It was an interesting weekend for the NoCat crew. In addition to turning on another public node (and point-to-point link), we got in some very good R&D time.
Implementing the Wireless Web
It all started last Friday afternoon. We were all buzzing about Cringely and his Passive Repeater. What a fantastic idea! The thought that he got one of these near-urban-legend devices working, on the first try, from up a tree, to low-gain panel antennas in downtown Santa Rosa, without a supporting ground crew (or even so much as a site survey), truly boggled our minds. Especially since it’s been a point of debate in various community group mailing lists for months as to whether such a design is practical at the low power level that client cards put out.
As he’s a local, I immediately wrote him and asked if he’d be interested in attending our next community meeting (on Wednesday) to show it off (naturally, cc:’ing our list in my excitement.) As he decided not to cc: the list on his reply, and evidently won’t be attending the meeting, I guess I’d better summarize his response rather than repost.
It seems that he doesn’t think he did anything special in the design. He mentioned that choosing the proper type and length of coax was critical in getting a working design going, and that it was based on the passive
repeater design he linked to on his site.
I’m even more intrigued that the design he links to doesn’t even use coax, and that he seems to be peculiarly evasive in providing technical details of his almost unbelievable accomplishment.
I can understand how busy one gets with a high-profile website: I get a huge number of “wireless tech support” email requests from readers every day, and I can’t possibly write them all back intelligently. And if Mr. Cringely has
in fact done the miraculous, I can certainly understand how he might rather keep his design to himself: such a design, if patented, could make a small fortune in the WISP market.
But since a device that works as well as he claims could dramatically change the world of wireless networking, I invite you to show your good will to the community who gave you the design, and come down to O’Reilly and
This past weekend, we found three key spots on the O’Reilly campus, forming a rough triangle. At one corner of the parking lot, you can see the single large white rock in our back yard (about 450 feet away). Between our warehouse and the center building, at about 600 feet away from the first point, you can see the same rock. There are two whole buildings between the end points, and in our preliminary tests, laptops with radio cards at these spots cannot “see” each other. We then put a battery powered repeater on the rock, which the two laptops could immediately detect.
If the Cringely design works as well as claimed, then setting up a “passive repeater” on the same spot should allow the two laptops to communicate with no trouble at all (in his article, Cringely claimed being able to pull 2Mbps
over a significantly greater distance).
Mr. Cringely: We would LOVE to have you demonstrate your design. In fact, I’d be happy to document and publish it online personally, as such a piece of gear would be a tremendous boon to the community wireless networking effort. If you’re willing to demo your repeater at O’Reilly and let us put your design specs online, I’ll happily give you a complimentary copy of my book, and I’m sure that you’ll have the sincere thanks of the wireless community at large. Please consider putting in an appearance at our meeting on Wednesday. Thank
Anyway, gentle reader, I don’t want you to go away from this little rant completely empty-handed. While at the new O’Reilly buildings this weekend, we decided to try a new antenna design. Actually, it’s an old antenna design, but it’s the first time we’ve gotten around to trying it: the Coffee Can.
(Yes, the Standard Disclaimer applies here as well: YOU are responsible for anything you connect to your radio! After all, this is the Web! Check your
work! I might be wrong, or lying, or both!)
We got the idea from a 2.4GHz video feed antenna from this site. The design is MUCH simpler and easier to make than the Pringles Can. You start with a 36oz can of “premium” coffee. We used href="http://yuban.com">Yuban, in the big brown can. Ah, Yuban. The Essence of Coffee ™. Whatever particular brand you choose for your project, make sure it’s a steel can (you’ll need something better than cardboard for this job.)
You’ll also need an N connector, and a short piece of heavy gauge copper wire. The length of the wire depends on how your N connector mounts to the can: you’ll want the top of the wire to be 1/4 wavelength from the inside edge of the can when the connector is installed (just as in the Pringles Can design.) In our can, this was approximately 1 5/8″. Simply cut the wire cleanly, and solder it directly to the connector. Now, where to mount it?
Looking at the can from the side, you’ll notice a number of lateral ridges. By a staggering coincidence of the universe, the bottom ridge on a 36oz Yuban can is almost exactly one quarter wavelength at 2.4GHz. Poke a hole
in the can big enough for the connector to pass through. (Just below the words 100% Colombian Coffee seemed ideal to us.) We used a hammer and nail to start the hole, followed by an awl, then a hand file. We simply poked mounting holes through for the connector hardware using a hammer and nail. Be quick and careful if you use a hammer, and don’t dent the can any more than you have to.
Before you screw in the mounting hardware, scrape the paint off of the area that the connector is bolting to. You want a nice, wide connection between the connector and the can’s body. Use the shortest screws that will hold the connector firmly in place.
Finally, put the plastic lid on the top (or glue it down, if you need weatherproofing ;) ) and you’re done! We made ours in under ten minutes.
How well does it work, you ask?
To find out, we went to the new O’Reilly site with the usual test equipment: some antennas of known gain, a couple of laptops, two tripods, and our coffee can. First, we set up a good shot between two points with clean line
of sight, using 24dBi dishes on tripods placed as far apart as the parking lot would allow. Open line of sight is amazing: At about 600 feet apart, using 30mW Lucent cards, we saw a whopping -36dB signal between the two points! (If you’re just joining us, that’s roughly enough signal to drive an ice cream truck over.)
Next, we removed the dish from one side of the link, and substituted our can. Once properly aimed, we saw a -44dB signal on both ends. This is about 8dB worse than our 24dBi dish, or a whopping 16dBi! And this is all without installing a single washer or cutting a single piece of pipe (as with the Pringles Can).
Considering that commercial antennas of comparable gain sell for well over $100, I think we got quite a deal. Commercial antennas don’t even come with complementary coffee!
So, don’t forget, kids: Tell Everybody Everything You Know. Life is too short for secrets. And Have Fun!
UPDATE: I’ve just received word that due to personal matters, Mr. Cringely won’t be able to join us for at least a month. I let him know that the invitation stands, and that he’ll have a very appreciative audience when he’s ready to present.