Research
I am currently doing research on High-Speed, High-Resolution Digital-to-Analog Converters (DACs). DACs are of interest to you too, whether you know it or not, because they are used everywhere. Your cell phone has several of them (yes, even your "digital" phone). So does your car, your DVD player and maybe your coffee maker. Since you're reading this online, there are countless DACs in the communication chain that relayed this page from some server at Stanford to your computer. Not to mention the ones in your hard drive, sound card, video card, monitor, wireless mouse...
Why work on DACs?
DACs and their counterparts, Analog-to-Digital Converters (ADCs), are interesting creatures. They have the difficult job of interfacing between the digital world of microprocessors and the analog world of sounds, vibrations, and electromagnetic waves. To make things more interesting for DAC/ADC designers, we generally have to piggy-back onto a manufacturing process that's geared for making high-speed digital circuits. So, you may ask, does a process geared for better digital circuits also help you make better analog circuits? The answer is yes, but the myriad of tradeoffs involved makes it anything but obvious. In fact, over the years numerous doomsayers have thought that analog circuit growth (particularly in CMOS) would be halted by various fundamental limitations. Fortunately, some smart people keep finding ways to prove the naysayers wrong.But I thought digital was cooler than analog...
Thanks to clever marketing by wireless phone companies, many people erroneously belive that analog is on the way out and digital is the way of the future (this is in fact why some of my friends and family are concerned that my research may be a waste of time). But to people who actually know what they're talking about, it is nonsense to ask whether "digital" or "analog" is "better." ALL circuits are analog! All circuits are comprised of elements that must, at some point, be analyzed in a completely analog context. However, once the analog behavior of a circuit is understood, we can often then deal with it at a more abstract level where we only care whether each transistor is "on" or "off." If we can usefully employ this level of abstraction on a circuit, then we can call it "digital." Thanks to this simplification, engineers can further abstract the idea of "on" and "off" into adders and multipliers and memory and then on to busses and networks and IP addresses and packets. But underneath, it's still analog.(Of course underneath that, it's all physics! Analog designers don't want to think about physics any more than digital designers want to think about analog. On the other hand, a physicist could never design a digital microprocessor by considering each electron individually. Each of us, physicist, analog designer, digital designer, software programmer, etc., has to work at a level of abstraction that gives us enough detail to solve the problem, but not so much that we get hopelessly buried in it. Some people call this specialization, but it makes just as much sense to call it teamwork).
Anyway, some cirucits are "more analog" than others. Basically, circuits that cannot be simplified into "on" and "off" are what most people call analog. This is the class of circuits that usually interfaces any sort of media (such as the magnetic patterns on a hard drive or the electromagnetic waves in the air). In vast summary: analog circuits are ones that can send/recieve information directly to/from the environment, such as an antenna or hard disk. DACs/ADCs bridge the analog-digital gap by converting that information from/to a digital format. Finally, digital circuits, such as microprocessors and dsp's, can process the digital data using digital techniques.