On Sat, 28 Apr 2012 19:16:27 -0500, amdx <
am...@knologynotthis.net>
wrote:
> But Jeff, you could do the wifi DIY's a big help if you could perfect
>the double double and still keep it simple.
My idea of simple is the Franklin/Amos antenna. See various examples
near the top of:
<
http://802.11junk.com/jeffl/antennas/>
The main problem with building a double-double biquad or AMOS is the
wire standoffs. Cheezy nylon doesn't work. Custom ceramic or glass
are tempting if you're into pottery or working glass. It's really
tempting to make the entire AMOS wire antenna out of a polysufone (not
G10/FR4) PCB (printer circuit board). That's easy for the AMOS as it
could be single sided, but with the double-double biquad, it would
need to be etched on both sides to deal with the crossover. If you're
going to build just one, all this is irrelevant. If you're going to
manufacture, it's vital.
I'll throw together a proper NEC model and offer some construction
ideas for do-it-thyself constructions. With semi-rigid coax, the
double double biquad is quite easy. However, no now. I'm busy with
paying work.
> I wonder if the pcb is a stripline that is used to transform the
>impedances? That would be the most difficult part for many.
I can't figure out what he's trying to do. Each double biquad section
has a 100 ohm feed point impedance. One could parallel two of these
to get 50 ohms, if they were on top of each other, or separated by any
multiple of 1/2 wave sections of mythical 100 ohm coax cable. (75
ohms would be close enough). Yet, the stepped PCB strip line in the
photo is doing something else, which I can't guess without dimensions
or a description.
The original design apparently came from this site:
<
http://www.zero13wireless.net/foro/forumdisplay.php?46-Antenas-Direccionales-Wireless&s=761a47b653870eb321cd3ed14ca58cb7>
but I can't find it.
>If you can't design it to eliminate the stripline transformer (if that's
>what it is) you could sell the pcb!.
No thanks. I can make money selling expensive antennas (or expensive
consulting), but not cheap antennas for experimenters. If you want
cheap and easy, think about multiple patch antennas on PCB material.
For example:
<
http://www.ebay.com/itm/170778036572>
14dBi gain is a bit optimistic, because it ignores matching section
losses. My guess is 12dBi maximum on a 5.5" x 5.5" PCB. I have
several of these, but have never bothered to measure the gain or sweep
the VSWR.
>I like it because the construction is so simple.
I don't think it's that simple. Permit me to remind you that things
become more critical as the gain goes up. One can be rather sloppy
when building antennas with gains of less than about 12dBi. Higher
gains require a pile of test equipment and some precision
construction.
>Can you really get 16-17 db gain in 7" x 7"? :-)
<
http://802.11junk.com/jeffl/antennas/DoubleBiquad/index.html>
Close. Doubling the size of an antenna yields 3dB more gain. The
model for the original double biquad claims 13.5dBi gain. Two of
these, properly combined, minus about 1dB combiner loss, would yield:
13.5 + 3.0 - 1.0 = 15.5dBi gain
Using lossy materials (i.e. leaky standoffs, lossy coax, and lossy PCB
material) will reduce this some unknown amount. The 15.5dBi is the
very best that can be done. Your mileage will be less.