|
| |
Phones of the Future
From: University of Michigan
| By:
Clark T.-C. Nguyen |
EDITOR'S INTRODUCTION |
All around us, machines that incorporate moving parts no bigger than the period at the 
end of this sentence are starting to make our lives easier, safer and more entertaining. Tiny sensors in cars trigger the inflation of air bags when they sense a sudden stop. Movies are projected on screen by reflecting off millions of minuscule mirrors. And in labs at the University of Michigan, Clark T.-C. Nguyen (right) and his colleagues are refining other kinds of microelectromechanical machines to improve the performance of wireless phones. In this interview, Nguyen describes the impact of this research on the future of communications. |
Fathom: How do wireless phones make use of these microelectromechanical machines, called MEMS for short? |
Clark Nguyen: One of the biggest issues with wireless phones is that you don't want 
someone else's conversation to interfere with yours, which can happen if you don't have enough filter selectivity. So you might be speaking into your phone, and suddenly you're hearing someone else's conversation, or you're just hearing noise from another channel on your phone. This is why you need extremely selective filtering, and this is where the MEMS technology comes in. |
Nguyen explains how resonating filters work.
|
Fathom: How do you think using MEMS could improve cell phones? |
Nguyen: Say you're trying to operate a wireless phone using GSM, which is one of the European standards for wireless communications. GSM occupies a certain frequency spectrum around 900 megahertz called a band. It's just one band out of many that divide up the total frequency spectrum--where each band is used for a specific purpose, such as radio, television or other wireless communications. |
To prevent interference from bands outside the GSM band, the first filter in your phone knocks out the other bands and leaves you with only GSM signals. Of the remaining GSM signals, your phone must still select the one (i.e., the channel) that corresponds to your conversation, so even more selective filtering is now needed. To do this, the signal frequency is first lowered from 900 megahertz to, say, 70 megahertz, and then more filters are used to further select the signal. So you start out at a very high frequency like 900 megahertz, and that's then translated down to a lower frequency of 70 megahertz, where filtering and signal processing can be done much more easily and much less expensively. So what we're doing in some of our projects here at the University of Michigan is making micromechanical signal processors to directly replace some of the filters and other signal processors that exist in current phones. |
Among the many direct benefits of these new microelectromechanical filters is miniaturization. Right now, if you took a look inside a wireless phone, you'd see that the filters and other passive components take up most of the phone's space. |
Miniaturizing the filters can get us closer to being able to get an entire phone on a single 
chip or maybe two chips, depending upon what other problems can be solved. In particular, saying "single chip" is pretty daring, because there are actually a few other problems in addition to miniaturizing the filters that must be solved before a true single-chip phone can be achieved. But even if all you could do is get the off-chip passive components down to a much smaller size, then this would be a huge step toward putting a wireless phone on a wristwatch or--and I'll be daring and say this--maybe even a ring on your finger. |
What we're doing with MEMS is replacing those components that so far could not be miniaturized with current technology. |
Fathom: Are there other benefits that MEMS have over the components commonly used today? |
Nguyen: Power savings is one of the main advantages of this technology. After all, what good is it to make the size of a phone as small as a watch if your battery still has to be as big as cell phone batteries are today? So you'd like your phone transceiver to be able to fit onto a wristwatch, but then also have a battery that's the size of a watch battery. The only way to do this is to lower the power consumption of the transceiver itself. |
It turns out that the micromechanical signal processors we're talking about can greatly lower power consumption in phones by taking advantage of a fundamental power-versus-selectivity tradeoff that wireless system designers encounter time and time again. In engineering terms, we often use quality factor, or Q, as a measure of the selectivity of a given resonant circuit or filter. The more selective the filter (in other words, the smaller the frequency bandwidth it can grab), the higher its Q. |
And it so happens that for wireless transceiver design, the better the frequency selectivity (or the higher the Q) of the filters, the lower the power consumption required in the circuit. This comes about because better frequency selectivity means that transistor circuits need process only a smaller bandwidth of signals. In general, when the bandwidth--or number of signals--to be processed is smaller, the total power of those signals is also smaller, so the transistor circuits can expend less energy to process them. |
It turns out that mechanically vibrating devices can generally achieve much better frequency selectivity (or much higher Q) than their electrical transistor or LC counterparts. Thus, a phone with more mechanical devices can be thought of as having a higher overall Q, making the job much easier for the transistors, so they can do their work while consuming less power. |
And in the end, mechanical components bring more than just size and power advantages. Specifically, the larger the number of more mechanical components used by a given phone, the better its overall performance. In other words, its sensitivity will be better, meaning that you'll see fewer problems with your phone dropping calls and fading and that sort of thing. |
And, if you're a system provider that's providing cellular phone service, maybe you'd have an easier time setting up your networks if you had the cheaper, better performing phones made possible by these mechanical signal processors. |
Nguyen explains why transistors have to work harder to select frequencies.
|
Fathom: You've gotten very high levels of selectivity in your work, haven't you? |
Nguyen: Yes. The level of Q that is available right now on-chip, using only conventional integrated circuit technology, ranges from five to nine. Compare that to the Q we get out of a micromechanical resonator, which is also an on-chip component. The Q's of polysilicon micromechanical resonators have been measured from 10,000 all the way up to 100,000. So there are many orders of magnitude in difference there. |
Fathom: What other uses for RF MEMS, besides phones, do you see in the future? |
Nguyen: Wireless networks of sensors could be used to monitor all kinds of environments. For example, think of a farmer who wants to use enough chemical fertilizer to make sure the vegetables are growing well on a huge farm. But the farmer doesn't want to use too many chemicals because then they can get into the water and cause problems in the environment. A wireless network of sensors would allow you to monitor the fields more exactly, so that you can make sure that you're putting in exactly the right amount of fertilizer that would be beneficial for the plants but not an excess that would harm the environment. |
But there's also the down side of this, too. We could be sitting in a room that would be covered with these tiny wireless monitors, and we may not know exactly what the monitors are doing. We may never have any privacy if this goes too far, so there are societal impact issues with this type of technology as well. |
Personally, I think the wireless technology is something that cultures will adjust to. We'll find ways to make things work out at the societal level. |
Nguyen: I think transistors are better at some things, and mechanical systems are better at other things. So I'm not talking about an eradication of transistors--that would be a truly wild idea. But I am talking about using both mechanical systems and transistors together to get the best performance you can out of a particular application. |
Fathom: Will tiny machines be susceptible to the same kinds of breakdowns that we see around us in bigger machines, such as cars? |
Nguyen: Because it's free to vibrate and because it's so tiny, people often think that a micromechanical device is fragile, and thus, can break very easily. They might think that it's going to be the component that is going to fail first. But this is not true in general. In the most common cases where nothing physically touches them (i.e., forces are applied remotely), these devices are much more resilient than even the macroscopic ones presently used. |
For instance, when NASA's mother ship orbits Mars, it fires in a device called a penetrator, which tries to get deep under the surface of Mars to look for signs of life or water. It then transmits its findings wirelessly back to the mother ship. However, a problem arises in that the penetrator hits the surface so hard that the inertial forces generated by the impact can break or change the behavior of some of the components within the penetrator. So NASA is looking for communication devices that are more resilient to high G forces--to huge impacts--and it turns out that the MEMS components, because they're so tiny, have so little mass that even large impacts do not generate large inertial forces on them. Because these small components generally experience much smaller inertial forces upon impact, they may actually be more resilient against high G forces. |
Nguyen discusses changes at NASA in recent years.
|
Fathom: What issues are left to solve to make MEMS technology more widespread? |
Nguyen: The biggest problem with MEMS right now is the packaging aspect of it. When you make something this small, it becomes more susceptible to the environment--to humidity, to temperature fluctuations, even to gas molecules getting onto the thing. Without proper environmental control (i.e., without proper protective packaging), these susceptibilities can make vibrating micromechanical resonator devices less stable in frequency than macroscopic vibrating components. This is why people in the MEMS industry are looking for inexpensive packaging solutions. |
If we are able to package these devices at a batch level instead of one at a time, then the cost is driven down and this becomes a much more marketable technology. People are looking at batch packaging, and it looks as if there are some techniques now that are getting closer to being manufacturable. |
|
| |
