On Fri, 17 Apr 2015 17:04:21 -0700, (Dave
Platt) wrote:

Incidentally, what's with the 33 cm wavelength notation? I'll start
using wavelength instead of frequency when HP, TEK, or some other
reputable vendor supplies a frequency counter or generator calibrated
in wavelength instead of the usual frequency or period.

Awww, I'll bet you even say "2 meters" sometimes ;-)

Sure, but when I feel like being a PITA, I often say "the 6 foot
band".

The EDZ is an omni (excepting pattern disruption from the tower or
mast it's mounted on). I'm not sure whether it, or an AMOS, would be
more useful in the sort of deployment I'm thinking about. Stations at
the outside edge of the event might benefit by the directionality of
the AMOS; the station in the center might want an omni.

Oh swell. Take two problematic J-Poles and stick them together and
hope that the result will be a better antenna. Two wrongs don't make
a right.

Ed Fong gave a presentation to the San Lorenzo Valley ARC a few months
ago on J-Poles. Here's the slide show from the presentation.
https://www.slvarc.org/wp-content/uploads/2015/03/dbj-2-slides_hawaii.pdf
http://edsantennas.weebly.com
http://www.work-sat.com/Antennas_files/FONG-DBJ1.pdf


The AMOS/Franklin antenna is NOT an omni thanks the reflector. It's
commonly used in sector antennas found in 800 MHz and up cellular
panel antennas. The stubs can be replaced by coils but it's still an
AMOS/Franklin antenna.

Neat! I like being able to build good, useful antennas from
indigenous materials (that's why I like the "cheap Yagi" design).

Rule-of-thumb. It's easy to build a low gain antenna ( 8dBi ) out of
junk parts, using sloppy construction, and loose tolerances. Things
become more critical above 8dBi. You could probably build a biquad,
AMOS-3, patch, or similar antenna using the aforementioned methods. An
AMOS-5 or AMOS-7 requires more precision.

What I have to work with at the moment is an 8640B (unfortunately
without the doubler),

I have 2 or 3 of those. Let me know when you need the gears fixed.
http://802.11junk.com/jeffl/pics/HP8640B/

plenty of scopes,

Yep. I have the same problem:
http://802.11junk.com/jeffl/crud/scopes-to-be-fixed.jpg
That's the "to be fixed one of these days" pile, which now has 4 more
scopes in the pile.

and a Systron-Donner spectrum
analyzer which goes up well above 2 gig.

You'll need a sweep generator. I have about 7, none of which work
well.
http://802.11junk.com/jeffl/pics/home/slides/BL-shop5.html
The logic was to buy 3 broken generators on eBay and make one that
works out of the parts. However, I goofed and bought one each
revisions A, B, and C, which are mutually incompatible. Argh.

I've been thinking that by
adding a decent directional coupler,

Yeah, that would work. However, I suggest you just get an RF sweeper
and a reflection coefficient bridge and be done with it. The real
problem with indoor antenna testing and the test equipment is that the
position of everything, including your body, has an effect on the VSWR
curve. You can move around my shop and produce almost any disgusting
pattern desired. It's much like RADAR, where the reflected signal
become the VSWR.

If you want, I can grind out the numbers for most any configuration
you want for 900 MHz.

I'd definitely appreciate it if you could crank out an AMOS or two of
varying lengths, for 902, and see what they look like.

Done using 4NEC2. See:
http://802.11junk.com/jeffl/antennas/AMOS-5-915MHz/
12.7dBi gain.
150 degree horizontal beamwidth.
10 degree vertical beamwidth.
VSWR = 1.15 to 1.50:1 in the 902-928 MHz range.
The wire is #12 AWG copper without any insulation.

Some details:
1. I changed my mind and made it for 915 Mhz instead of 902 MHz.
2. I goofed and made the reflector too short. It will work as is but
I need to recalculate and re-optimize for a slightly longer reflector.
3. The antenna is designed for 200 ohms and will require a 4:1 coax
balun. Something like this:
http://www.brest-wireless.net/albums/AntenneAmos/balun.jpg
but using 1/4 wave electrical coax at 915 MHz.
4. Most of the dimensions are a radius, not a diameter and that
measurements are to the center of the wires, not the edge.
5. The numbers are all in wavelengths. Scraping the NEC file:
SY W=.003 'Wire radius #12 AWG = 2mm dia.
SY Rx=0.50/2, Rnx=4 'Reflector width / 2 , number of wires
SY Ry=3.70/2, Rny=20 'Reflector height / 2 , number of wires
SY Dl=0.335472 'Half of center dipole length
SY Dh=0.216148 'Feed distance from reflector to dipole
SY D=0.147626 'Stub Length
SY E=0.183988 'Stub width
SY B=0.605056 'Length of 2nd element
SY C=0.512391 'Length of 3rd element

To convert these to cut lengths, take the wavelength at 915 MHz,
convert 915 MHz to mm wavelength = 328 mm, and multiple by the
calculated wavelengths. For example, the reflector is:
0.50 wavelegths wide or 0.52 * 328 mm = 171 mm wide
3.70 wavlengths long or 3.70 * 328 mm = 1214 mm long (about 4ft)
and so on.

I also claculated a larger AMOS-7 (14.3dBi gain) which is about 6-7 ft
long. That might be a too big for portable use. Might be worthwhile
if you really need 1.6dB more gain. Bug me if you want it posted.

I've got a few sheets of FR-4 and (certainly better) some Rogers
low-loss PCB stock, and could probably etch or cut-and-trim to make a
patch antenna of this size... once I figure out what the matching
section would need to be.

I don't think you'll be able to do much without a pile of test
equipment. That works ok for 440 MHz, but is twice as tricky for 915
MHz. At 2400 MHz (wi-fi), cut-n-try is quite difficult.


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