802.11b Reception Tricks
by ddShelby
Since the article Comprehensive Guide to 802.11b in 19:2, I dove headfirst into wireless. I would like to acknowledge Dragorn for a well-written article. I also would like to acknowledge oriellynet.com, seattlewireless.com, and turnpoint.net for the information contained in this article.
Supposedly because of a dispute with Time Warner and the landlord, a cable Internet connection is not available in the apartment building in which I live. DSL is available but seemed a bit steep at $70 a month for a 128k line. So I considered wireless. However, NYCwireless nodes on the Upper East Side of Manhattan are few and far between and my rather anemic Netgear wireless can't reach the nearest node.
So I looked around for an 802.11b card that has provisions for an external antenna and settled on the Lucent ORiNOCO Silver. It's a 40-bit WEP card only but it was cheap on eBay, so to me it did not matter. I picked up a four-foot pigtail cable that adapts the connector on the ORiNOCO card to an N male connector from FAB Corp.
Some Connector Basics
There are several types of connectors used in the 802.11b world that need mention.
The most common is the N connector. These are usually found on the antennas themselves and it seems that this is the norm. The antennas I have come across thus far are all equipped with a female N. The other side of the cable (pigtail) has the connector that will attach to whatever device you are connecting to. Here is where it can get a bit hairy.
Devices like access points or wireless bridges can come with a BNC, TNC, or an SMA connector. Connectors on the Wi-Fi NICs depend on the model and manufacturer of the card. To complicate things just a bit, all of these connectors are available in reverse-polarity. Simply put, the small gold pin in the center of a BNC is a male pin.
On a reverse-polarity BNC, the gold pin is female. The reverse-polarity connectors are usually indicated as an RP-BNC for example. Just for reference, BNC is an acronym for British Naval Connector, TNC is a threaded BNC, and SMA is Subminiature Type-A connector. All of these connectors, I suspect, originate from the military.
A search on Google revealed a few sites with information on antennas for 802.11b. O'Reilly had the most extensive information I could find (www.oreillynet.com) and is a great place to start if you're new to this like I was.
My first antenna was the famous Pringles Yagi.
I constructed it exactly as laid out on the www.oreillynet.com/cs/weblog/view/wlg/448 web site and found significant gains as compared to the ORiNOCO card without any external antenna. A total gain of 11 dB was the best I could do with the addition of a Pringles can as compared to the ORiNOCO card itself.
The other antenna choice is the waveguide antenna.
The construction of the waveguide is easier since it does not involve the use of a threaded rod and washers as the Yagi does. The can itself and the addition of an N connector with a piece of copper wire is all that's needed.
For the copper wire I used a piece of grounding wire from common household electrical wire. With the simplicity of the waveguide construction, you can sacrifice many coffee cans at no significant cost, especially if you're a caffeine nut like myself.
The ideal waveguide antenna for 2.4 GHz is about a 3.25 inch diameter and just shy of 10 inches long. Good luck trying to find those dimensions in a coffee can or anything for that matter on the grocery store shelf. But this being said, there is no harm in experimenting with what you have lying around the house.
I first tried an 11-ounce Maxwell House can. I mounted the N connector accordingly at one quarter wavelength from the back of the can as calculated by the handy script located at www.turnpoint.net/wireless/cantennahowto.html.
As compared to the Pringles can, the Maxwell House can gave me an additional 3 dB for a total of 14 dB.
Keep in mind that every 3 dB is a doubling of the signal. A loss of 3 dB in noise is as good as an overall gain of 3 dB with respect to the signal-to-noise ratio.
Interesting thing happened though; using NetStumbler I picked up three more access points that I did not see before. This could be due to the additional gain but I thought it might be the type of antenna construction lending to a wider pattern. So I tried again with a larger diameter can to see if my theory was in fact correct.
I chose the Folgers 39-ounce can and cut a hole according to the handy script on turnpoint.net. I reused the N connector from the 11-ounce Maxwell House can to avoid unwanted variables.
As it turns out, the gain fell slightly to 13 dB but I again noticed two additional access points according to NetStumbler. With the 39-ounce can, I now picked up a total of 11 Access Points (APs) as compared to nine APs with the 11-ounce can.
Of these APs, by the way, four show up in the list printed in the Fall 2002 edition of 2600, and still remain unencrypted.
For those of you into warchalking, the larger wave guide from a 39-ounce can seems more appropriate than the Pringles Yagi or a waveguide closer to the 3.25-inch optimal diameter.
Although you may prefer something omnidirectional like a mast antenna, the overall gain is typically lower. So if you are looking for directionality in the signal, then stick with narrow diameter waveguides or Yagis.
If broad coverage is what you're after then go with wide diameter waveguides or mast antennas.
Having established the difference in gain and beam pattern associated with the size of the can, I launched a quest for the ideal 3.25-inch diameter can. I needed as much gain as I could get just to reach the nearest NYCwireless node closest to my apartment.
Blocking that node are three high-rise apartment buildings, two parking garages, countless brownstones, and three blocks. After a near exhaustive search for a 3.25-inch diameter can at 10 inches long, I decided to just spend the dough for a commercial 2.4 GHz antenna.
It's a dish style that has an advertised gain of 14 dB. The noise on this commercial antenna is slightly lower than any homemade antenna I had constructed, so the overall signal-to-noise ratio was in my favor by about 3 dB. Despite this, the signal-to-noise ratio was still not enough to get a consistent connection, and dropouts were still too common. So then I thought, do I have to spend even more money for a higher gain dish? Well, not quite.
Dream Cans?
Ah... well, sort of.
It certainly looks like the 3.25-inch by 10-inch ideal. While shopping for new roller blades at a Sports Authority on Long Island, I noticed a tennis ball can.
Most tennis ball cans are now made of the same plastic as soda bottles. But this one is a bit different. Wilson makes an oversized tennis ball for the geriatric crowd, that just so happens to come in a steel can that's 3.25 inches in diameter. And the icing on that cake is that the length is just about 10 inches. My three tennis balls were about $6 and the N connector was $2.
I punched a hole in the can at 2.49 inches from the bottom and mounted the N connector as the turnpoint.net script calculated.
The result was 17 dB gain, just enough for what I needed to get a clean signal to the AP. Now 17 dB for a tennis ball can is more gain for the money than you might imagine.
A commercial antenna at 14 dB like the one I bought cost up to $80 and does not include any green fuzzy things to play with. The drawback is that I had to sit near the window with my laptop. My pigtail would only let me stray four feet.
Two Weeks Later
This new Linksys WET11 is neat.
The Linksys WET11 is sold as a bridge, not an AP, essentially giving a Cat 5-only device the ability to go Wi-Fi or, using two of these WET11s, to connect wirelessly to each other to bridge two wired networks.
I got to thinking and wanted to experiment to see what else this thing was good for.
I wanted to connect through the Linksys WET11 with an AP I already had lying around. So, I picked up a reverse-polarity SMA to N male pigtail from FAB Corp to hook up my Wilson antenna to the WET11.
First, the WET11 output is rated at 71 mW, which is more than most Wi-Fi cards and more than twice the rated output of my ORiNOCO Silver. With an antenna other than the rubber duckie mast provided, there is the potential for some serious range.
Also, I wanted to see if I could set up a kind of repeater. So I took the 10-base output from the WET11 and plugged it into my el-cheapo Netgear AP and set the Netgear to a different SSID from the WET11.
The results: The Netgear AP worked locally as any AP would, the signal goes to the WET11 via Cat 5 crossover cable and to the AP that it's aimed at a few blocks away.
The connect speed was good enough to give me Internet access in my New York City apartment wirelessly. And with the WET11 sitting on my window sill and the antenna on the fire escape, I have the ease of surfing from my kitchen table or anywhere in my shoebox apartment without having to contend with the limitations imposed by the four foot pigtail that connects my antenna directly to my ORiNOCO card. And with the higher output and increased sensitivity of the WET11 versus the ORiNOCO card, I can use that dish I bought without feeling guilty for spending 80 bucks for it.
Another Waveguide Idea
There is another design in waveguides that can pull up to 18 dB if constructed carefully.
If you take shortcuts or if it's poorly constructed, you can still obtain 13-14 dB. The details on its construction can be found at www.seattlewireless.com.
It's constructed using a peanut can and some stovepipe fittings from Home Depot. Stovepipe is thin sheet metal and not much different from the material used to make your typical soup can off any supermarket shelf.
In this case it's an adapter (sometimes referred to as a reducer) to go from a 5-inch diameter to a 4-inch diameter. This acts to increase the radio waves collected before they enter the can amplifying the overall gain by as much as 6 dB.
Experimenting with various sizes and lengths can be worthwhile and who knows? You might stumble onto something.