Ed is an EE, PE, and author in Poughkeepsie, New York. You can contact him at [email protected].
Any group of people working on a project will evolve unique terms and expressions that encapsulate complex concepts. Even though outsiders may find the jargon impenetrable, that's generally not the intent.
But sometimes it is. As serious radar development got underway in the mid-1930s, everyone realized that mentioning exact frequencies might reveal critical information to folks who, even if they weren't quite enemies, might become more than competitors. So engineers working with particular wavelengths described them with letters in the hopes of mystifying the outsiders.
This encryption was so successful that, more than half a century later, electrical engineering students still blow exam questions and drivers still puzzle over the differences between X-band and Ka-band traffic radars.
The always-on, always-connected, network-ready embedded device model makes RF a hot topic for systems designers who must now figure out how to shoehorn a radio and antenna into their hardware budget. Let's take a look at some of the jargon and physics that lies beyond the firmware and between the antennas.
First Contact
In radio's early years, around 1900, transmitters operated in what we'd now call spread-spectrum mode, emitting "DC to daylight" RF energy that could be detected by untuned receivers. Only one transmitter could be active at any one time within a geographic region defined by the transmitter's power and the receiver's sensitivity, which became a major inconvenience as both ends of the link improved and more organizations got on the air.
The notion of confining a transmitter-receiver link to a specific frequency originated shortly after 1900. Common knowledge held that long-distance communication required long-wavelength radio energy, with wavelengths on the order of tens of miles.
Steam-driven electrical alternators were the only source of AC power in those days, typically running near 60 Hz (Hz stands for Hertz, the guy who discovered RF. In those days, 60 Hz was known as "60 cycles per second"). By 1918, General Electric's Alexanderson designed monster alternators that produced 100 KHz at power levels reaching 200 KW. Radio stations could now communicate around the world, although their antenna arrays filled entire valleys with intricate metal cobwebs.
New government agencies evicted amateur radio operators to the useless "short wave" frequencies below 200 meters to prevent interference with commercial and military stations. Hams soon observed that their "high frequency" signals propagated over unexpected distances, which eventually led to the discovery of how RF bounced off the ionosphere.
Edison patented the Edison Effect in 1883, but didn't recognize its significance. Fleming turned that effect into the vacuum tube (aka valve) diode in 1904 and De Forest added a third electrode in 1910 to produce an amplifier or oscillator (sometimes both at once, a circuit tendency familiar to any analog designer). Within a few years, mechanical alternators became landfill and, amid a blizzard of IP litigation, electronics was in full swing.
In the 30 years from 1910 through 1940, the maximum usable frequency jumped six orders of magnitude to nearly 100 GHz and sensitive vacuum tube receivers reduced the need for rock-crushing transmitter power. Best of all, the field accumulated theoretical backbone behind the inventions, which often far outpaced the design equations.
Getting the Wavelength
Table 1 shows the chunk of electromagnetic spectrum we know as radio with frequencies and wavelengths covering 11 orders of magnitude. Although ELF and VLF signals lie in the same frequency range as human hearing, they are not directly audible. On the high end, the EHF band is a factor of 1000 below the frequencies of visible light.
The speed of light, about 300,000 km/s, relates the frequency (f) and wavelength (
) of radio signals. With c representing the speed of light, the equation is easy enough:
c=f×
Pop Quiz: Verify that Alexanderson's screaming alternators produced 100-KHz RF signals with a 3-km wavelength, and cellular phones in the 800-MHz band have wavelengths near 37 cm.
If it seems strange to call a 10-MHz signal "high frequency"—remember that those names date back to Alexanderson's day. When the PL-259 and SO-239 connector series was designed, "UHF" meant anything over about 30 MHz. They're now regarded as junk for signals beyond the middle of the VHF band.
Much of the RF action these days lies in the decade of frequencies above 1 GHz, with wavelengths most conveniently measured in centimeters. You can convert between f and in your head by setting c=30 cm×GHz. For example, an f=2-GHz signal has a =15 cm wavelength.
Table 2 shows the cryptic letter designations picked by early RF engineers on the basis of wavelength. It seems that L meant long and S meant short (with C a compromise between them), X marked the spot for targeting radars, and K-band came from the German kurz (short). You can find enough folk etymology to buttress any argument, so feel free to believe all, some, or none of the stories.
Because the band letters have no official basis, different groups used the same letter for slightly different wavelengths. The right-most column purports to show the new standardized band designations based on frequency, but as nearly as I can tell, these have no traction whatsoever. Of course, they conflict with the older designations, so you have no way to know whether C-band means 1 GHz or 4 GHz.
We'll just use the numbers.
Why Wavelength Matters
To transmit and receive RF energy, you must have antennas connected to your circuitry. Although nearly anything can (and will!) act as an antenna in a pinch, an antenna's size determines its capture area, a measure of how much RF energy it can intercept. For a given frequency (and, thus, wavelength), a larger antenna works better.
This poses a major problem at lower frequencies where a one-wavelength antenna requires Texas-scale real estate. The lowest U.S. amateur radio frequencies, 1.8 to 2.0 MHz, are known as the 160 meter band—525 feet. You can see why those early 100-KHz radio antennas employed suspension-bridge technology!
Mercifully, most applications don't call for a one-wavelength antenna. For reasons beyond the scope of this column, the most-favored portable antenna spans one quarter of the operating wavelength: a 1/4- whip. Your 850-MHz cell phone antenna is about 3.5-inches long, although various electronic tricks can adjust the length one way or another.
Pop Quiz: Work it out. (Hint: 850 MHz is 0.85 GHz.)
Making an antenna shorter for a signal of a given wavelength reduces its capture area and the amount of energy it can intercept, but the marketing folks for your handheld device may set an upper limit for the length that's independent of physics. Whip antennas have not yet become fashion statements: Note the number of people who use a cell phone with a retracted antenna.
Embedded devices also put severe restrictions on antennas. Most wireless network links for laptop computers have a patch antenna inside the thin plastic tab that sticks out of the slot. The antenna may be somewhat more complex than just an etched pattern on a printed circuit board, but it suffers from small size, bad position, and lousy orientation. That's as good as it's permitted to get, however, because teensy little rabbit ears haven't caught on, either.
Because the optimum antenna size is proportional to wavelength and, thus, inversely proportional to frequency, the operating frequency for newer devices has been rising. After all, if a 4-inch cell phone antenna works at 850 MHz, you can get the same results with an inch or so at 2.4 GHz. Even better, the marketing folks can edict that the antenna shall be buried in a nicely rounded, nonextending plastic radome.
Unfortunately, nothing in life is free, particularly free space path loss. Under some reasonable assumptions that aren't germane here, an RF source in free space (think arm's length) with an antenna length proportional to the operating wavelength suffers a path loss to a distant receiver with a similar antenna of:
Path loss=20× log4
d/
This means that the loss varies with the square of the distance and inversely with the square of the operating wavelength. If you double the frequency (which halves the wavelength), the power at the receiver drops by 6 dB: a factor of four. Quadruple the frequency and the signal drops by 12 dB, a factor of 16.
Now we can begin to understand why RF engineers have that perpetually worried look.
Thinking Inside the Box
Let's suppose you're building a handheld gizmo that requires wireless connectivity. Call it a cell phone, a hyperthyroid MP3 player, or a converged technology PDA—they're all facing the same issues.
Handheld devices must be hand-sized, ideally without eye-poking antennas. A flat-panel display and keyboard occupy most of the front surface, behind which will be the circuit board. The gizmo's external shape will be dictated by styling rather than the internal lumps that make it work.
The operating frequency can't be chosen arbitrarily, as it must fall within one of the specific bands allocated for that particular use. Not all governments allocate the same bands for the same purposes, so you may be forced to use the least-worst compromise frequency rather than the best one for your globalized gadget.
A higher frequency improves the efficiency of whatever size antenna fits on the box by a linear factor while simultaneously worsening the path loss by the square of that factor. Higher frequencies also interact with the nonideal real world in peculiar ways, as conductive objects begin to look like wavelength-sized antennas that absorb and reradiate RF energy.
For example, both the LCD and the circuit board look remarkably like RF shields that are a significant fraction of a wavelength. Any conductive material close to an antenna changes its radiation pattern, usually for the worse, and slabs 1/4- or 1/2- on a side produce serious problems.
From the viewpoint of radio signals, human beings can be modeled as irregularly shaped bags of salt water: Their moderately low resistance dissipates RF energy reasonably well. Handheld gizmos clamped against a head devote most of their signal energy to warming up the tissues and leave very little for actual communication.
Increasing the transmitter's output power to compensate for absorption and path loss reduces the operating life for a given battery size and technology. A handheld gizmo constrains its battery size, so you must choose a better and more expensive battery technology. A more costly battery gets you in trouble with the marketing people who have already picked a target selling price.
Firmware can help by allowing a more complex coding scheme that reduces transmitter airtime, but that generally requires a complex CPU or DSP with increased power consumption. Choosing a lower power and slower CPU leaves fewer cycles for user interaction, which can trigger devastating product reviews of your disappointingly sluggish gizmo.
There are no easy answers. The best you can hope for is a compromise that doesn't badly short-change any single feature. You must account for all the features when you're allocating the budget, though.
One last data point illustrates what can happen when you leave something out. The German Kurz-band radar effort during World War II produced a technically advanced 1.5-cm radar. That short wavelength provided excellent target range and size resolution, particularly compared to Britain's already obsolete 10-meter Chain Home radars.
The first units began operating in the winter (of 1944, I believe) and, by spring, reported decreasing effectiveness. Clouds and precipitation reduced the radar's range to the point of uselessness. Were the Allies jamming their new radars?
The Germans had inadvertently picked a frequency very near water vapor's strong absorption at 22.235 and 23.87 GHz. If you need long-range targeting data, K-band radars just aren't the right hammer for the job.
What exquisite little inconvenience have you overlooked?
Reentry Checklist
Recall that human flight went from the Wright Flyer to the B-29 Superfortress during those same three decades. There must have been something in the air.
You'll enjoy the well-written history of radar and RF electronics at http://www.vectorsite.net/indextt.html and Varian's invention of the klystron at http://www.varianinc.com/cgi-bin/nav?varinc/docs/corp/history.
If you have a big printer you can run off a poster of the U.S. band chart from http://www.ntia.doc.gov/osmhome/allochrt.html. A smaller chart of cellular frequencies lives at http://www.privateline.com/PCS/CellPCSchart.html.
The ARRL Handbook for Radio Communications (http://www.arrl.com/) will get you started on radios, bands, circuitry, and theory. It's not just for radio amateurs any more!
Historically minded reader Charles Gallo tells me that a group of freight railroads bought Conrail in 1997, cracked it into useful components, and handed CSX the tracks along the Erie Canal. If you've ever worked with legacy code, you'll understand why I was misled by the many boxcars still sporting Conrail livery.
DDJ
