NetStumbler.com Forums - Post Your Homebrew Antennas = Here

From: Subject: NetStumbler.com Forums - Post Your Homebrew Antennas Here Date: Mon, 22 Jul 2002 11:53:54 -0700 MIME-Version: 1.0 Content-Type: multipart/related; boundary="----=_NextPart_000_0023_01C23176.74159A70"; type="text/html" X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 This is a multi-part message in MIME format. ------=_NextPart_000_0023_01C23176.74159A70 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Location: http://forums.netstumbler.com/showthread.php?s=&postid=20618 NetStumbler.com Forums - Post Your Homebrew Antennas = Here

=20 =20

=20

=20

=20

NetStumbler.com=20 Forums > Wireless=20 Forums > Hardware=20 > Post Your Homebrew Antennas = Here

Pages (4): = =AB=20 First ... =AB=20 2=20 3=20 [4] Go=20 to first unread post

=20 Last=20 Thread Next=20 Thread 3D""=20

=20

Author

Thread

IL GIACO
Mini=20 Stumbler
3D""=20

Registered: May=20 2002
Location:
Posts: 11

quote:
Originally posted by gcrocker
Not that = appropriate for stumbling (it's 20" high and PVC), = but I may=20 mount it on my crappy van for just that = purpose:

http://www.guerrilla.net/reference/antennas/2ghz_collinea= r_omni_lowpwr/

First=20 one I built took 3-4 hours, and I think I can do = them in an=20 hour now.

-glenn

I've build many of them and I = can=20 made it in about 30 mins 3D""=20

However I've also build one with 5 = elements=20 instead 4 (usings the same measures) and It work very=20 good.
For example with airtools from the same = distance from=20 a AP the 4 elements antenna give me 10 of quality, = with the 5=20 elements I can reach 17-18 of quality.
I would like = to know=20 what a 6 elements could do.
PS: I've used rg174, = it's the=20 only cable that I can find here.

__________________
Se puo' essere ferito, = puo'=20 essere ucciso!

Report=20 this post to a moderator | IP: Logged

07-09-2002 03:06=20 AM

=20 =

fordem
Tropical=20 Stumbler

Registered: Apr=20 2002
Location:
Posts: 186

quote:
Originally posted by lincomatic
i = wonder if=20 it's ok the way i did the feed connections instead = of using=20 rigid coax. performance seems ok but i don't know = what my=20 setup does to VSWR...

I don't have access to the rigid = coax=20 either, but after seeing the Trevor Marshall site = which showed=20 one version built directly on the back of an N = connector, I=20 decided to go ahead with it.

I just need to get = a round=20 tuit - as soon as I find one of those 3D""=20

I also got a friend with one of = them little=20 DBS dishes that she's not using, so maybe I'll even = try it=20 with that.

Report=20 this post to a moderator | IP: Logged

07-09-2002 04:29=20 AM

=20 =

sparafina
yatta!

Registered: May=20 2002
Location: AUS/LAX/DCA
Posts: = 168

Re: = success!=20

[QUOTE]Originally posted by lincomatic=20
[B]ok, where best to get the answer than from = the=20 horse's mouth. i fired off an e-mail to trevor, and he = gave me=20 the answer quickly. my driven element & feed were = all=20 wrong. in the pic above, i have the thing shorted = together at=20 the middle and only 1 wire connecting. it's supposed = to be=20 open at the middle w/ 1 side connected to the signal = and one=20 to ground. works great now...about 3dB better than my = comtelco=20 patch clone when both are rotated for best match of=20 polarization.[QUOTE]

Nice job! I also got it = wrong, I=20 only connected the driven element. I also used a piece = of=20 stiff thinnet, which is a bit floppy and a pain to = work with=20 at 15 mm. I need to redo the driven element with just=20 wire.

Report=20 this post to a moderator | IP: Logged

07-09-2002 12:33=20 PM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

Re: Re:=20 success!=20

quote:
Originally posted by sparafina
Nice job! = I also=20 got it wrong, I only connected the driven element. I = also=20 used a piece of stiff thinnet, which is a bit floppy = and a=20 pain to work with at 15 mm. I need to redo the = driven=20 element with just wire. [/B]

it'll make a dramatic diff in the = performance when u hook it up the right = way.

btw, i=20 asked trevor about the double 8 design:
http://patraswireless.00sf.com/antennas.html
and=20 he claims that 2 loops gives best performance.

__________________
~lincomatic

Report=20 this post to a moderator | IP: Logged

07-09-2002 02:40=20 PM

=20 =

fordem
Tropical=20 Stumbler

Registered: Apr=20 2002
Location:
Posts: 186

Re: Re: Re: = success!=20

quote:
Originally posted by lincomatic
btw, i = asked=20 trevor about the double 8 design:
http://patraswireless.00sf.com/antennas.html
and=20 he claims that 2 loops gives best performance.

How are you counting loops? I'm = thinking=20 that 2 loops would be a bi-quad.

Report=20 this post to a moderator | IP: Logged

07-09-2002 02:51=20 PM

=20 =

sparafina
yatta!

Registered: May=20 2002
Location: AUS/LAX/DCA
Posts: = 168

Re: Re: Re: = success!=20

quote:
Originally posted by lincomatic
he = claims that=20 2 loops gives best performance.

I've also seen another design = with all=20 the loops stacked in parallel, I'll post the url when = I find=20 it again. It was based on a ham radio antenna design = scaled to=20 2.4 gHz

Report=20 this post to a moderator | IP: Logged

07-09-2002 02:59=20 PM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

Re: Re: Re: = Re:=20 success!=20

quote:
Originally posted by fordem =

How are=20 you counting loops? I'm thinking that 2 loops would = be a=20 bi-quad.

yes, i'm calling a biquad 2 = loops. trevor=20 told me that the biquad is optimal...adding more loops = doesn't=20 enhance performance. so where do these double 8 = designs come=20 from?

__________________
~lincomatic

Report=20 this post to a moderator | IP: Logged

07-09-2002 03:31=20 PM

=20 =

fordem
Tropical=20 Stumbler

Registered: Apr=20 2002
Location:
Posts: 186

Re: Re: Re: = Re: Re:=20 success!=20

quote:
Originally posted by lincomatic
so = where do=20 these double 8 designs come from?

I wondered about that myself - I = spent=20 some time researching quad, bi-quad and the good old = cubical=20 quad and doodling on a scrap of paper and theory says = the=20 voltage at the center point of a double 8 will be = pretty much=20 zero at the resonant frequency.

It's possible = that a 6=20 loop version might work reasonably well, but so far = that's=20 only on paper.

quote:
Originally posted by sparafina
I've = also seen=20 another design with all the loops stacked in = parallel, I'll=20 post the url when I find it again. It was based on a = ham=20 radio antenna design scaled to 2.4 gHz

Is this it?

The Frars guys = had a very=20 brief snippet of information on their webpage about a=20 quad-bi-quad and had promised a writeup that never=20 materialised.

What I had envisioned was four = bi-quads=20 mounted in a single plane equidistant from one long=20 rectangular reflector.

Report=20 this post to a moderator | IP: Logged

07-09-2002 04:00=20 PM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

seattlewireless has posted a new version of = the coffee=20 can waveguide calculator:

http://seattlewireless.net/index.cgi/DirectionalWaveguide=

__________________
~lincomatic

Report=20 this post to a moderator | IP: Logged

07-09-2002 09:21=20 PM

=20 =

sparafina
yatta!

Registered: May=20 2002
Location: AUS/LAX/DCA
Posts: = 168

quote:
Originally posted by lincomatic=20
seattlewireless has posted a new version = of the=20 coffee can waveguide calculator:

http://seattlewireless.net/index.cgi/DirectionalWaveguide= =20

The calculations seem to have = changed=20 with respect to waveguide length, am I = wrong?

They also=20 have links to Christoph Cattelain's antenna site below = the=20 calculator. Ill try and visit him in September when = I'm Liege.=20 I'm particularly interested in his slotted waveguide=20 experiments.

Report=20 this post to a moderator | IP: Logged

07-09-2002 09:48=20 PM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

quote:
Originally posted by sparafina

The = calculations seem to have changed with respect to = waveguide=20 length, am I wrong?

if you put in channel 6, the = numbers are=20 very similar to greg rehm's calcs (at least for = 3.25"). i=20 looked into greg's calculator and he's using channel 6 = (2437)=20 as the reference freq.

__________________
~lincomatic

Last edited by=20 lincomatic on 07-10-2002 at 09:33 AM

Report=20 this post to a moderator | IP: Logged

07-10-2002 08:58=20 AM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

as a=20 crosscheck, i decided to re-implement the cantenna = calculator=20 based on formulae from the ARRL antenna book.
in = case=20 anyone is interested, i've uploaded C++ source and = Win32 EXE=20 at:

http://www.geocities.com/lincomatic/software.html#cantenn= a

seems=20 like the seattlewireless calculator is a little bit = off. my=20 calcs are almost exactly identical to greg rehm's = calculator=20 when you plug in channel 6.

found some = interesting=20 stuff in running my program. w/ a 3.25" (82.55mm) dia=20 waveguide optimal probe position varies by a whopping = 5.5mm=20 between channels 1-11! but with a 92mm dia guide, the = diff is=20 down to 2.63mm. can't go much higher than 92mm because = the=20 TM01 cutoff starts to get too low to do channel = 11.

btw=20 the ARRL book recommends 2-3 guide wavelength for the=20 waveguide length, so the 3/4 guide wavelength value = should be=20 treated as a minimum.

__________________
~lincomatic

Last edited by=20 lincomatic on 07-10-2002 at 11:47 AM

Report=20 this post to a moderator | IP: Logged

07-10-2002 11:44=20 AM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

cantenna finally beats the = biquad=20

i've been=20 trying to hook up my brother who is a professor at a = local=20 college into the campus wireless network. he lives in = college=20 housing just down the hill from the campus. the IT = department=20 has several antennas pointed at college buildings off = campus,=20 but unfortunately, my brother's house is around the = corner=20 from them. they were supposed to install another AP = near my=20 brother's house, but haven't gotten around to it after = a year=20 of waiting, so i've been trying to figure out how to = get a=20 signal to him.

a couple of nites ago climbed up = on his=20 roof and pointed my biquad at a yagi (which was = pointed away=20 from me) a few hundred feet away which was behind a = tree and=20 amazingly still got a 5dB SNR...marginal at best = especially=20 because the tree may grow yet more foliage.

i = built=20 another cantenna tonite using 3 3.25" cans...makes the = length=20 about 1.75 waveguide wavelengths. the driven element = is=20 30.75mm long and mounted the 64mm from the back of the = can.=20 the conical collector is 7.25" in diameter on the big = side, w/=20 a 30degree flare. this is just a quick prototype so i = made the=20 collector out of 2 coat-hanger circles, separated w/ = four 4"=20 long coat hanger supports covered in aluminum foil. = the final=20 design will need to be more durable to stand up = against wind=20 & hail. here are the SNR's i got across the street = from my=20 AP.

orinoco built in: 26
2 cans w/o = collector:=20 36
trevor marshall biquad: 39
3cans w/ = collector: 43=20 (!)

this is the highest gain antenna i've built = yet. i=20 hope it's good enough to get a clean signal at my = bro's house.=20 can't wait to try it out.

dragged my jornada = off a=20 table while connected to this %$* thing and it fell on = the=20 floor. lucky the card & jornada are ok, but i = broke off=20 the end of my pigtail 3D""=20 3D""=20 3D""=20

lincomatic has = attached=20 this image:
3D""

__________________
~lincomatic

Last edited by=20 lincomatic on 07-11-2002 at 08:52 AM

Report=20 this post to a moderator | IP: Logged

07-11-2002 03:45=20 AM

=20 =

ew123
Mini=20 Stumbler
3D""=20

Registered: Apr=20 2002
Location:
Posts: 8

Hello=20 Lincomatic,

your new cantenna is surely = impressive and=20 very usefull for sniffing with kismet but for the = purpose of=20 connecting your brother to the campus it might be not=20 ideal.

The horn collects more energy and = improves=20 reception for this antenna. However when used for = sending, the=20 horn spreads the TX-signal over a larger area thus = decreasing=20 the received power at the other end.

So you = will=20 improve reception of the campus signals while = simultaneously=20 decreasing reception of your brothers signal at the=20 campus.

Such an antenna is only usable for two = way=20 communication when used on both ends. Then you have = the=20 gain-gain (or how would you call improved gain? ;-) = for=20 receiving as well as for transmitting.

Report=20 this post to a moderator | IP: Logged

07-11-2002 04:52=20 AM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

quote:
Originally posted by ew123

The = horn=20 collects more energy and improves reception for this = antenna. However when used for sending, the horn = spreads the=20 TX-signal over a larger area thus decreasing the = received=20 power at the other end.

thanks for the info. = bmoore314=20 made the same observation with his wire mesh = collector, but i=20 was hoping that by keeping mine smaller (his has a 11" = aperture and mine is 7.25")
that my Tx power would = stay=20 high enough.

in case it doesn't work, what kind = of=20 antenna would you recommend? i can't find any decent = designs=20 w/ enough gain ... may resort to buying a parabolic = grid=20 antenna if all else fails.

__________________
~lincomatic

Report=20 this post to a moderator | IP: Logged

07-11-2002 08:51=20 AM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

It works!=20

just got back=20 from my bro's. from the same place i got the 5dB SNR = using the=20 biquad, i got 8dB w/ the 3-can & cone cantenna. i = fired up=20 pocketIE and was able to get into hotmail with it and = surf a=20 little. by moving to a different spot on the roof that = wasn't=20 blocked by the tree, i got a 12dB SNR. i'm still = amazed=20 because i'm pointing it at the Lucent yagi from about = a 30-45=20 degree angle from the rear....didn't realize they = radiated=20 that strongly to the rear.

my question is do i = have=20 enough signal margin to operate this thing reliably? = what's a=20 good minimum SNR for reliable communications? i don't = want to=20 drop the connection every time the atmospheric = conditions=20 change.

__________________
~lincomatic

Last edited by=20 lincomatic on 07-11-2002 at 02:59 PM

Report=20 this post to a moderator | IP: Logged

07-11-2002 10:48=20 AM

=20 =

Thorn
DooMed=20 Stumbler

Registered: Apr=20 2002
Location: T=EDr na n=D3g
Posts: = 384

Kewl.=20

lincomatic

12dB should be good. Just = for grins=20 though, what are the actual figures for the noise = level,=20 signal, SNR?

__________________
Thorn

Report=20 this post to a moderator | IP: Logged

07-11-2002 10:59=20 AM

=20 =

lincomatic
UberNetbumbler

Registered: Apr=20 2002
Location: Venus
Posts: 880

Re: = Kewl.=20

quote:
Originally posted by Thorn=20
lincomatic

12dB should be good. = Just for=20 grins though, what are the actual figures for the = noise=20 level, signal, SNR?

thorn,
12dB was the most = stable SNR=20 reading i got...the antenna pretty directional (though = less so=20 than the bare cans)...hard to keep it steady holding = the=20 bazooka & jornada in 1 hand and the stylus in the = other=20 whilst balancing on a pitched roof! looking back into = my MS=20 log, here are the peak values it = recorded:

Signal+:=20 -85
Noise-: -103
SNR+: 14

i'm also trying = to=20 allow for signal loss in the cabling & connectors = to the=20 eventual location inside the house. how low do you = think it=20 can let the SNR go at the connection to orinoco card = in the=20 house and keep a stable connection? the AP is also an = Agere=20 unit.

__________________
~lincomatic

Report=20 this post to a moderator | IP: Logged