The "Cantenna" has not 12 dBi!!!

snipped-for-privacy@gmx.de hath wroth:

Show me. I do designs and models.

The final gain really depends heavily on the waveguide geometry. The typical coffee can will do about 9dBi. Pringles cans are about 7dBi. Extra long extensions with horns can be as high as 14dBi.

I show 8.5 to 9.8 dBi gain:

Note that the gain varies with frequency:

The NEC model is directly stolen from Trevor Marshall's at:

That's also without any additional losses such as surface conduction losses, connector loss, coax loss, etc. The final antenna at the end of a few feet of lossy coax cable will probably be about 2-3dB less.

If you paid for 12dBi and are only getting 8dBi, ask the vendor for your decibels back.

If you want to do something useful, figure out what I did wrong with the coffee can design speadsheet. It's close, but not perfect.

Dear Jeff,

I have seen your cantenne model. But where is your NEC-file for the cantenna? I will show you, where you have model errors. Did you ever performed an average gain test? If not, you should do this allways first. Also all warnings in your model should be removed or checked! Otherwise, you will get wrong results. Especially, you should not have surface patches (tin can) nearby your feed point!

Regards, Aziz

Hell Jeff,

I have found your excel sheet and did the model calculation. Your feed have to be optimized. Also, some parts (surface patches) nearby the feed, must be removed.

My model with surface patches: Tin can diameter: 100 mm Length: 135 mm Lg/4: 44.452 mm My model gets round about 7.7 dBi with the same dimensions. Calculated total gain is 8.3 dBi, average gain test correction is 0.6 dBi. 8.3 -

0.6 = 7.7 dBi.

Model with wire grid (more accurate): Total gain is 8.4 dBi without error (average gain test=1.0 (-0 dBi correcting factor)).

Of course, with 198 mm diameter flare for the cantenna, you get 12 dBi gain. But the original cantenna can not be have 12 dBi!!!! It is a cheap marketing issue.

Aziz

snipped-for-privacy@gmx.de hath wroth:

Sorry. My Jalbum 7.1 photo album software buried the NEC2 files. Some of the various NEC2 files are:

'll make them visible when I figure out how.

Sure. I'm interested. However, I would prefer you find whatever we did wrong with the Excel speadsheet which can generate the NEC file.

Yep. It fails with a value of about 2.0. 1.0 is perfect. I haven't chased down the cause. I've made it worse, but not better. L.B. Cebik offered some sage advice from about 2.5 years ago on the model, which none of the perpetrators (including me) have done anything about:

You may need to register to view this. Please ask if you want me to email the message.

I did, but couldn't determine how to fix it. Believe me, I tried. I think the problem is fairly universal because all of the coffee can models I've found on the web have exactly the same problem. The model is even included as a sample with several packages. Despite the errors, the field test results closely resembles the model.

4NEC2 shows:

etc...

That's from the overlapping wires at the base of the conical feed. It was too much work to try to get them each positioned so there's no overlap. I'll replace the cone with a single wire feed (as in the spreadsheet), which gets rid of that error but narrows the bandwidth.

Some warnings can be safely ignored.

Yep. It is quite close, but not touching. I'll try punching a hole in the side of the can, under the conical feed, and see if that helps. However, not for a few days. I've got too many projects going and this one is going to take some real work.

The construction of a waveguide requires a minimum amount of understanding of E and H fields. In order to realize the maximum gain quite few principles must be gotten right..My opinion, "it's not for everybody.."

Hello Jeff,

I have sent you some cantenna models via email. They are very long, I couldnt post it here.

Here is the source code of my cantenna model (not 4nec2/nec2/nec4 compatible!): As I mentioned in email, the source antenna model have to be compiled.

--------------------------------- CM Cantenna WLAN Antenna Model at 2.437 GHz CM (C) 2007 by Aziz Oeguet CM File: Cantenna2G4Hz.txt CE

#include "Common.txt" #include "PP4NEC2Defs.txt"

ofsx =3D -0.05 // Offset x-Achsenverschiebung (zur besseren 3D- Darstellung)

//--- Kontrollschalter (Switches) f=FCr Modellkonfiguration --- #ifndef CANTENNA_TRICHTER CANTENNA_TRICHTER =3D 1 // Trichter-Blende (0=3Dkein, 1=3Dmit) #endif

#ifndef CANTENNA_WRMODE CANTENNA_WRMODE =3D 1 // Drahtmodus (0=3DSurface-Path, 1=3DWire-Modus) #endif

#ifndef CANTENNA_CALC CANTENNA_CALC =3D 0 // Cantenna parameter berechnen (1=3Dberechnen,

0=3Dauf manuellen Wert lassen) #endif

#pragma symode =3D _SYMODE_REUSE

//--- Feeder(Erreger)-Parameter (Lambda/4-Element) --- Feed_Durchmesser =3D 0.0015 // Antennendrahtdurchmesser Feed_Rad =3D Feed_Durchmesser/2 // Antennendrahtradius Feed_nSeg =3D 9 // Anzahl Segmente des Antennendrahtes Feed_Tagnr =3D 1000 // Draht Tagnummer-Start Feed_Segnr =3D 1 // Segment-Number des Erregerst=FCcks

#if !CANTENNA_CALC Feed_Len =3D 0.0308 // Antennendrahtl=E4nge #else Feed_Len =3D lambda/4 // Antennendrahtl=E4nge (Wellenl=E4nge/4) #endif

Feed_Delta =3D Feed_Len/Feed_nSeg // Segment-L=E4nge (Antennendraht)

//--- Cantenna-Parameter (Tubus) --- TubusnSegLen =3D 15 // Anzahl der Segmente in Tubus-L=E4nge TubusnSegRad =3D 6 // Anzahl der Segmente im Tubus-Radius TubusnSegMantel =3D 32 // Anzahl der Segmente im Tubus-Mantel

TrichternSegLen =3D 12 // Anzahl der Segmente in Trichter-L=E4nge TrichternSegMantel =3D 32 // Anzahl der Segmente im Trichter-Mantel

TubusDurchmesser =3D 0.1 // Tubus-Durchmesser TrichterDurchmesser =3D 0.198 // Trichter-Durchmesser TrichterLen =3D 0.086 // Trichter-L=E4nge

TubusRadius =3D TubusDurchmesser/2 // Tubus-Radius TrichterRadius =3D TrichterDurchmesser/2 // Trichter-Radius

#if !CANTENNA_CALC //--- Cantenna Parameter manuell --- TubusLen =3D 0.135 // Tubus-L=E4nge Feed_Dist =3D 0.044452 // Erregerabstand zu hinteren Wand #else //--- Cantenna Parameter berechnen --- DLc =3D 1.706*TubusDurchmesser DLg =3D 1/sqr((1/(lambda*lambda))-(1/(DLc*DLc))) TubusLen =3D DLg*3/4 // Tubus-L=E4nge Feed_Dist =3D DLg/4 // Erregerabstand zu hinteren Wand #endif

Feedzpos =3D -TubusRadius // Erregerfusspunkt (unten)

//--- Feeder --- // Erregerspeisedraht GW Feed_Tagnr Feed_nSeg 0 0 Feedzpos-Feed_Delta/2 0 0 Feedzpos

+Feed_Len-Feed_Delta/2 Feed_Rad

// winziger Anschlussstummel unten (NEC2 Warnung umgehen) GW Feed_Tagnr+1 1 0 0 Feedzpos-Feed_Delta/2 0 0 Feedzpos-Feed_Delta/

2-Feed_Delta Feed_Rad

p1x =3D Feed_Dist-0.04 p1y =3D -0.03 p1z =3D-0.03 p2x =3D Feed_Dist+0.11 p2y =3D 0.05 p2z =3D 0.03 rad =3D 0.04

//--- Hotspot (am Erreger in der N=E4he keine Objekte generieren) --- #Hotspot point=3D_HS_OUTSIDE Feed_Dist 0 Feedzpos 0.005

// F=FCr hintere Wand des Tubus n=F6tig, da Vierecke im Mittelpunkt zu Dreiecken werden. #pragma limits=3D1, minwirelen=3D0.001, minareasize=3D0.000001

//--- Tubus generieren ---

#if CANTENNA_WRMODE // Parameter f=FCr Wire-Modus #pragma mode=3D _MODE_WR, wrnumseg =3D 1, wrtagstart=3D2000, wrtaginc =3D 1, wrradius =3D Feed_Rad #else #pragma mode=3D _MODE_SP #endif

//--- Tubus um Erregerabstand nach hinten verschieben ---

#Move -Feed_Dist 0 0

w1 =3D 180 w2 =3D w1+360

// Tubus generieren #Pipe TubusnSegMantel TubusnSegLen 0 0 0 w1 w2 TubusRadius TubusRadius TubusRadius TubusRadius TubusLen _R1|_R2|_A1

// Hintere Wand (Ring) generieren #SPArc TubusnSegMantel TubusnSegRad 0 0 0 w1 w2 0 0 TubusRadius TubusRadius _A1

// Trichter generieren #if CANTENNA_TRICHTER #Pipe TrichternSegMantel TrichternSegLen TubusLen 0 0 w1 w2 TubusRadius TubusRadius TrichterRadius TrichterRadius TrichterLen _R2| _A1 #endif

#Transform // Transformation abschliessen (Move) //--- Ende Tubus-Verschiebung ---

// Achsen-Transformation (um aus dem Nahfeld rauszukommen) #if !CANTENNA_TRICHTER GM 0 0 0 0 0 -TubusLen+Feed_Dist+ofsx 0 0 0 #else GM 0 0 0 0 0 -TubusLen+Feed_Dist-TrichterLen+ofsx 0 0 0 #endif

// End-Of-Geometry GE 0

//--- Wire-Load (HF-Speisung) --- EX 0 Feed_Tagnr Feed_Segnr 0 1.0 0.0 LD 5 0 0 0 62900000

// Extended-Wire-Kernel einschalten EK 1

//--- Frequency Parameter and Execute --- FR 0 1 0 0 freq 1 RP 0 91 181 1001 -180 0 2 2 EN

-----------------------END-------------------

cantenna with maximum 24dBi? --> 'WLAN-SAT-FORUM Forums-viewtopic-Ultimate wifi antenna TinCan Enhancers - part-2 ...'

------------------------------------------------------------------------ View this thread:

fablept hath wroth:

This is not exactly a cantenna, but a horn antenna uses the same principle. 16dBi gain:

How to grind the numbers for a horn antenna:

The cardboard and aluminum foil horn antenna:

SETI "horn of plenty"

gain at 1500MHz. My guess(tm) is about 24dBi at 2.4GHz. Note the size of the horn required to get that gain as compared to the Slovanian cantenna.

it´s some kind of "booster" to a cantenna...i didint get 24dB, but around 17dB..it´s not bad for very low-cost antenna.

I use this freeware for horn antennas / conical horns /parabolic /lens, its great because gives a template according to the frequency we want. i import the postscript file to corel draw.

hdl_3b4 --> 'Download hdl_3b4.zip - DepositFiles.com - the best upload and file sharing service!'

------------------------------------------------------------------------ View this thread:

fablept hath wroth:

I've never heard of a booster.

The download site is full. I suggest you get it from the authors web site at:

(see other programs) It is a nifty program for designing horns, but it's not terribly interactive or educational. For example, you have to know in advance the size of 2.4GHz WR-340 waveguide. Hint:

Although this dish feed horn design spreadsheet doesn't generate sheet metal templates, it's much more useful:

More of the same:

a booter is what i call a c>

great sites, thanks!

'Wi-Fi. Krasnodar. Horn antenna. Ðóïîðíûå àíòåííû.'

--> here it is a rectangular Horn antenna for 2.4ghz with 33dB, it´s not very pratical to use...but it´s impressive.

at this moment i´m using Patch microstrip antennas, very simple to build...and give good results, the problem in building PCB patch antennas is to find the correct type of PCB due to his dielectric values, i have a 14dBi in a bad bakelite-PCB that works great, but in other antennas dont work very well...maybe FR-4 PCB would have better results.

------------------------------------------------------------------------ View this thread:

fablept hath wroth:

Nice. There was also some ham radio 2.4GHz horns with a similar case of elephantitis. It's actually not a horrible as it looks and is very easy to fabricate. It's also not a critical in construction and tuning as a parabolic dish. However, a horn is big and ugly.

Bakelite? I think you mean phenolic, which is the light brown paper and glue cheap junk that absorbs moisture like a sponge and is very lossy at 2.4GHz. Even G10/FR4 is better than phenolic. However, if you want quality, look into air dielectric (standoffs), polysulfone, or PTFE (RTDuriod).

check out the FA-20 antenna >

i know its terrible for 2.4ghz...but yet it´s the cheapest PCB i can get.

i know this is getting very offtopic, but i have this link for a omni PCB antenna: 'David Kittle's Collinear Antenna'

i´m trying to understand the design to use in different PCB´s..

i think the omni is made according to this:

-for the 57mm section= (lambda/2)*copper velocity factor

-for the 27mm section= (lambda/4)*copper velocity factor

-for the microstrip width i just use a simple software like AppCAD to calculate a 50ohm microstrip line according to the PCB specifications, right?

-regarding about feeding the antenna, a balun to connect the two sections of the antenna? in the link nothing is mentioned how to feed the antenna.

I appreciate any help...

------------------------------------------------------------------------ View this thread:

fablept hath wroth:

It's crap. G10/FR4 is cheap and commonly available.

Correct. Although there's a bit of a question of from where to where to measure. Hmmmm.... No NEC cards included so I can't run his numbers through 4NEC2. The first paragraph on the page is speculative baloney.

Nope. The traces between the feed point and the beginning of the dipole elements might be 50 ohms (they're not, but close enough). However, the other elements a fairly high impedance and should be quite a bit thinner than shown.

Yep, 1:1 balun is needed. That can be made from coax cable. The "extra" two quarter wave sections between the feed point and the dipole is rather questionable and mostly useless.

What are you trying to accomplish? Duz it have anything to do with the original question?

fablept hath wroth:

Here's another large 2.4GHz horn antenna used in the 2003 WiFi long distance (DX) contest:

(Near bottom of page). Made with a metal tubing frame and covered with aluminium window screen wire mesh.

how can i f>

sorry, this has nothing to do with the original topic of cantenna..

i which to build that omni PCB, but i only recently started at RF..so i dont have enough knowledge to make the antenna.

that´s a very big horn, how much dBi that antenna will have?

the world record wireless distance is 237.36miles, 'link'

------------------------------------------------------------------------ View this thread:

fablept hath wroth:

Lots more under stripline and microstrip.

Well, ok. Low gain antennas are easy enough. High gains antennas are difficult.

Dunno. I need dimensions to run a model. It's late. I'm tired. Maybe later...

Amazing.