Like many of you, I have a modest 802.11 wireless network at home. One link–going over a distance of about 25 feet (7.5 meters) and through two thick, old-fashioned plaster walls–was marginal. I was always getting readouts of low and very-low signal strength, as well as link breakups due loss of sufficient signal. The wireless router at the hub PC has a small, non-removable, rotatable antenna, about 3 inches (8 cm) long; the antenna at the peripheral PC is similar but can be removed via the SMA connector at its base.
I needed a solution to this frustrating problem. I had even tried using a cookie sheet behind the antennas, to add some directionality and gain, but that didn't help. My next plan was to get a range extender, which is an active signal-booster box (about $40) that is placed along the link path. The other possibility was to go “retro” and get wired, by running an Ethernet cable along the floor. It would take a 50-foot (15 meter) cable and it might be ugly, but it should work (about $30).
So I went down to the local Officemax to see what they had, and while browsing found a different solution: a high-gain antenna plus extension cable designed for this situation. The ANT24-0700 from D-Link (www.dlink.com) is a 10-inch (25 cm) antenna which claims 7 dBi omnidirectional performance, and also includes a 5-foot (1.5 meter) RG-174 extension cable with those tiny SMA connectors. (If you don't know what dBi means, you should; one brief discussion is here). In this design, the gain flattened the radiation pattern, increasing performance 360° around the antenna while reducing the “up and down” performance, according to the diagrams on the package (very thoughtful of the vendor to show those!).
It's an entirely passive solution that not only would add some gain, but also allow me to move the antenna itself around a little, to find some local lot spots and avoid some dead ones. Experience and analysis show that you can have mini-zones of each type in real-world situations, due to RFI/EMI, nearby physical objects, other wires, and many other mysterious factors.
Long story short: this antenna solved the problem very nicely. Signal strength now consistently reads good to very good. Even though the new antenna assembly seems overpriced at $40, it's actually worth it. It's a no-hassle solution to the problem, since it does not require running a long Ethernet cable, adding software, performing initialization, adding another box and its power brick, nada. It's 100% transparent to the system, and it is truly “plug and play”, to use that often-abused cliché.
But the whole incident got me thinking again about the term “gain” with respect to antennas. As a passive device, an antenna does not have gain in the sense that an amplifier does. What antenna gain means is that the radiation pattern of the antenna is focused in some way. It's really an energy-redistribution scenario rather than the simple gain of an amplifier. As such, you have to be careful when using and exploiting antenna gain, because you enhance performance in one direction, and reduce it in others. Of course, in many applications, the reduction in some directions is a very good thing, as it minimizes unwanted signals from some sources, or wasting energy towards non-receivers. Thinking about and working with antenna gain is a much more complicated topic than it is for amplifier gain.
One more thought about my larger-antenna adventure: I know that small, and even internal antennas are popular, especially in applications such as cell phones. And I know that they can deliver decent and acceptable performance. Yet my gut feeling, experience, and the efforts of others show that, sometimes, what you really need is a physically larger antenna, both to shape the pattern and to capture more signal energy, and thus boost both your S and your SNR. Just as a larger lens intercepts more photons, rather than providing mere image magnification, “more metal” can sometimes be the right way to go (see also “Rice-grain antenna piques my interest, but confounds my intuition”). ◊