Interview with Bill Messner

Featured Engineer

Bill Messner

Bill Messner - Professor, Carnegie Mellon University

image of Bill in 1966 working on an engine.

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

I had an interest in things mechanical from an early age. I really didn’t start working with electronics until college. There I built an optical scanner using scavenged parts from a computer paper tape reader.

What are your favorite hardware tools that you use?

Good old fashioned multimeter and oscilloscope, when I get into building stuff. Those and pencil and paper.

What are your favorite software tools that you use?

MATLAB. It’s utility for control systems analysis and synthesis is unequaled.

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

Back in the days of the 386, I had programmed a robot controller to perform a manipulation task once. Every once in a while it would repeat the task over and over . It turns out I had only allotted 6 characters for the file name in which to store the experimental data. When I entered file name was longer than 6 characters, it overwrote part of the program memory, causing the program to loop. It took a while to figure out that one!

What is on your bookshelf?

“A History of Mathematics” by Carl Boyer, and “Guesstimation: Solving the World’s Problems on a Cocktail Napkin” by Lawrence Weinstein and John Adam, “Night Comes to the Cretaceous” by James Powell, and “Guns, Germs, and Steel” by Jared Diamond.

Do you have any tricks up your sleeve? (special way to analyze circuits, special process you use to make something, etc.)

I love tricks for doing back of the envelope calculations without an calculator. For example, I have a technique to simply derive the factor equivalents of integer decibel values between 1 and 20 without a calculator. Since decibels are a logarithmic scale they are useful for quickly approximating powers, roots, and products.

I’m also a big proponent of classical loop shaping for controller design. I’ve developed a number of new compensator structures (e.g. the complex lead compensator) that provide new capabilities for precision loop shaping, and new visualization tools that are modifications of the traditional Bode plot that provide closed-loop performance information along with the open-loop frequency response.

What has been your favorite project?

Working on the 2004 and the 2005 DARPA Grand Challenge. (See my non-speaking role in Nova’s “The Great Robot Race.”) My primary contributions were some shock isolation technology—the March issue of Scientific American shows my quad-tripod shock isolation design—and some math for the path planning.

The idea of the design was simply to support the 550 kg electronics box like a sedan chair. The provided good symmetric support. Each support was a tripod of coil springs over shocks like one finds in a car or on a motorcycle. The tripods were arranged in equilateral tetrahedron, which provided a good balance of vertical and lateral stiffness.

The excitement and intensity of that project were absolutely exhilarating.

Do you have any note-worthy engineering experiences? (blowing up things, getting shocked, etc.)

About 17 year ago my group had a project on control high speed tape transport for advanced tape drives. Due to an in accuracy in the dynamic model for the winding tape, our initial controller was unstable. Since the reels were turning (over 20 miles per hour) at such high speed, the machine quickly threw dozens of feet of tape around the room—a big mess!

What are you currently working on?

Right now I am working on a few different projects. One of them is a low tech methods for improving the efficiency of apple harvest, specifically moving fruit from trees to the apple bin. This system is a totally passive mechanical system. Apples are particularly susceptible to bruising and so it is a great design problem—a huge design space but with lots of cost and durability constraints. It’s been lots of fun getting to know growers and learning about the clever approaches they have for lots of tricky problems of planting, pruning, cultivation, and harvest. Also, everyone who hears the about the problem “gets it,” and people immediately have lots of ideas, many of which are quite good.

I also have a very interesting collaboration with one of my Carnegie Mellon colleagues and a colleague in biology at the University of Pittsburgh on the dynamics of embryonic development in the frog Xenopus laevis. My contribution has been the control of the flow in the microfluidic network to control the local chemical environments of excised samples of embryonic tissue placed in the network.

We developed a novel long-term, high-speed pressure control mechanism that employs modulation of mechanically coupled fluidic resistance and squeeze pump fluidic between a fluid reservoir and the microfluidic network. Using a DC motor to actuate a 4-bar linkage that constricts (or releases) a narrow tube to modulate fluidic resistance while release (squeezing) a small variable reservoirto modulate fluidic volume in network capacitance. Compared to using a variable resistance alone, our flow control is 500 times faster, with time constants of just 40 milliseconds. We use a PIC microprocessor and MEMS pressure sensor to implement the feedback control of the pressure. Commands are issued from a PC running MATLAB.

Since our reservoir can be of arbitrary size and is separate from the control mechanism, we can run experiments that last hours or days, much longer than using standard syringe pumps that have limited volume and bandwidth.

When working on a real world application, do you model the system you are designing?

We do model the systems. For mechanical systems, we may use Newton’s Laws or Lagrange’s method. For the fluidic systems, we develop electrical analogies—very handy to think of pumps a current sources, pressurized reservoirs as batteries, and fluidic resistances as resistors!

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

Well, my business is education. I predict a move away from textbooks. Students no long use them as references, and instead rely (appropriately) on Wikipedia and other resources on the web. However, I still don’t see what will take their place. I’m open to suggestions.

What challenges do you foresee in our industry?

Global competition for resources, e.g. rare earth elements and energy. It’s worrisome that the electronics industry and the world are so reliant on sources of some critical raw materials from countries that have already indicated a willingness to restrict supply. It also appears that many predictions that the world has reached or will soon reach peak production of certain raw materials, such as oil, are more correct than not. (See the 25 March 2011 issue of Science for a news focus feature on the peak production of oil.) The industry is going to have work harder and smarter to produce more with less.