Hot-headed over cell phones

[Today's guest blogger, Mr. GK of San Diego, CA, works for (cellular telephone company). When he first started there, he had the job of measuring the effect that cellular telephones would have on the brain.]

So everyone's walking around with these cellular phones stuck to their heads and some people are wondering about possible effects. One obvious effect is that people on the phone drive slower and swerve a lot. But what about the biological effects of placing a transmitter against the head? Beats me what the effects are, but I can tell how we to measure the power entering the brain and try to design around it.

First off, a little basic intro into the world of cellular phone antennas. Anything that can carry current on its surface is an antenna. [Antennae are the rods that stick out of a bug's head, and some antennas resemble them, but not all.] By the laws of physics, the currents on the surface will create radiating waves, and radiating waves striking the surface will create currents. Of course, the amount of energy crossing over could be very low, making a bad antenna, or close to all the energy could be converted, making a good antenna. Every mobile phone has an antenna, even if you can't see it. Actually, the phone itself makes up a large part of the antenna.

In a typical situation, there is a port or line near the top corner of the phone circuit board to which attaches a length of straight line (a whip antenna) or a metal spiral (a helix antenna). Frequently you'll get both, with the whip retractable and the helix hidden in a plastic stub at the base. So the current goes on the antenna part, right? Nope. When broadcasting, the current comes out of the port and spills all over the place since there's nothing to stop it. Some goes on the antenna, and some goes on the circuit board. So actually the whole phone is the antenna. A dipole antenna has two symmetric metal rods with a gap between them where the source goes; here, the whip or helix is like the top part of the dipole, and the phone body is like the bottom part. What about phones that don't seem to have an antenna? In those cases the antenna is inside the casing, usually as a wire or a little conducting plate along the top edge of the phone. Well gee, you may be wondering, why not always use nice, compact, invisible internal antennas? Much like Mister Ed being a horse, it's because of money, of course. Not the money involved in making the internal antennas, which is not that much, but the money in building more base stations so the phones will be in a higher-signal environment. It's when you're far from a base station and the signal is weak that you need a big antenna. In most metropolitan areas the signal is strong enough that you don't have to extend the whip antenna. There's also a psychological factor. Big antennas that aren't needed are sometimes called "placebo antennas".

Designing an antenna usually doesn't involve much since the phone is so small and the price is very constrained. There's just not much you can do, so as a cell-phone maker, you'll go to an antenna manufacturer and order an antenna from the catalog. The only real freedom in the design is in the matching circuit. That's a small collection of inductors and capacitors placed at the antenna port to try to maximize power transfer. There's a device called a network analyzer that sends a signal into the port and measures the reflection. So you sit there with the phone and try different positions, including holding the phone and putting it up to your head while measuring the reflected energy. Hopefully, the matching network will still work somewhat well in the different positions. Then come the radiation pattern tests. We take the phone and put it on a rotating stand in an antenna chamber. That's like a quiet room but the echoes being absorbed are radio waves. Imagine a dark little room with big, blue foam cones covering every surface. On one end is the rotating podium with the phone, and at the other end is a horn antenna. The network analyzer is attached to the phone and the horn antenna, but now it's measuring the power transmitted between them, rather than the reflection at one port. The podium rotates the antenna and we get a picture showing how much energy goes in which direction. Those patterns are always measured, and they're always a half-step above worthless. That's because the pattern changes completely when held in the hand up to the head.

There are fake heads that can help with the measurements. These are plastic heads with mannequin-like general features and filled with brain-imitating goo. The phone is strapped to the fake head and the antenna pattern measured. This is a better indication of how the phone radiates, although the best test is to have a person sit in the chamber and hold the phone. For that we take out the podium and place a chair on the rotating platform. The subject sits there holding the phone and gets rotated. I went to a conference presentation once where a university presenter mentioned that they couldn't get permission to do this kind of test since it involved human subjects, even though the power involved is actually less than when the phone is in operation.

That's the antenna design process to get it to radiate the most power; now we get to the opposite process, where we try to minimize power going into the head.

The important parameter when considering radio waves entering body tissue is called SAR. That stands for specific absorption rate, and is indicated in W/kg or mW/g [Watts per kilogram or milliWatts per gram]. There are various SAR limits set by the FDA and some agencies in other countries. The limits depend on frequency, exposure time and other factors, but for the frequencies we're interested in, the limit is 0.2 mW/g for the whole body. Since that's averaged over the whole body, it's actually not difficult to meet this limit. Perhaps you've got an antenna irradiating your left arm, but the rest of your body has no radiation and the average could be minuscule. So we don't worry about this limit. The rule that sends us to the measurement labs is the local SAR limit for brain tissue. That's 1.6 mW/g, and we're only allowed to average over a single gram cube. Some other countries follow the same limit, but for Europe and Japan it's a much easier to meet 2 mW/g averaged over a 10 gram cube. So how do we measure the brain tissue absorbing energy?

Imagine a plastic sink shaped like half of a human head. It's like someone cut the mannequin head in half, laid it on its side, and put an edge on it to avoid spills. Into that we put our brain-like goo. That goo is home-made, and actually quite simple. It's just water, salt and sugar. We start with water in a bucket, add some salt and sugar and then take a reading with a device to measure the conductivity and permittivity of the liquid. That device is a network analyzer again! This time it has an attachment that can be stuck into the liquid. It again puts some energy out into the attachment and measures the reflection. Roughly speaking, we add salt to increase the conductivity and add sugar to increase the permittivity. (What's permittivity? That's related to optical index, as in refraction, although this isn't in the optic range. It's like a measure of how electrically dense the liquid is.) If you go over the target values, add more water.

Once the fake zombie-food is ready, we pour it into the hollow head. Then comes the stirring. The ingredients will settle towards the bottom and throw off the carefully calibrated density. So about every hour the mixture has to be stirred with a long stick. The stirring has to be very slow; if bubbles develop, we'll have to wait for them to dissipate. A few minutes of slow stirring should do it. Now we're ready to take a measurement.

The phone is attached underneath the fake head around the ear and turned on. We program the phone to just transmit at maximum power. The measurement is done by a big robot arm set up over the head. The robot arm holds a long straight rod, and at the end of the rod are a couple of little dipole antennas. That's the near-field probe. The robot arm puts the probe into the head and starts measuring the field strength. The operator can kick back and relax as the robot arm sweeps the probe in a 3-D pattern over the head volume. We then look at the results. Did the SAR exceed the limit at any point? If not, we're set, right? Once again, the answer is a big opposite-of-yes. While the FDA sets the SAR limits, they don't say much about how the measurements should be made. How should the phone be positioned against the fake head surface? Can we include a fake hand? How about a fake ear? Is the whole concept of using a fake head with fake brain-goo valid? Who knows, but it's all we have to go on. And what happens if we measure areas where the limit is exceeded? That's when the witchcraft really begins.

To get the SAR down we can do things like metalize the inside of the plastic casing. Try different patterns; maybe one will work. There are metal screens available on the market that you can put over the earpiece that will supposedly cut down on radiation into the head. They must be meant to be used in the bizzaro universe, since by those pesky laws of physics they won't do anything in our universe. Unless the metal screen is attached to a ground, it will be invisible to static fields, radio waves, and any electromagnetic wave up to nearly optical frequencies. The metal patterns we put on the inside of the casing have to be attached to the circuit board, and they work by drawing the current away from the ear area rather than by shielding. Or maybe there is some shielding going on. It's all trial and error.

In the usual whip or helix antenna configuration, the electric current hotspot is around the ear. That's also the area where the SAR will be the highest. An antenna at the bottom of the phone would bring the hot-spot down toward the chin, but there isn't enough room in the phone for it. Probably the only way of reducing energy absorption into the brain is through distance, like with those ear-piece cords. Although on the down side, when you use those you look like you're walking around talking to yourself.

The issue of long-term low-level exposure is another mysterious area. SAR deals with heating of tissue. As long as you're at the SAR limit, the tissue will not heat up (there's actually a safety factor of about an order of magnitude, so you could go over the SAR limit and still not experience heating). Bad effects from radio waves caused below the heating level have so far not been found. A typical experiment involves putting a dipole against a mouse's head for a long time and then dissecting the mouse. That hasn't turned up anything. There has been the suggestion that long-term exposure could cause a benign (non-cancerous) tumor to develop around the ear. It's hard to talk about long-term exposure since cell-phones have been developing and power levels are dropping. An old analog phone is like a portable FM radio station, pumping away almost continuously. And 10 years ago it could have been trying to reach a distant base station. Digital phones are able to operate at lower power (analog maximum is 600 mW, digital maximum is 200 mW). They also switch on and off to save power, so across any second it may only be on half or a quarter of the time, further lowering power levels. Also, there are base stations everywhere now, so the phone doesn't have to raise its power output much.

Does this mean that there is now no danger? Don't know. So I can only offer these words of advice: Trust us.