If this were a live FonBar, you would log in to the hotspot through the form above.

Wifi Reflections [1 EDIT]

I’ve been playing around with a pair of 6-inch parabolic reflectors, based on the “Windsurfer” design found here at Freeantennas.com, and thought i’d share my luck so far. I have my trusty Linksys WRT54GL router inside the house, and myself positioned about 40 (12m) away, on the front porch. I use DD-WRT, so I have boosted the transmitter’s power output to 125mW. The house is all wood-frame construction with no obstacles other than a metal gas stove in the line of sight. I use an Orinoco Gold A/B/G PCMCIA wifi card, which has a reputation for reasonably accurate strength readings when using NetStumbler.

Click the photo for a large version:

At first, I just ran NetStumbler with the parabolic dishes both pointed toward my computer. I get about -63db, which is an excellent signal. Second, I turned both dishes in the opposite direction. The signal plunges to about -78db. Third, I took the parabolic reflectors off to see the signal I get without assistance. This is around -68db which is fair to excellent, but it experiences occasional complete dropouts. I rarely get these dropouts now that i’m using the reflectors. :)

Here’s a Shockwave Video tutorial on building the Windsurfer. For my own reflectors, I used tape, cardstock paper and an Exacto knife, and got neater looking results. When you’re done, your router should resemble K-9 s head.

There were some very interesting pictures on Freeantennas.com that you might be interested in seeing: Deep Dish Cylindrical Parabolic Template and Lots More Pictures.

An excellent parabolic reflector design is also provided by fellow Fonero Kyros (.pdf format). While the Windsurfer design states that the template can be scaled up proportionately, Kyros argues otherwise.

Since my 6 reflectors are delivering less of a boost than the notes on the Windsurfer template suggest, i’m going to take my chances and build a pair at double size.

EDIT: I made reflectors at 150% the scale of the small ones (about 10 ). They’re so big that they can’t both point in the same direction without overlapping a bit. While the new reflectors are still better than none, the smaller reflectors outperform them by about -5db. Can anyone offer advice about what “sweet sizes” these reflectors should be built at? If they can’t be scaled up proportionately, then what is the next larger size I should construct at?

0 Responses to Wifi Reflections [1 EDIT]

  1. Mike Puchol says:

    Shameless plug – you could use the antenna comparison tool built in StumbVerter ( http://www.sonar-security.com/sv.html ) to see which antenna is working better. Nice post!

    I also agree with Kyros in saying that reflector-based antennas don’t scale linearly, you have to take into account the wavelength of the target frequency, otherwise you risk out-of-phase reflections which negatively affect the signal.

  2. Jeff says:

    I may be stating the obvious to you, but someone else will come along, see “125mW” and get a crazy idea…

    Jacking up the power does not necessarily improve things. At some point it will just increase the noise and cause bleed-through to other channels. (there’s a great graphic out there somewhere, I can’t find right now). Your laptop may have an easier time hearing the AP, but it does nothing to help the router hear your laptop’s tiny output. And, more power = more heat. More heat = more noise.

    As the old saying goes, “a dime in the antenna is worth a dollar in the transmitter” — the Linksys omni antennas are only 2dBi IIRC. That said, reflectors do redirect unused energy where it can do more good. This can also restrict the signal to avoid spillage where you *don’t* want it going (maybe not much interest to a Fonera though).

    Line-of-sight is also important, obviously. You also might also experiment with antenna diversity, although I haven’t seen much in that department. I myself am going to play with Fontenna — I just got one this week, but it’s too early to tell.