Electronic Design

Millimeter Waves Beckon

We're not out of spectrum yet

What do you do when you run out of radio spectrum? Like real estate land, they are not making any more of it. What's there is there. The solution over the past decades has been to push applications to higher and higher frequencies. This in addition to tightening regulations and adopting advanced wireless technologies to make more efficient use of what pitifully small amount of spectrum we have left. The number of wireless applications continues to expand, so we are finally getting a glimpse of the technology and components needed to make use of the higher frequencies.

Beyond the current near saturation of the lower microwave bands lie higher frequency microwaves we refer to as millimeter (mm) waves. Use has been limited for years for lack of components that operate in this range. But, thanks to semiconductor technology, usage has grown significantly over the years. And now, designers are finally taking a hard look at how to use these previously unoccupied bands.

What is a Millimeter Wave?

The name itself accurately describes this type of wireless. The wavelength of the signals is only a few millimeters (mm) or less. But since most of us think in terms of frequency, rather than wavelength, the mm wave range is 30 to 300 GHz, according to an international agreement struck in 1947. Traditionally this range is referred to as extremely high frequencies (EFH). Some people include frequencies just below 30 GHz to frequencies just above 300 GHz in mm wave discussions. In any case, the wavelengths are really short. Since wavelength in meters is 300/fMHz, then the wavelength of a 30 GHz signal is 300/30,000 = 1/100 or .01 meter. That is one centimeter or 10 mm. 300 GHz is 1 mm. A half-wave dipole at 30 GHz is only 5 mm long.

To further put this into perspective, the mm wave band lies between the upper end of what we normally call the microwave region (30 GHz) and the lower end of the optical band. The optical band extends from about one micron, or 10-6 meter (one millionth of a meter), up to past 0.4 x 10-6. The optical band includes infrared (IR), visible light and ultraviolet (UV). Both IR and visible light are used in wireless (mostly IR), but UV is not. The visible light range is roughly 0.7 x 10-6 (red) to 0.4 x 10-6 (violet) or 400 to 700 nanometers (nm) where a nm is 10-9 meter. The most widely used IR wavelengths are in the 700 to 1600 nm range.

The upper end of the mm wave band is 300 GHz, but that is 1 mm or 1 x 103 nm. So if you think about it, there is a gap between the low end of the IR band and the upper end of the mm wave band. It is sort of a no man's land or dead zone. I am not sure what it is called...quasi-optical, ultra-low optical, extreme mm waves or whatever. No doubt there is some research being conducted there, but to date no practical components or systems exist. Ah, yes. Room for growth.

Finally, the microwave bands have been divided into bands and given a letter designation (see the table). The mm waves are included.

More Room At the Top: Benefits and Downsides

Obviously the higher in frequency you go, the greater the available bandwidth for a given application. You might think in terms of a 6 MHz TV channel. In the 1 GHz range, that one channel represents 0.6 %. At 300 GHz the percentage is only 0.002%. So there is lots of extra space for lots of services, subscribers, etc—if you can get the allocation from the FCC.

In addition to more space for your apps, you can transmit data at higher rates. With gigabit data rates today, you need lots of bandwidth and a higher carrier frequency. It takes a carrier in the mm wave range and a channel bandwidth of 2 GHz or so to deliver a 1 Gbit/s data stream reliably.

If you are using the signal for radar, a higher frequency gives you much greater resolution. This is because shorter wavelengths can help you see smaller details in the target.

One unexpected benefit of mm waves is the very short antennas needed. You can readily make patch arrays with high gain and steerability. Most apps probably still use a parabolic dish. Even horns are practical at these frequencies. A small dish that measures about one foot in diameter can produce huge gains at these frequencies. High gain compensates for low power and helps extend the range.

But, of course, the down side of mm waves is the fact that signals travel a distance inversely proportional to the square of the wavelength. As a result, the potential range is much shorter—even than the lower microwave bands we have become so familiar with. If you use enough power, you can go farther. A good example are the Ku band satellites used in direct broadcast satellite (DBS) TV from EchoStar, DirectTV, and a few others. The distance out to a geosynchronous satellite is about 22,300 miles. That is about as far as anyone needs to transmit or receive, but it takes high power and very high antenna gains at both ends of the link—not to mention the very low noise requirements at the receiver. In any case, we do it routinely today.

One of the main downsides of mm waves is oxygen absorption. The size of an oxygen molecule is about the same size as one wavelength at 60 GHz. This means that the air absorbs the radio signal— introducing attenuation beyond the normal radio signal loss over distance. The loss does not prevent the use of 60 GHz radio, but it does reduce its range because of the additional 15 dB or so attenuation per kilometer. Some say that this is okay, because the signal is made more secure (since no one can actually intercept the signal at any great distance). Of course, given that there is no oxygen in outer space, oxygen absorption for inter-satellite communications is not a problem. And above and below the 60 GHz band, the absorption attenuation is no longer much of a factor.

A really key thing to think about is the components available to actually implement mm waves. It used to be that there were no tubes, transistors, or diodes that would work in that range. But today, with sub-micron semiconductors and special materials like SiGe, GaAs, and InP (and a few other compounds— including some with carbon), we can make bipolars and FETs that easily reach into that range. Some researchers even report transistor fTs out to 300 GHz. Not bad. So, yes, having suitable components is a critical need, but it has been filled slowly over the years. It used to be that the leading edge technology was physical waveguides and Gunn diodes in cavities for oscillators and varactor multipliers. These are still widely used, but are slowly giving way to MMIC (monolithic microwave integrated circuits) amplifiers, mixers, oscillators and other circuits in IC form, some in standard sub-130 nm silicon. Put these on a ceramic or plastic substrate and you can make a mm wave radio.

Finally, cost may be an issue. Both silicon and SiGe designs are still very expensive, but prices will fall as volume for these products builds. Millimeter wave gear remains pricey simply because of the exotic components and low volume. But if you really need the benefits that mm wave bring, the additional cost is generally not a knockout factor. Example: military.

Millimeter Wave Applications

Probably the main use of mm waves is super fast backhaul. It is used to wirelessy link two buildings up to about a mile away. Speeds to 1 Gbit/s are possible. A good example of a typical product is the Proxim GigaLink series. It operates at 60 GHz and can deliver 1.25 Gbit/s Ethernet up to about one half mile with only 8 mW of power. It uses a very high gain parabolic dish with a beamwidth of one degree.

Another application is military radar. To get ever more resolution and detail, designers have gradually increased the frequency of operation in some aircraft radars to mm wave frequencies. Not only can these radars see smaller objects (like missiles); they also can make out fine details in other aircraft and ships scanned.

Another radar use is automatic braking in automobiles. Some high-end cars from Lexus and Mercedes Benz have a 77-GHz radar behind the grill or bumper to detect the range in distance to any vehicle in front of it. It then automatically adjusts the brakes if the distance gets too short.

A newer application of mm waves is weapons detection. Scanners using 75, 94, or 140 GHz can detect guns or other weapons on a person at a short distance. Not bad. Wall penetrating radars, which allow soldiers or police officers to look behind doors and walls before taking action, are also available. Ground penetrating radars also let engineers inspect roads and bridges for flaws in construction.

Inter-satellite communications is another mm wave application. MM wave modems are very small and light, making them perfect for satellites. Communications between satellites in arrays or larger systems over short distances is commonly handled via mm wave radios.

One unexpected application is using mm waves in new models of radio telescopes to explore the universe. A special 12-meter parabolic dish has been constructed with a supercooled receiver front-end to observe distant constellations and to analyze their gas content. The National Radio Astronomy Observatory (NRAO) operates four of these radio telescopes in Arizona, New Mexico and West Virgina for research and education.

The table linked linked heresummarizes the more common mm wave applications and their frequency assignments.

As semiconductor devices get smaller and faster and newer materials emerge, mm waves become cheaper and easier to use. It is still a rather exotic technology, but it’s becoming more practical every day. Look for more of it in the future.

My special thanks to Dana Wheeler, Senior VP and general manager of Terabeam-HXI, and his associates Earle Stewart and Dave Russell, who provided a substantial amount of information for this article. Terabeam is the parent organization of Proxim Wireless a leading microwave equipment company.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.