Critical Dimensions for Parabolic

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

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.

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

calls for height and width to be the same-a square.

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

Below is a reflector that I made using the bottom pattern. Worked pretty well for a simple setup.

I like the

reflectors. They are cheap and easy, and make a good starting point... oops, somebody already said that.

I do wonder what the OP did with the reflector that he made. It doesn't fit so well on the Orinoco PCMCIA card that he mentioned, unless there are some pieces missing that I don't understand.

If you have marginal signal at some location, adding a reflector might be all you need to do. Maybe adding one at both ends. Or, if adding the reflectors gives you a marginal signal where there was none before, you can guess at what dB gain you've added over stock, and guess again at how much gain you might need for a stable connection.

Now you can buy a real antenna, of, say 24dBi gain, knowing that is

10dBi or more higher than the reflector on a stock antenna, and have some confidence that it will work, since the free reflector already sort of worked.

Your half-height reflector might not be so good, according to Mike, the author of the site. "Errors of 1/4" are unacceptable at these frequencies." "Parabolic reflectors also loose gain if your finished reflector varies much from the correct curve." I found the curve too hard to maintain, and I prefer the "Windsurfer" EZ-12 from the same site. "you will lose roughly

3 dB for each halving of reflector height" Mike leaves out all the calculations. Jeff offered to crunch numbers for you if you measured the one you built. Was the image square still square? I didn't have a problem with that, but Mike mentions it as a problem area, throwing off the curve.

David Taylor made a free standing antenna using the parabolic template for his reflector.

has some other interesting antennas on the site.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Thanks for the reply.

See the url, there is a diagram and photos of the one I built. Looks like half a tin can with rubber ducky places at focal point. More below:

Guestimate, stable connection at 33-40% was at about 1km, now at about

4km. Line of sight, hahahaha, u jest. Just walked around with the laptop to see where I could find the strongest signal. Cannot actually see the transmitter, but the signal must be pretty strong, cuz I am marginally connecting now, even at 4km, but it is too iffy. The Senoa card, the one most are using with the prism chipset.

Just double the size of the template diagram at the url. Cut thin sheet metal, fold around foam supports to form the parabola in the template diagram given. Only they are using a square piece of sheet metal and mine is twice as wide as high, roughly 12 inches by 6 inches high. I don't have a ruler handy but the exact dimensions can be gotten off the template diagram at the url. I am not at home, so I don't have a ruler with me here, but the dimensions can be gotten off the template at the url, only I doubled the size and made the height half as high as the width.

Thats greek to me, haha. Using a single rubber duckey 6 inch 5Dbi whip as the focal point foam supported in a half cylinder parabola (similar to 1/2 tin can but shaped more precisely) See approx. dimensions above. Connector between antenna and card is about 6 inches long, so loss minimal there.

You might as well be speaking Lithuanian to me ;-). Please take a look at the "parabolic reflector template v.3" at the url above. Double the size of it, as I enlarged it by a factor of two. Now bend your metal over foam supports conforming to the template curve and insert a roughly

6 inch rubber ducky into the foam vertical at the focal point of the parabola. That is what I have, EXCEPT that my metal parabola is one half as high as it is wide, yet still covers the complete heighth of the rubber duckey, whereas there's is a square-as long as it is wide. Does that explain what I have done any better?

Mine looks similar to this:

doofus hath wroth:

The URL is for someone elses reflector. I wanna see what you've done.

I think you're doing just fine if you can get a connection at that distance, especially without line of sight. That's much too small an antenna to do that range reliably. It's just too far. The small dishes are good for perhaps 300-500 meters maximum. Any further improvements will need to be in either the size of the antenna, or the construction (i.e. feed illumination angle).

Well, thats a fair test. However, you'll probably notice that the signal goes up and down radically with traffic and position. That's because of the Fresnel zone and reflections.

At that range, you should have some altitude. For example, at 4km (2.5 miles), you should have 9 meters (30ft) of clearance which means both ends should be at least 9 meters off the ground.

Ok, no model numbers. Sorry, no help.

Got it. 12 inches long and 6 inches high. That give you an aperature size of about 6". Maximum gain for a 0.15 meter (6") parabola is about:

Maximum gain for a 0.15meter diameter dish: gain = 9.87 * Dia^2 / wavelength^2 * (feed efficiency) gain = 9.87 * 150mm^2 / 125mm^2 * 0.4 gain = 5.7 dBi = 10 log(5.7) = 7.5 dBi gain This is the best that can be done with such a small dish. The 40% efficiency is probably optimistic. Because your feed method is very inefficient, it will be much less than 7.5dBi.

That's fairly close to the original. The problem is that the rubber ducky feed radiates in all directions. Only a part of that radiation hits the reflector and goes in the desired direction.

The 6 inch 5dBi rubber ducky is another problem. The original 2dBi version only radiates in the upper half wave (6.25cm) section, so the vertical height of the reflect can be as short as 12.5cm. However, the 5dBi mutation radiates over perhaps 12.5cm in length. It should be at least 15.6cm (1/2 wave plus 1/4 wave above the feed tip) wide.

18.8cm would be better if you can extend the reflector another 1/4 wave below the ground end. Translation: Make the height somewhat longer than the feed antenna.

Will broken Polish, French, Hebrew, Yiddish, Russian work any better?

Done.

Is there some problem with supplying your dimensions? I can't tell if you doubled the width, height, volume, depth, f/D ratio, wire thickness or what.

No, but I can guess that you're really asking is does the aspect ratio have to be 1:1 or does the height have to be the same as the width? The answer is no. The height should be 1/4 wavelenth (3.125 cm) above the tip of the antenna, plus possibly 1/4 wavelength below the base of the antenna to insure sufficient illumination. The width will need to follow a parabolic curve. If you extend the width of the reflector, just follow the general pattern for the parabola, or extend it with a new and larger template. There is no direct relationship between the height and the length.

If you're getting connections at 4Km with that arrangement, I am really impressed. Whatever you're doing is working just fine.

Incidentally, I once built a WX satellite parabolic reflector out of a long piece of aluminum roofers flashing. The frequency was about

1.7GHz instead of 2.4GHz. I used a piece of 4x8ft plywood, jig sawed into a parabolic shape. I then stapled the flashing to the parabolic curved cutout forming a reflector that was about 12" high, but perhaps 12ft long. Initial measured gain was about 18dBi, which is far less than the 26dBi I had expected. It worked amazingly well, but was mechanically unstable, difficult to aim, and was far too ugly to sell commercially. However, it might give you an idea. Sorry, no photos.

snipped-for-privacy@59.usenet.us.com wrote in news:f89eo1$n4t$ snipped-for-privacy@blue.rahul.net:

Thats a invalid link.

snipped-for-privacy@59.usenet.us.com wrote in news:f89eo1$n4t$ snipped-for-privacy@blue.rahul.net:

I used the same connector that comes from netgate with their wubber duckey. They have a pigtail for all kinds of pcmcia cards.

don't own the other end; did not see any pix of additional reflectors-i'm guessing you mean in addition to the parabola and not sure how that would work with a wubber ducky as focal point. I usually only see reflectors on dish types.

Well,I've gone this far and i'm guessing that the next step up in commercial stuff would be a dish antenna which introduces aiming problems, especially in light of the fact I don't know if/what the obstructions are between me and the transmitter.

I think he mean errors in positioning of the wuber ducky at the correct focal point, not errors in overall antenna measurements. CORRECTION 4 Jeff: The measurements are actually 9" by 9" so I do have a square after all (don't know how I got that wrong).

BOTTOM LINE?: Make it bigger with the same curve. Right?

Will check, thanks.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Ok, the measurements are actually 9" by 9" so I guess I do have a square after all. I'm guessing that just making it bigger with the same curve might give me the extra oomph I need.

yeah, it's just too far.

Yes, ur right, very sensitive to position and angle. Is on an odd channel for here, 11, so don't have to worry much about interference. I'm at about 18 feet and can't go any higher cuz it's a crappy apt. building.

Here's the exact hardware I am using sans antenna:

a9187acc6d8d7e92205725

Card is: Senao 2511 CD PLUS EXT2

The card is super. very powerful. and netgate is the best.

sorry I got it wrong, it's 9 X 9" Will try building it bigger.

One thing I don't get, maybe you can explain. Why is the card software reporting the same percentage of connection 33-40% as I was getting when I was 3 times closer and had a stable connection, yet now, no good connection?

Yeah, don't know how I'd capture the rest of it without getting rid of the rubber duckey which is what my present and only pigtail connects to.

Already is about 1/2 higher on both top and bottom. Guess I have to go wider.

even though ur usually talking above my head, I do learn stuff from you.

only problem is that I don't have internet at home now, that's all.

Very unstable, only works maybe 20% of the time. Going to increase the width of the parabola by point 5 or more. Also were having wet weather so that's another variable.

I suppose a dish is out of the question, since I don't know what obstructions are between me and the transmitter; can't even see the transmitter which is on a tower about 50 feet up from memory when I was alot closer.