originally written by stygen some time in or before 2002, and published at http://members.iinet.net.au/~stygen/Quad.html and in the original WaFreeNet wiki;
content recovered via the wayback machine.

2.4GHz Cubic Quad Antenna


Introduction

The Cubic Quad antenna is a commonly homemade antenna in the range of about 150 odd MHz. Our little project was to design one of these for use in the 2.4GHz range for 802.11 wireless LANs. The reason these are seldomly used for 2.4GHz is the size. The picture below is a 4 element cubic quad for the 147MHz range. Large isn't it. The one we are going to build for 2.4GHz will only be 6cm long!

quadvan1.jpg

Before We Start

I'm not even going to pretend to know how all this stuff works. So I won't be explaining it to you either. The following sites may help those of you who want to be 'in the know'. I don't care.

I made my design by taking bits and pieces from the last two links.

The Design

I scratched together an initial plan on how I was going to set about putting this together. The measurements came from the second (or third) link above. While each element was made the same as in the design, the support structure was changed to a much easier one. This was about the only advantage of building a really small antenna.

Materials

1 hot glue gun
1 soldering iron
1 soldering god (enter ChrisK)
short length of coax with connector
60cm of builders wiring (stripped to get one solid copper wire ~ 2mm thick)
3 cotton buds (hehe I'll get to that bit later)

Construction

After stripping the copper wire, we constructed the four elements as per the measurements I pilfered from the java application on the previously mentioned page. We bent the wire with a pair of pliars against a small anvil. The reflector and director elements were soldered closed by ChrisK, the driven element left open for connection to the coax.

quad-materials.jpg

From left to right.... Reflector, Driven element, Director element 1, Director element 2 The white sticks are cotton buds with the cotton crudely removed.

Each element differs in size from the next. From the reflector through to director 2, the sides of the squares get smaller by only ~0.1mm. Human error can really screw this up. As this is only really a prototype we are not overly concerned. However, when it comes to building the real deal, we have decided that getting a computer driven robot to cut out some copper on a fibreglass board with some precision in length and squareness would be a goer.

The next step is to solder the driven element to a nice thick and chunky bit of LMR400 :) We did this on an angle to prevent the 'direct' short.

quad-driven.jpg

Here lies problem number two. The space created by the gap between coax core and outer is huge in comparison to the size of the element. We decided that keeping the length of wire for the element was more important than the shape, so it is also not really square anymore...prototype. This would also ideally feed into a balun rather than directly onto the coax. We just need to figure out how.

We then built the rest of the elements onto the driven element with the assistance of a hot glue gun and some cotton buds. When you put cotton buds in the microwave for one minute next to a glass of water, they do not get hot. Ideal antenna construction material! The elements were distanced according to the java application. However, it should be noted that increasing the distance between the elements will increase gain at the expense of bandwidth. The final version will hopefully be totally adjustable for tuning.

We used three cotton buds and the hot glue gun to hold it all together. It is messy... prototype... but it is also very small. Hehe. You can still see the leftover cotton wool on the ends of the sticks ;-)

quad-size.jpg

OK. So once all done, we did some very quick testing.

It worked! We didnt keep logs of the test as we intend to do it properly soon, but it gave a dramatic increase in signal, S/N and reduced noise. I will add test results to this page when they are ready.

Here is a closer look at the prototype.

quad-cqprotoype.jpg

Test Results

Two laptops with wireless cards were moved apart at such a distance where signal could be improved. One of the laptops was then given a balaxy dish (a galaxy dish that ChrisK modded to have a different balun and dipole). The balaxy dish was then replaces with the prototype cubic quad. Results were logged and the peak of all results were as follows;

Rx

Tx

Signal

Noise

SNR

Signal

Noise

SNR

Internal

-78

-97

20

-78

-94

17

Galaxy

-61

-99

38

-61

-94

34

Cubic

-70

-99

28

-70

-93

25

I am very encouraged by these results. The prototype cubic quad was a complete bodge job with very little precision. More precise elements may give better results. It was not adjustable due to the hotglue used to stick everything together. With tuning these results may be better. And there was no balun used, on account of my having to figure out how to make a balun for this little beasty. A balun would hopefully give me another 3db.

Future Directions

The elements need to be more precise. Having them properly machined would be ideal. The support structures should be threaded. This will allow us to put plastic washers at each bend with some plastic nuts, giving us the ability to tune it for maximum gain/bandwidth.

A balun is required (perhaps). The signal is skewed about 15 degrees to the right (guestimate). We also need to figure out how to design the connection to the balun/coax in such a manner that will cause the least hassle to the shape and length of the driven element.

v1.2 is under way. v1.1 was scrapped before I put it together because I am still unhappy with the elements. We have some good ideas on where to go from here, so watch this space for developments over the next week or so.

Contact:

If you have any constructive comments, and especially if you could comment on the test results please give me a yell at;

justin at wantree dot com dot au

I welcome all constructive and informative feedback.

Kindly hosted by DarkNet - For all your hosting needs