doofus <doofus@nomathwhiz.org> hath wroth:

>I built one of these for a unstable weak connection a couple miles away.

One of these whats did you build? Photo?

>I improved the connection, but it is still not stable; transmission
>appears to be better than reception using a orinoco pcmcia card (one of
>the best tested for receive sensitivity; power 200mw). Interestingly the
>software reports a fair 33% - 40% signal strength and link quality,
>which is as high as I got when I was alot closer with a good stable
>connection, but NOW that I am much farther away the connection is made
>to a web page but the data transfer stalls and I get nothing. It is
>related to signal strength cuz when I am close to some other access
>point, everything works fine.

33-40% is marginal for any connection. You may be trying to go too
far. Any particular distance in feet or meters? I kind prefer real
numbers to "alot closer". Anything in the way along the line of
sight? Are you very close to the ground, in which case you might not
have sufficient Fresnel Zone clearance? Any particular model Orinoco
card? How is your dish positioned between the card and the antenna?
The dish reflectors don't work very well with PCMCIA cards. Photo?

>Here is the template I used, but I doubled the size, EXCEPT I made the
>heighth about 1/2 the width (curve of parabola remains the same) The
>template from http://www.freeantennas.com/projects/template/ calls for
>height and width to be the same-a square.

That kind reminds me of my high skool math problems, where the purpose
was to obscure the actual measurements. Since this isn't a quiz,
could I trouble you to supply actual measurements in feet or meters
instead of the math quiz format?

>Since my math only goes through Trig, how critical is the height/width
>ratio and does it have to be a square to be maximally effective.

Assuming vertical polarization, the height has to be at least 1 full
wavelength for a dipole (1/2 wave) feed. That's 12.5 cm at 2.4GHz.
Larger is better, but less will be a problem. If the feed is longer
than 1/2 wavelength, the total height should be at least 1/2
wavelength longer than the feed.

The width is a function of the f/D (focal length to diameter) ratio
and the feed illumination angle. It can vary substantially. Since
neither of these are particularly easy to determine with a PCMCIA card
feed, the optimum size is currently mostly guesswork. I can go more
into detail once I determine what you've actually done so far.

>I am
>just on the borderline now of getting a stable signal. What can I do to
>increase the strength, particularily the receive end?

Lose the PCMCIA card unless it has an external antenna connector.
Replace with a PCMCIA card that does have an external antenna
connector. Build or buy a real parabolic dish antenna, yagi, panel,
biquad, cantenna, Franklin, AMOS, whatever antenna. Buy a pigtail to
connect between the PCMCIA card antenna connector and the dish antenna
connector. The antennas on FreeAntennas.com were designed to somewhat
improve the signal by redirecting the RF in some general direction.
The gain is not huge but certainly useful. It's also cheap and easy,
which makes it a good starting point. However, you are apparently
going for the DX record over some unspecified distance. That's going
to take a more complex system, possibly using commerical antennas.

>Can someone take a
>look at the template at the url above and tell me where I am going wrong
>or what I can do to improve the antenna?

The template is fine. Post a photo of what you've done and some
numbers and we'll try to help.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558