Impossible antenna, only $50!
Originally published 2007 in Atomic: Maximum Power ComputingLast modified 03-Dec-2011.
If you want more range for a wireless network, you buy a bigger antenna.
There are repeaters and power boosters and other such things, but a bigger antenna's usually part of the recipe. More power only allows a wireless device to be heard farther away, after all; a bigger antenna should allow it to hear farther away, as well. Since a working network requires the nodes to be able to make themselves heard and to hear other nodes, just cranking up your transmit power isn't enough.
The thing is - most add-on WiFi antennas are impossible.
They can't exist.
What's written on the box is contrary to physical laws.
Most consumer products don't have this problem. That can of cola may not actually make you thin and sexy, and that burger may not look like the picture above the counter, but nothing about the product descriptions violates any rules of nature. There aren't a lot of negative-350-millilitre cans or square-root-of-minus-one-pounder burgers out there. Well, not on planet Earth, anyway.
And yet it is quite easy to purchase "omnidirectional" WiFi antennas, with gain.
Gain is a way of describing how much stronger an antenna's output signal and input sensitivity are than some reference signal. In the case of WiFi antennas, gain's specified in "dBi", which stands for "decibel isotropic".
A 3dBi antenna has a 3dB, or twofold, gain advantage over a theoretical perfect isotropic radiator. That radiator spits out all of the power that goes into it evenly around it in a perfectly even sphere, and is similarly perfectly evenly sensitive to incoming signals.
That perfect isotropic antenna also wastes no power at all. Every picowatt of power that goes into it is emitted. Its gain is zero dBi.
So an "omnidirectional" antenna... with 3dBi gain... would have to somehow emit twice as much energy as you pumped into it.
Unfortunately for people keen on the ongoing search for machines that run on nothing, this is not actually what a 3dBi omni does. What it actually does is just... cheat. Omnidirectional antennas are not actually omnidirectional at all, and the higher their "gain", the less omnidirectional they are.
This is connected to the fact that the theoretical perfect isotropic radiator is another thing that can't actually exist in the real world. This is partly for reasons of relatively advanced electrophysics, but it's partly also just for the same reason that you can't make a perfectly isotropic light bulb, or a perfectly omnidirectional microphone. Awkward facts of the physical world mean that the structure of a light bulb blocks the light the bulb emits in some directions, and that the structure of a microphone blocks some of the sound that's trying to get to the mic diaphragm.
And so it is with antennas. Your typical "rubber ducky" WiFi antenna, as seen on most access points, has a quite noticeable "null" lined up with its axis. If you point the antenna (or antennae) of your AP at your WiFi computer, you'll probably notice the signal strength fall quite a lot. So the antenna is omnidirectional, but only in two dimensions.
The vertical null is actually a good thing, because it prevents the antenna from wasting power on the sky above it and the ground below it, where WiFi clients are unlikely to be. But the higher the gain of an "omnidirectional" add-on antenna, the wider become the cones above and below the axis of the antenna where it doesn't work well, or even at all.
People expect this sort of thing from an antenna that's supposed to be directional, like the famous "cantenna" or more sophisticated Yagi designs. Those are antennas-as-spotlights; they have a quite narrow transmit-and-receive "beam width", and if you're not pointing them in the right direction, you get nothing.
But it's easy to assume that an antenna labelled as "omnidirectional" will work the same no matter what direction you point it in. The higher the gain, though, the less true this is.
If a high gain "omni" antenna's vertically aligned, and you're roughly level with it, then you'll get more signal than you would from a smaller "omni".
If you're not sitting in the narrow signal disc around the antenna created by its high gain, though - if the antenna's a floor above or below you in a building, for instance - then you're likely to be very dissatisfied with the performance of your new super-antenna.
It's not even unknown for higher-gain "omnidirectional" antennae to have distinct gaps in their horizontal coverage. This is seldom actually a problem if you know to just try turning the antenna 45 degrees if some clients can't connect, but if you don't know that you've bought an antenna with a coverage pattern that kind of looks like a butterfly, you're going to be deeply confused.
All of this is, of course, only the tip of the iceberg of mystic antenna voodoo. But most of the rest of the deep and confusing world of radio frequency electronics design doesn't announce its existence with a blatantly paradoxical product description.
But, you know, I don't think I'd mind if this sort of thing happened more often.
People might stop and think before buying $500 jeans if the label said they were Klein bottle pants that could make your bum disappear.
