Research

I am currently doing research on high-speed, high-resolution digital-to-analog converters (DACs). DACs are of interest to you because they are used everywhere. Your cell phone has several of them (yes, even 'digital' phones). So do cars, radios, beer openers that play college fight songs, and big-mouth bass that sing 'Take Me to the River.'  DACs and their counterparts, analog-to-digital converters (ADCs), are analogous to transmitters and receivers.  You use a DAC to send a signal from a microprocessor out into the world, and you use an ADC to bring real-world signals back inside the microprocessor.  Without DACs and ADCs, all the wonders of digital technology would be trapped inside tiny chips with no way in or out.  What would an iPhone be without ADCs to register when you are touching the screen, or DACs to drive speakers or headphones?  Or ADCs notice when you turn an iPhone on its side or shake it.  Most importantly DACs and ADCs allow iPhones to broadcast and receive data on the cellular network (or WLAN).  Without ADCs and DACs, digital electronics would be deaf, blind and dumb.  

Why work on DACs?

DACs and ADCs are fun to design because they involve so many crazy challenges.  Some are related to fundamental physics and put hard bounds on what you can ever possibly achieve.  Other challenges are related to manufacturing limitations.  Still others are related to physical constraints, or battery limitations or other concerns.  Because the constraints are always changing, DACs are always changing too, which keeps people like me busy.  As an example, DAC/ADC designers generally have to piggy-back onto a manufacturing process that's geared for making ever faster and faster digital circuits. So, you may ask, does a process geared for better digital circuits also help you make better analog circuits? The answer is (a mildly qualified) yes, but the myriad of tradeoffs involved makes it anything but obvious.  Figuring out how best to use the good (and overcome the bad) characteristics of new integrated circuit technologies is a big part of the fun.  

But I thought digital was cooler than analog...

Thanks to clever marketing by wireless phone companies, many people believe that digital is the way of the future and that analog will be relegated to the garages of strange hobbyists.  But to people who actually know what they're talking about, it is pretty much nonsense to ask whether 'digital' or 'analog' is 'better.' Partly this is because ALL circuits are analog. Even digital logic must, at some point, be analyzed in an analog way. Once the analog behavior is understood, we can sometimes deal with it at a more abstract level where we only care whether each transistor is, say, 'on' or 'off.' If we can usefully employ that level of abstraction, then we call the circuit 'digital.' In other words, there is a lot of analog design done on digital circuits.  But, the other reason analog isn't going away is that analog techniques are fundamentally required when data has to move over long distances (and 'long' is getting shorter every day).  Really, the right perspective on all this is to think about how to combine analog and digital techniques in order to obtain a very high performance system.  In fact, this is exactly what I do.  It's called mixed-signal design.