Interview with Richard Mellitz

Featured Engineer

Richard Mellitz

Richard Mellitz - Principal Engineer for Intel Corporation

image taken in an Agave field in Guadalajara while I was there giving a speech on a signal integrity paper.

How did you get into electronics/ engineering and when did you start?

I started working on the Apollo Navigation systems and U2 spy camera analysis as a Co-op for Perkin Elmer.

What are your favorite hardware tools that you use?

It depends on what I’m doing of course. For physical channel analysis I tend to favor a VNA although TDR has its place.

What are your favorite software tools that you use?

Software can be a real black hole. I’ve used a wide variety of software. Right now I mostly use MatLab.

What is the hardest/trickiest bug you have ever fixed?

Bug… I would say problem. One problem comes to mind is when we were testing embedded resistance layers and noticed drop out. When viewed in 3D rendition the data outlined a hand print. It was obvious from what the rendition showed that the process handling was the problem.

What is on your bookshelf?

EM, Communication, and Signal Integrity books.

To be a good SI engineer what fundamentals or theory do you feel is most important to master and why?

A broad understanding of electrical engineering and physics. SI really a conglomerate field. In 2003 we were just introducing high speed differential signaling. Now for 100Gbs interface a whole plethora of topics are comprehended under the SI umbrella. A few might surprise some folks. Examples are statistical communication theory, statistical lifetime processes, and design of experiments, and DSP.

Regarding SI models, do you find extracting models from CAD drawings useful or do you use measured data to create models?

This is the classic answer “it depends”. One thing folks need to come to grip with is there is no universal model. The issue is we are talking about a problem that ranges from sub microns to meters.

Do you have any tricks up your sleeve?

Think outside the box.

If you were to pick one issue, what item would you have layout designers or engineers pay attention to that they might be overlooking?

Engineer and Layout designers have completely different outlooks. The attention is really the communication between the two.

What has been your favorite project?

Setting up a Graduate Signal Integrity Program at the University of South Carolina. At the time Intel needed a significant number of SI engineers and was having a difficult time hiring them. So Hao Li and Brian Knotts talked me into creating a course based on Steve Hall’s, Howard Heck’s, and Maynard Falconer’s work at Intel on training new hires for the SI role. I agreed to teach the first year and hundreds of slides later, the program has spawned a number of SI PHD’s and Masters. I believe Intel has hired over 25 out of that program.

The first slide set from that program can be found here.

The program was subsequently offered by Dr. Paul Huray and Dr Yinchao Chen. As result of the program fundamental research on practical transmission line theory was refine and publish by Dr. Huray et al. In addition from this involvement Dr Huray wrote a SI text book. A bit about the program is included in the preface of the book.

Do you have any note-worthy engineering experiences?

In every generation we find phenomena that we should have known about in hindsight. CIRC 2000 it was return path. Lately via coupling. This is reflected in scale order range that SI addresses. i.e. nano meters to meters.

What are you currently working on?

High speed differential IO SAS-3 12Gb/s, 100 GB/s Ethernet CU, and others.

What direction do you see your business heading in the next few years?

Better understanding of how design affects quality. Example not 1 or 0 but likelihood of 1 or 0. i.e. BER

What challenges do you foresee in our industry?

Speed/power walls. SSD.

Visit Richard's EEWeb Profile

Previous Next

Previous Spotlights

Featured Products

PCB Tools

Passive Tools

RF Tools

Math

Extras