going to have a real impact,” says Mukherjee, who is now pursuing his Ph.D. “The collaboration helped us make something that was real and useful.”