Wlan Antennas

(Colinears, Troughs and Dishes)

Directional or Omnidirectional, this will be one of your easiest choices. If the signal needs to be broadcast to everyone in the area you require an omnidirectional type. (circles in a pond). If it is to be focused to another receiver only, you will require a yagi, dish, etc. So "Verticals" are the ticket to broadcast to everyone in a local network.

What about polarization? - The radiating field has an electric and magnetic component which exist because of each other. The polarization of this field is a term used to describe at what plane the electric field exists in. (it's maximum field). The coaxial Collinear has a vertical polarization, this polarization is generally the easiest method to obtain omnidirectional radiation, and it means the electric field is in the vertical plane. Free space can be considered as a "transmission line" with an impedance of 377 ohms. ( E (electric component) / H (magnetic component) = Z (impedance)). Antennas can be considered as transformers which couple our transceiver's transmission line to the "space" transmission line, correct impedance matching is important, as is correct field alignment for maximum efficiency. A vertically polarized antenna will loose 3 dB of signal (1/2 of it) when received with a horizontally polarized antenna, (and vica versa), and a circularly polarized antenna will be received by a horizontal or vertical antenna equally. (averaged signal) . Terrain can also modify polarization over long distances, and generally the electric field is used in the horizontal plane for long distance communications with such antennas as "yagi's and parabolic dishes".

The choice of vertical will largely depend on a number of factors, performance, windage, EMP immunity, cost, ease of fabrication, life-span, cost, environmental aesthetics, cost and so on. We could choose from the ground planed verticals like a "rubber duck", 1/4 wave, 5/8, 1/2 wave, (or to a full vertical dipole), to phased verticals such as collinear or phased arrays employing an element type, (eg. stacked dipoles), or some of the variations such as "J" poles. Without a doubt a 1/4 wave vertical would be cheap and easy to whip up, would have a long life, but unfortunately the trade-off is low performance. So how much gain does the antenna need?? Some examples below are given as a guide (somebody please check my maths. . . it's been a while, and please disregard the "top" diagram below)

A Wlan example (see second figure above) - Suppose we have a wlan card which output 18 dBm and will run to their maximum speed with -80dBm receiver signal input, each have 10 meters of LMR400 coax feeding an 8 element vertical coaxial collinear antenna, through 2 connectors (say another 0.2dB loss each). Neglecting the loss in the pigtail that feeds the LMR from the card, what is the maximum distance they could transmit over without data speed loss.

Coax loss in LMR400 is 0.22dB / metre at the rated frequency, so a 10 metre transmission line has 2.2 dB loss. 2441 Mhz was chosen because it will be the centre frequency of the Wlan cards range.

The problem is (from TX card to Rx ) . . .18 dBm - 0.2 - 2.2 - 0.2 + 6.35 - path loss + 6.35 - 0.2 - 2.2 - 0.2 = - 80 dBm
25.5 - path loss = -80 dBm
dB now applied to counter path loss = 105.5 dB, now working from the path loss equation
105.5 = 100.25 + 20 log d
5.25 = 20 log d
d = 1.83 Km

Now suppose we disconnected the homebrew antenna systems and replace the stock (rubber duck) antennas that come with the card. Assume the antenna has about 1 dBi and the card connector about 0.2 dB insertion loss.

The problem is (from TX card to Rx ) . . .18 dBm - 0.2 + 1 - path loss + 1 - 0.2 = - 80 dBm
19.6 db - path loss = -80 dBm
dB now applied to counter path loss is 99.6 dB
99.6 = 100.25 + 20 log d
-0.65 = 20 log d
d = 0.93 Km . . . . . . . ( incidently if the antennas were 0 gain, connectors had no losses then d = 97 metres)

Now lets use some better antennas (vertical 16 element collinear), can't do much about the connectors (except they are the lowest insertion loss type you can get), or the feeder. The numbers go like this. . .

(from TX card to Rx ) . . .18 dBm - 0.2 - 2.2 - 0.2 + 10.55 - path loss + 10.55 - 0.2 - 2.2 - 0.2 = - 80 dBm
33.9 - path loss = -80 dBm
dB now applied to counter path loss = 113.9 dB, now working from the path loss equation
113.9 = 100.25 + 20 log d
13.65 = 20 log d
d = 4.8 Km . . . . . . . . (Dats better . . . . . .)

Now lets use a 1 Watt bi-direction amplifier, some large...ish high gain dish antennas (22 dBi) and the same transmission line and connector set up. .
(dishes are directional, point to point, not omnidirectional).

(from TX card to Rx ) . . .30 dBm - 0.2 - 0.2 - 2.2 + 22 - path loss + 22 - 2.2 - 0.2 - 0.2 = - 80 dBm
68.8 - path loss = -80 dBm
dB now applied to counter path loss = 148.8 dB, now working from the path loss equation
148.8 = 100.25 + 20 log d
48.55 = 20 log d
d = 267.6 Km . . . . . . . . (Demonstrating the distance achievable when a "focused" antenna system is employed for a point to point communication system and also that a brick can fly if a big enough engine is strapped to it.)

It must be remembered that these examples are "line of sight", which means no obstacles are in the way and it is not beyond the horizon.

Wlan parameters and problems

Australia has a "legal" limit of 4W EIRP (+36dBm) for 2.400 to 2.463 ghz and 200 mW EIRP (23 dBm) between 2.463 GHz and 2.4835 GHz. (For operation of devices under the "low interference potential devices class licence": a maximum radiated power of one watt EIRP is authorized in the 2.4-2.45 GHz band for telecommand and telemetry transmitters and radio frequency identification transmitters; and a maximum radiated power of 10 milliwatts EIRP is authorized in the 2.4-2.463 GHz band for all other transmitters.) - exerpt from SMA.

This means that for wlan channels 1 - 11 the radiated power is restricted to 36 dBm, and above that it is 23 dbm, (also for general antenna computation the "centre" frequency for Australia's wi-fi is 2.4315 Ghz , that is a free space wave of 123.38 mm, a half wave of 61.7 mm and a 1/4 wave of 30.85 mm).

Channel . . . . . . . . . Freq (centre)
1 . . . . . . . . . . . . . .2412 Mhz
2 . . . . . . . . . . . . . .2417 Mhz
3 . . . . . . . . . . . . . .2422 Mhz
4 . . . . . . . . . . . . . .2427 Mhz
5 . . . . . . . . . . . . . .2432 Mhz
6 . . . . . . . . . . . . . .2437 Mhz
7 . . . . . . . . . . . . . .2442 Mhz
8 . . . . . . . . . . . . . .2447 Mhz
9 . . . . . . . . . . . . . .2452 Mhz
10 . . . . . . . . . . . . .2457 Mhz
11 . . . . . . . . . . . . .2462 Mhz
12 . . . . . . . . . . . . .2467 Mhz
13 . . . . . . . . . . . . .2472 Mhz

The EIRP figure include (as seen from the above examples), your antenna gain, any inline bi-directional amplifiers as well as the diving source and any losses, that is your effective radiated power at the antenna. This creates a dilemma because most people will simply go to the max to assure good transmission, with a high gain antenna, a big amplifier, or the biggest AP they can find. This philosophy is not quite correct however, because the more people using high power the more background "noise" there is generated on the band. Also the more unorganized overlapping frequencies in use will mean more collisions with less overall through-put for everybody (including you). The band space is a natural resource there for everyone like air and water, (contrary to politicians belief that they own it and can sell it to the highest bidder!), so we should all try to manage it properly so the bandwidth is maximized so we are all able to use it. By applying the examples above to your situation, and adding an additional operating margin of 10-15 dB for "fade", a minimum guaranteed specification can be made, which is fair to all users and provides a working link to yourself. Also select "more" sensitive receiver products over something with higher Tx output, use the correct antenna type for your application (not brute force, or needlessly radiate in directions not required) and minimize any system losses in transmission lines as much as possible.

The wlan channels are incremented by 5 Mhz , and because the mode of communication is DSSS (Direct Sequence Spread Spectrum) and has a bandwidth of 22 Mhz, (11 Mhz either side of the channel frequency), interference from adjacent channels is of great concern. To alleviate this problem users within their particular area should use channels at least 2 apart to avoid this problem. Where Metropolitan Area Wireless Networks have been established strategies have been developed to minimize these problems. "Cells" are organised, where 3 channels (widely spaced over the entire wlan bandwidth) service hexagonal matrixes of areas and users, so no interference is generated into the next channel. The channel used in one cell doesn't extend into the next, (or by much), and the next cells channel, (further out - middle channel), is spaced enough away to avoid interference with the first and also to the next adjacent one out, and the further one out uses the last channel within the bandspace allowed, Channels further out from this use the first channel again and so on. . . (hexagons cells are used because "circles" don't nicely fit together). This method of organising is called frequency re-use, or channel re-use, it provides a "buffer" between interfering channels and has been widely used in phone networks and other wireless networks.

The 2.4 Ghz band is publicly shared with many commercially available devices, such as TV extenders, garage door openers, telephones, audio links, and it is also shared with Radio Amateurs. This arrangement causes a lot of background noise for this band, but what it must mean is local community users, commercial manufacturers and Amateurs must cooperate together to squeeze all they can out of the tiny band space allocated to them. (not that I'm complaining . . . I'm sure that we all pay much less tax from our respective governments because of those that do have to pay for other band space :) Because much of the commercial equipment available uses vertical polarization, the use of horizontal polarization could be employed within congested areas to possibly provide a solution to interference (horizontal polarization is usually employed for Dx work as the EM field degrades slightly less with terrain than with vertical).

The Wlan transmission employs many different methods of signal encoding and modulation, (BPSK, QPSK, DBPSK, DQPSK, CCK, OFDM, etc - the hardware.) to provide data rates from 1 to 108 Mega bits per second, (Mb/sec). The software protocol beneath provides multipoint networking, (the IEEE 802. whatever standard), the device also has a unique "MAC" address which allows an unlimited number of active devices to be on a given network (each node is unique), and the solution of contention for network access is handled through CSMA/CA, (carrier sense multiple access with collision avoidance). Security is yet another software layer in which encryption is usually employed, but even with this unauthorised access might take place, and the algorithm decryption certainly slows data through-put down. Newer methods of security are being explored such as WPA (Wi Fi Protected Access). The result of all this can be a very fast network, or a dog of a one. (Nodes of differing preset speeds sorting out a mode and speed, packet decryption, packet collisions from other nodes, "other" interference / noise, data errors and error checking, path fade, etc etc. - With these normal overheads expect an 11 Mb / sec link for example, to have a "real" through-put of 5 to 6 Mb /sec). Devices are now available (mid 2004) for multi mode and dual band operation, which means they can lock to 2.4 Ghz or 5 Ghz transmissions at various speeds. (108 Mbps is capable of realtime video and audio streaming.)

Another mixed blessing is "antenna height". Most antenna systems work better the higher they are, and Wlan systems and other microwave systems are no exception and also require additionally, (optimally) a "line of sight" between antennas stations. The draw back with "height" however, is transmission line losses which eat into your EIRP power. Inserting an in-line bi-directional amplifier will compensate for coax losses, but it does also introduce a lower signal to noise ratio at the antenna than one just coming from the plug at the source. Signal to noise is important too if we want to maximize the throughput of the link, . . . a clean low level signal might have more through-put than a noisy high power link.

Lastly it is a good idea to join your local wi-fi community network / "club". Many of these organizations provide valuable databases of wi-fi systems perhaps in your area. This information is valuable in appraising whether a link can be established, and many also provide tools and other information to enable you to painlessly join the network.

"My" Antennas.

As my Wlan system was fairly well "outside" the "Brismesh" node system (Brisbane, Aust.), I decided on building a decent omni-directional antenna and see what happens. I thought that 32 elements, (lots of gain) while it seems a lot, wouldn't be too hard to build using brass tube because the element lengths weren't too long (cost factor), and a high degree of accuracy could be achieved using some small pipe cutters fairly easily and reasonably quickly. I also decided on using element insulators between sections so I could achieve a better theoretical model, as close as possible to a 1/2 wave. I also decided that making a jig for assembly would further increase accuracy, so I thank the guys on the web that come up with that. The 32 element antenna should give 12.65 dBi. This was my first antenna for wi-fi use, and after testing with mixed results several more were to follow, and are described further down the page.

The results of the 32 element antenna were fairly poor, and I believe this was because the beamwidth was too narrow (about 2 degrees). It became too difficult to align to another receiver, (The problems of gain Verses beamwidth of a vertical antenna was something I hadn't considered really to be a problem until I looked up some figures on this). I found myself in the position where my rubber duck PCI card antennas wouldn't go further than 80 metres, (and that was doubtful too), and the 32 element collinear high up on the roof was projecting well over the top of everything and difficult to quantify, (but I still optimistically think it worked). I could drive around the neighbourhood with a laptop and perhaps never find my signal, and as I have very little test gear and no GPS to quantify my, or another's antenna position, I then decided to break the antenna into smaller sections and try them out, (2 x 8 elements and a 16). These would provide broader beam angles and be easier to align to another node. This also high lighted to me that "casual" installation will more often than not, not work, and ideally Tx and Rx antenna positions need to be plotted, elevation angles (and heights) calculated, antennas of the correct beamwidth and gain be employed, and rigid antenna mounting points be provided. (no swinging in the breeze stuff)

Some typical figures for collinear systems are:

2 element . . . . . . . . 2.15 dBd gain. . . . . . . 4.25 dBi gain. . . . . . . . 32 deg elevation Beam-Width (approx)
4 element . . . . . . . . 4.25 dBd gain. . . . . . . 6.35 dBi gain. . . . . . . . 16
8 element . . . . . . . . 6.35 dBd gain. . . . . . . 8.45 dBi gain. . . . . . . . . 8
16 element . . . . . . . 8.45 dBd gain. . . . . . . 10.55 dBi gain. . . . . . . .4
32 element . . . . . . . 10.55 dBd gain. . . . . . 12.65 dBi gain. . . . . . . .2

Another embarrassing "hole" I fell in was with "N" type connectors. I purchased a roll of cable and several connectors after explaining what I was going to do from a company (who will remain nameless), only to find (eventually, grrrr) that the pins in the "N" type connectors were incompatible with each other. The male pin from the pigtail was too small for the female contacts (the type which fitted my coax, from the roll), and being high quality and well made, the male pin was concentric within the female but with a 5 thou gap all around it. . . . .it looked as though it was right, but no cigar! . . . . don't get caught with a hole in your transmission line, buzz it out.

I next rebuilt the antenna as an 8 element collinear, and again re-tested it's performance against a "known" antenna. Again, I was disappointed with the results. It was much lower than a commercial vertical collinear (stated as 8 dB).

I would estimate the 8 element coaxial version at a lousy 4 to 6dBi. (This estimation has been gauged by antenna comparison noting the PC link quality and signal strength. It must also be stated that the power available to the antenna after 12 metres of LMR400 coax and supplied by a wlan card was probably pretty feeble, and an ant crawling up the wall probably affected the field pattern I was trying to watch).

I decided to play with the antenna at this point, I removed the 1/4 wave tip and top 1/4 section to find an increase in gain, but it was still well below the commercial one. The adjustable collar seemed to work ok too, as I was able to peak it for best performance and it was around the "theoretical" position as well, (short at 1/4 wave from feedpoint.) . .

The photo below shows some of the coaxial colinears built, as well as a commercial collinear, and a "trough" which a friend loaned to me to try. The first collinear in picture is shown with a section of "dry" dowel cable tied to the elements to "hold it all rigid and straight". Tests were performed with and without the dowel and found to make no difference to the results. The second collinear into the picture is the "homebrew" Aerialix design. After these I decided to build a microstrip type since it seemed to be the best omni vertical design I had tested.

Photo shows some of the antennas, "Aerialix's" design is lower down the web page, as is the coaxial, but I think the one to build is the Microstrip type . . .

So at this point I have built several colinear antennas, but none of them blow away the commercial stripline job. . . .( gee I hate that . . . when a plan doesn't come together, but they all work, . . .a bit, . . . .maybe,. . . . . just,. . . . . . . . . aw ok their shit!). Anyway maybe you could enhance them, (and please let me know if you did, or the results of any test gear on them.). Testing in an "un-interfered" location is a must, and perhaps part of my problem (frequency interference from phones, garage door openers, tv repeaters etc etc.), . . . so find I nice big paddock away from the city and do some good . . . .

Microstrip Antenna . . a better way!

As I had good results with the commercial "microstrip" style antenna while testing and comparing it with the others (verticals), I decided to have a go at one of these. These comprise of strings of PCB elements connected together with PBC striplines (transmission lines). The striplines are calculated to 50 ohms impedance, and don't radiate (leak power), as they are straight conductors without discontinuities. (ie fringing electric fields cancel out). They seem to work very well probably because they do away with lossy phase shift coax sections or phase coils.

The antenna consists of PCB rectangles 31mm long and 14mm wide which are connected together with microstrip lines also 31 mm long and 2.5 mm wide. (widths may vary depending on your PCB material, dielectric material, board thickness, copper weight, etc. - consult your microstrip transmission line calculator). You will require about a 54 cm by 14 mm piece of double sided PCB material, 1.5 mm thick, a bit of LMR-200 about 100 mm long, and a suitable male N type connector. I painted a "positive" with enamel paint after marking it out, and then etched away the unwanted bits. This should give you a vertical omni-directional antenna with hopefully better than 8 dB gain.

First paint one side of the PCB, (note I had to make mine in two sections as my PCB material wasn't long enough, I am going to join them. - see photo). Wait until the enamel paint is dry , then do the other side. Then the copper is etched away leaving the wanted bit. (Don't make your acid to strong (or hot) as the paint may fail and undercut the tracks. Better to be patient, and don't forget about it either like I did once, and found a nice clean piece of fiberglass sheet)

After etching use some thinners to remove the paint, and also clean up the copper etching if there are some faults with a sharp trimming knife etc. It is important that your tracks have nice straight edges, furry edges will encourage radiation leakage = Losses. (If your tracks undercut, make them slightly wider on the next go and trim them back). Prepare the tail of LMR-200 to a male N connector to have about 50 to 60mm of coax from the connector end. Cut the coax back so 30mm of shielding is exposed with a 1 to 2 mm inner foam extension and a 10mm pigtail. (ie cut plastic 42 mm from end to expose braid, cut braid to 30 mm, cut 10mm pigtail)

Drill a hole to allow the tail to be soldered to the first microstrip track underneath - see photo. Drill another hole in the centre of the first segment through to the track underneath. And another hole in the last copper segment, centered as well. Solder a piece of tinned wire through the last two holes to the tracks underneath. Pare away the track from the solder point to the extremities if you made the trackwork go to the extremities. (So outermost segments solder to a 1/8 wave length point of track) Push the coax pigtail through the "first " hole drilled, solder it to the track underneath (bend it over). Solder the coax braid in several place to the first copper segment. Offset the coax to avoid the soldered wire through to the other side. Thats the antenna . . . All that is required now is to make the radome and a mounting bracket.

Feed arrangement

Radome - This consists of a piece of plastic conduit (electrical or plumbers), long enough to accept the antenna and wide enough to slide the "N" connector into. Styrene foam or low density "packing" foam can be used in a couple of places to stop the antenna rattling around if you wish. Check a sample of conduit and foam out in the microwave before committing yourself to it. Finish the radome with a cap glued to the top, and in most cases there are end caps for all sorts of different conduit ranges.

Top view

Mounting bracket - Obtain a N type chassic mount female to female adaptor. Bend up a piece of aluminum or galvanised steel plate at 90 degrees with holes for the "N" adapter, and other holes to accept an exhaust clamp or other "U" shaped clamp. Screw the antenna onto the top adapter female, the other goes to the transmission line, bolt it to some pipe or box section.

Back view

The radome may detune the antenna due to capacitive coupling, so test your antenna uncovered, and then completed with the radome. If you have a marked reduction in performance try selecting a different conduit "type". (ie material and/or diameter) The beam width of this antenna is only several degrees so remember to check out the "target" is within it's radiation zone, or it may shoot straight over the top!

The 31 mm long sections are 1/4 wave sections, length may be experimented with if you wish. There will be an optimim length as the raydone material of your choice will detune the antenna. A progressive lengthening is worth trying too, to broaden the bandwidth, also current drops in each section towards the tip. The antenna could be scaled to use 1/2 wave through to 5/8 wavelength sections too. (Let me know of any results there thanks)

Phase Coil Type

I also found a design by "Aerialix" which interested me greatly too. They provide a kit, and I would have bought the kit in a heartbeat if I could afford the Australian international postage, . . . so I made it from scratch. (but their kit looks pretty good)
This antenna is easier to build than the coax versions, mine consists of 1/16 inch brass rod and 3/32 brass tube sections. The rod is used to fabricate the phasing coils and the tube is used as the "element" sections. The complete details can be downloaded from "Aerialix" and presumably instructions come with the kit.

I looked closely at the "Aerialix's" photo's and thought the phasing coils were actually 4.5 turns even though their instructions say it is 4 turns, so I made mine with 4.5 turns. On refections I think I will rebuild this antenna with 4 turns, and maybe play with the diameter too. (they also mention that the coils can be wound with a "screw", this can be seen in the photo, some other designs just use 1 turn of about 10 - 11 mm diameter). All elements are "inline" with each other, I used some fishing line to apply tension to hold it straight within the conduit radome. The photo on the right shows the isolating section, a bit of water hose makes a nice seal with the "conduit" radome to push on to, a rubber bung is used at the top end of the radome to weatherproof it, with a small cap off a bottle which sits on top to cover everything. (A small hole in the side at the bottom of the sleeve is a good idea to allow any water to drain from it, that way it should not work it's way down your coax centre if the worst does happen). Aerialix also supply a positional jig PDF which is well worth grabbing for assembly.

My local store again supplied me with the materials, and you get about 7 coils from a 3 foot length of 1/16 rod and 1 length of tube is ample for 8 , 1/2 wave elements. It consists of alternate 1/4 wave sections incorporating a phasing coil with 1/2 wave sections. Decoupling is done with a "sleeve" (Aerialix suggests 11/32 inch brass tube (8.75mm) 30mm long), 1/2 wave length away from the feed point and the sections are just stacked to give you your required gain. (If using 1/2 copper water pipe as the sleeve like I did, use 4.7 mm OD tube from the connector to the first coil (photo above), and my sleeve was 32.5mm long to cater for the physical construction of the "n" connector I used - but it was 30mm electrically long!). I tested the completed antenna against my "commercial" one and found it just below in gain (but better results than the "coax" versions though, - wish I had bought the kit. . . .) )

Construction of the coaxial collinear, - 32 , 16, 8 or 4 element are all made the same way

(I would avoid this type and try the "aerialix" design for better performance, this type is better suited at lower frequencies.)

The coaxial collinear is constructed with 1/2 wave sections which are alternatively reversed connected ( see diagram below), to generate an "in phase" field pattern. The more 1/2 wave sections included into the string the higher the gain of the antenna, by doubling the number of elements, the gain is doubled. (There is a practical limit to this theory though as the current distribution in each element gets smaller the further out along the string - the field becomes weaker, each additional element increases gain by a lightly smaller amount each time) This system has been constructed with coax with the outer shield retained and with brass or copper tube replacing the outer braid and shielding. Brass / copper tube is certainly more durable, more rigid and possibly easier to solder together but it probably doesn't really affect the performance of the antenna when both types are built neatly and accurately. "Foam" dielectric is better for this application over "solid" dielectrics because this type is generally a low loss coax (current is converted to field and not heat) and allows the element length to be "electrically longer" which slightly increases the capture area of the antenna, and helps increase the gain. (higher velocity factors). By not employing a connector at the bottom of the antenna stack further losses can be avoided as each connector within the transmission line produces a "bump" in the standing waves which adds an insertion loss.

(photos below show the elements being assembled with plastic insulators, this was changed for mica, and the jig is about 1.5 m long with slots cut along to let the insulator washers drop into them. Right picture shows the completed 32 element antenna (less radome - which was some plastic conduit).

I started with an insulation thickness of 0.75mm (mica from a toaster element, tried plastic, melts too easy), and worked back from that. Using LMR400, with a velocity factor of 0.85) the 1/2 wavelength is 52.5 mm (using a centre frequency of 2.4315 Ghz for the Australian band space), and allowing for the insulation width my brass tube sections were cut to 51.75 mm, the LMR was stripped back of shield and cut into 67.5 mm pieces, giving me about 8 mm pigtails to play with. I slid these into the brass tube sections and evenly spaced the ends out each side, ring-barked the dielectric where the tube finishes and that became a completed 1/2 wave element. (one end may need a taper shoved up it (ie, a centre punch), to undistort the tube from the cutter to get the LMR through it, the "rounded edge does tend to lock it into position though).

It was then a simple matter of threading on the insulation washers, placing the sections in the jig and soldering it up with each sucessive element spaced at 52.5 mm . The sections were checked with a vernier caliper and a sand block quickly takes off that 0.2 or so the cutter adds on for you. (as it stretches it). The mica insulators were stamped out using a "wad" punch, then 2 holes drilled into them to take the inner conductors.

The final section (shown above) was made using 1/4 wavelength of tube using coax, (26. 25 mm), and the tip was also a 1/4 wave brass tube section, (29.3 mm), which has a velocity factor of 0.95 applied because of an "air" dielectric. (This can be made using tube or wire, note the 2nd photo down and on the left). The centre conductor of the 1/4 wave coax section is bent at 90 degrees to come out a groove in the tube tip section to solder to the outside brass section, or conversely soldered to the outer and a wire tip section. It too was installed using the assembly jig. I found by cutting the tube tip to say 34 mm, flaring the end by driving a taper up it (before cutting your slot), the tip would fit nicely over the 1/4 wave section tube, it could then be cut to 29.3 mm as you can see the end of the other section through the slot and then solder it into place.

The isolating balun section presents the most difficulty, the sheath is replaced with a brass sleeve 3/4 wavelength long so the collar can slide around a 1/4 wave node point for tuning. The sleeve "short" (a 2c piece), is positioned 1/2 wavelength from the terminating end which will be soldered to the 1/2 wavelenght string of phased elements. The 5/16 brass tube becomes the "inner" conductor for our new section of transmission line. The inner / outer diameters for the sleeve should ideally fit this equation Zo = 138 LOG D / d, so plugging in the inner (approx 8 mm) and a Zo of 50 ohms, the outer should be around 18.4 mm. The length of the collar should be 29.3 mm as well because it is essentially an air dielectric, if you want to solder it into position and make the collar longer to cut back for tuning that should be ok too. The new 3/4 wave brass outer of the coax is terminated to the shield braid by winding some tinned wire around the end and solder into position. (see photo above).

11/32 brass pipe fits snugly over 5/16 pipe so makes a good choice for the sliding sleeve inner tube. The 11/32 pipe is soldered into a hole in the copper short (2c piece), concentrically to the outer tube. The correct diameter tube for the sleeve was difficult to obtain so I used some copper sheet to form the correct diameter. After soldering the seam of the sleeve, tie a piece of copper wire around it (low) to act as a heatsink for the soldering operation to the copper short, otherwise the seam may spring apart. The tube ratio for 50 ohm impeadance is 2.3 times, ie the outer tube I.D. needs to be 2.3 times larger than the inner tubes O.D. The photo above shows the finished antennas, the 2 eight element versions, and a 16 element.

"A 2.4 Ghz trough reflector".

A mate gave me a loan of his trough to try out on Wlan. (thanks Pete) , well I put it on the bench and the dam thing fell apart! . . . really . So I had a quick peek.

Oh, and I threw the calipers over it too. . . The commercial version works very well, it is claimed to be 10dBi with a beamwidth of 40 deg so pointing it shouldn't be much of a problem (reflector width = 125mm, omitted from the drawing).

The homebrew version. . .

I thought I would try building the antenna with cheap, easy to get components and see how it performed against the commercial version. As seen from the photos the reflector was made from some scrap aluminium (with a couple of small holes in it, but that won't affect things), the dipole made from a strip of copper sheet. The radome is made from a plastic utility box with a piece of 16mm conduit glued into the side of it. The feeder was made from a 750mm length of RG-213. I followed all the measurements of the commercial unit, but varied the radome and the way the dipole was all held together because the commercial radome is hard to acquire.

The dipole must be accurately cut. The photo on thefar right shows a section of RG-213 coax (or LMR400 is better) with the 1/4 wave balun and dipole. The inner coax conductor is soldered to one side of the dipole at the bend. The section from the coax soldered joints to the bend with the soldered joint is the 1/4 wave stub which is the isolating section and stops rf from travelling down the coax

The utility box had some slots on both sides which I used to hold some plastic strips into place. These strips were slotted to a depth which held the dipole assembly central in the conduit inner space, and central within the box, ( junior hacksaw cut width is a perfect fit). A small amount of "silastic" was used to hold things into place, check this doesn't affect the tuning, I've been told some glues etc., are very capacitive at microwave frequencies. A wooden wedge was also used to lock the RG-213 into the conduit too, ( not too much as to distort the conduit shape or damage the coax). The radome conduit slid up into a 25mm box section which was angle bracketed to the reflector. (brackets on the rear of the reflector and not visible in the photo). A PK screw screwed into one side of the box section locks the conduit into position. (the "point" is flattened off). The dipole is 39.5 mm from the bottom of the reflector and held parallel with the reflector sides by this screw. (this also allows the dipole deck height to be adjusted), final tuning can then be achieved if you have the test gear.

The first tests were a bit sad but after checking a few things out I found the plastic which I used to hold the dipole and reflector into position (above and on the left), wasn't microwave and dishwasher safe, this has now been corrected (see the photo above in the middle). The photo below essentially shows how little there is in this antenna, and the cable tie around the radome is just there until testing is finished.

Two out of the 3 antennas finished and ready for testing. Things to note here are, the commercial antenna has 2 water drain holes strategically placed (one has a red stopper)

A Parabolic Trough.

One of the things a did while playing around with the above trough was to check the geometry of the reflector as many troughs use angles of 60 deg through to 90 degrees, but I had never seen one at 40 deg. (the one above) So I asked another person to aim a laser light pen from some distance away into the apeture (moving the beam from the side to the middle) and analyzing the focus with a sheet of white paper held in the middle. I found the results weren't very focused so I decided to make a parabolic trough and compare the results between the two.

.

The dipole and isolation section was made using a chassic mount "N" connector, a length of 1/2 inch copper water pipe, a length of brass tube scavenged from a telescopic antenna (4.7 mm) , and some 2mm wire. The reflector was made from a strip of 1/8 aluminium sheet which I had lying around (of no particular dimension, reason? - I didn't have to cut it, but was 96 x 640 mm) which was initially bent around a gas bottle and then reformed by hand to fit the curve of a parabola as drawn out on a template. I found the smallest tube recovered from the telescopic antenna fitted nicely over the N connector tab, a small section of this helped in providing a good fit from the "N" connector tab to the feeder tube section. (the 4.7mm brass tube). A 50 ohm impedance section is required, and this is achieved using the 4.7mm OD brass tube and 1/2 inch water pipe with a 11 mm ID.

I then cut off a section of 1/2 copper pipe (62 mm), flared one end slightly, cut a slot down the end of the non-flared end centrally and equally to 31mm. ( with a single hacksaw blade). This "cut" section becomes the "isolating" section. The copper pipe was then soldered to the "N" connector on the flared end, ensuring it is positioned perpendicular and concentrically. The section below the cut is just a 1/4 wave 50 ohm transmission line from the "N" connector to the isolating section. The inner pipe is then cut off just protuding the 1/2 pipe. The copper wire "dipole" was then soldered into position, one side of the dipole to the outer of the copper pipe, (as shown in the photo) and the other from the inner 4.7mm pipe, (held concentrically), to the outer 1/2 inch pipe, ensuring the dipole is "in-line" and at the same height. (Picture below shows the dipole close-up) The dipole assembly is mounted in such a way that nuts can easily adjust the perpendicular-ness (if there is such a word) and height above the reflector. The wire dipoles are positioned exactly at the focal point. (50mm) and can be "tuned" to length if you have the test gear. (If your like me and flying by the seat of your pants, just cut it so both are equal, and with an overall length of 61mm, two 1/4 waves and a bit less to cater for "end effect", {more guesses}).

The above photo shows the final "fit" to the parabolic curve on the template. (strong fingers are required at times)

The focus was checked with a laser pen and found perfect all around the parabola (except for near the tips). The "arm" was also just a bit of aluminium sheet lying around. It had a flap already bent on it which I thought would give it strength. I just bent it in a vice and filed the edges up with a belt sander, finishing it with a file.

Because the dipole is so far out from the groundplane (the reflector), I decided to make a sliding radial ground-plane which I could move and lock into position anywhere along the dipole assembly to adjust the antenna impedance. My initial set point is 1/4 wave from the dipole, (ie radials perpendicular at the end of the balun slot).

This has provided good results with the antenna out performing the commercial 10 dB trough by quite a margin. The ground plane radials consist of 6, 1/4 wavelength long bits of 2mm copper soldered equally spaced around a copper ring which clamps the 1/2 inch copper pipe dipole assembly. (see photo below) A 1/8 metal thread clamps it into position. I am very happy with the initial results of this and shall continue to experiment with this to see if I can squeeze anything more out of it. (Unfortunately I have to rely on the "other end" for signal reports as my wireless driver in ad-hoc mode tells me absolutely nothing, but I shall try to gauge relative performance against the 2 commercial antennas available to me, (the 8 dB vertical collinear, and the 10 dB trough), and publish these results.

I also suggest that a laser pen might find uses in lining up dishes etc by carefully taping it to the dipole or another perpendicular part if the antenna , (After 1500 metres or so the spot will probably be getting a bit large though). I would also suggest that this would be better done and seen at night, and please be careful of your eyes. Superbright leds, laser leds are bright enough to damage your eyes permanently. I also have to turn the mounting arm around 90 deg to work my friend up the road with a vertical. - updated photo's and results coming. . . .

(this can be knocked up very rapidly and would make a good point to point antenna project, for between friends, school projects - between school buildings etc, making "microwaves" fun if that is at all possible. . . )

Corner reflectors, troughs and things.

The corner angle can be 45, 60 or 90 degrees, but as the angle is decreased the length of the sides must be accordingly increased. For a 90 deg reflector the driven element spacing should be from 0.25 wavelength to 0.7, for a 60 deg reflector a 0.35 to 0.75 spacing, and for a 45 deg reflector it should be 0.5 to 0.8 wavelength. (for "flat sheet" it should be 0.1 to 0.3 wavelengths). Closer spacings yield lower feed point impedances. A folded dipole (approx 300 ohms) could be used for wider spacings. (both ARRL and Jessop have the same nice drawing depicting resistance verses wavelength spacing). Reflector lengths should be a minimum of 0.6 wavelengths and if the reflector is made of mesh or of a grid, spacings should be no more than 0.06 wavelengths. A "mesh" system is ideal if the surface geometry is accurate because it reduces wind loadings, and surface imperfections/geometry should be kept lower than 1/10 wavelength so it doesn't affect gain (particularly for dishes and other parabolic types). Smallest mesh size or solid surfaces will yield the highest front to back gain ratios and best efficiencies. Forward gains of 10 to 15 db are typical for corner reflectors and troughs, dishes and parabolic troughs can have much higher gains depending on the diameter (and operating wavelength of course).

The type of driven element will determine largely the bandwidth of operation too. "Fat" cylindrical elements or a triangular dipole (small wavelength to dia ratio), will give greater bandwidths than one constructed with "thin" elements. Wider spacings between the driven element and the reflector also give greater bandwidths (and a trade off with radiation resistance).

"Trough" reflectors are "corner" reflectors with the vertex chopped off. With this done the physical size of the antenna can be substantially reduced without affecting the gain terribly. The dimensions of the "cut" have to be adhered to so the antenna performance is maintained.
90 deg. . . . . . S= 1.5 wavelengths. . . . . . Gain= 13 dB. . . . . . . T=1.0 to1.25 wavelengths
60 deg. . . . . . S= 1.25 wavelengths. . . . . Gain= 15 dB. . . . . . . T= 1.0 wavelengths
45 deg. . . . . . S= 2.0 wavelengths. . . . . . Gain= 17 dB. . . . . . . T=1.9 wavelengths
(where T is the distance from the vertex to the truncated flat plane, S is the distance from the vertex to the driven element)

The commercial trough above uses a 40 degree angle, an S of 1.07, a T of 0.74, driven element to reflector of 0.32, Side length 1.23, aperture 1.46, width of 1.02, wavelengths, . . . . and it seems to work very well.

Some info for this from the bible(s), (ARRL antenna handbook, and Jessop VHF and UHF man.)

Some notes for Element spacing / phasing

The ARRL Antenna handbook shows a graph depicting the gain of two 1/2 wave colinear elements verses the spacing of adjacent ends. No one would be surprised to see a maximum at a 1/2 wavelength, at about 3.4 dB, but what is interesting is at 0.4 and even 0.3 spacings the gain is still over 3 dB, and then rolls down to 1.9 dB with zero spacing. So to optimize the gain of colinear elements the spacings need to be between 0.3 to 0.6 wavelengths. Another aspect in the design of phasing sections is that radiation occurs as a result of accelerated charge, and it appears that this begins to get pronounced after about 1/10 of a wavelength (36 degrees), so phasing sections should be designed with this in mind. The physical construction can also provide leakage of RF by discontinuities and unbalanced structures, (as is exploited in "microstrip antennas"), where the RF "fringing fields" are not balanced by the physical properties of the conductor radiation can occur.

Impedance Matching Systems

It is fairly obvious that an open circuit or short circuit condition will provide no power transfer, it is only when the line from the source is "matched" to the connected load that maximum power transfer can take place. To complicate the situation loads and transmission lines are rarely purely resistive but rather a mixture of resistive and reactive parts.

Impedance is a term used to quantify the resistive and reactive terms so it can be mathematically depicted and solved. A transmission line (from a source) may have a "Characteristic impedance" ("Z") of 52 ohms, but it also has capacitive and inductive parts which must be considered when matching it to a load, (the load itself has resistive and reactive parts too). These reactive parts will place bandwidth limits over which good power transfer can take place, but manipulation of the reactive parts can sometimes resolve some of these problems. (example - changing reactance of matching system, changing dipole parameters to change reactance). When a perfect match is achieved a "standing wave" within the transmission line is established where maximums of current and voltage can be found along the length of the line, and can be predictably found knowing the frequency of the source (it''s electrical wave length) and the "velocity factor" of the transmission line.

Stub tuners are a section of transmission line that incorporate an adjustable "short" which is connected across a transmission line feeding a load. By adjusting the short the reactive component of the stub can be tuned making the transmission line match the load. "Double" stub tuners are quite often used so the tuner doesn't require a specific electrical distance from the load for it to work. Stubs are usually placed at odd multiples of 1/8 wave spacings.

Series section / linear transformer transmission lines -
An impedance transformation can be made by using a length of "different characteristic" impedance transmission line from that of the feeder transmission line so it may match a load. (example - a 50 ohm source impedance may be matched to a 75 ohm load by using two series sections of coax, 29.3 degrees long (0.081 wavelengths). The 50 ohm source to the 75 ohm coax section to the 50 ohm coax section finally to the 75 ohm load) The ARRL Antenna handbook Ch 26 -14 covers the maths to work any coax section to another for any inputs and loads.

Taper sections,- Klopfenstein, Hecken, exponential and triangular tapers gradually step the characteristic impedance over the length of the section. In an analytical model where the taper is broken into many parts (steps), each step can be analysed as a microstrip line with it's own characteristic impedance over the length of the taper. {"Horns" (antennas) are a structure which exploit this principal giving it a very broadband use and are often used to efficiently couple to waveguides}. Charts are usually consulted for the width, area and length of the taper for design as the maths involved is heavy going. This type of matching is used in "helical" antennas to reduce the helical impedance (140 ohm) to coaxial impedance (52 ohm) with an 85.3 ohm transformer.

A 1/4 wavelength transmission line transformer, can transform any load impedance to any desired value at the input of the 1/4 wave line.
Z (char. impedance of line) = SQR( Z {impedance at the input end of the line} * Z {impedance at the end of the line})
(A match between 52 and 600 ohms is achievable with common coax impedance sections)

The "T "match system is a balanced system coupling to a dipole where parallel lines (>1/4 wavelength bars ) attached the dipole and to the feeder. The parallel bars usually have clamps to the dipole which may be adjusted to tune the matching system. Because the matching structure is inductively reactive, the antenna is usually "shortened" from resonant length to provide some capacitive reactance to counter balance (another example of the manipulation of reactive components) Max impedance of the dipole is about 40 - 60 % out.

The gamma match is similar to the "t" match but employs (variable) capacitors to couple to the dipole. It also is a balanced matching system and has found popularity in "all metal" construction antennas. It can be difficult to tune as each dipole leg is tuned separately, but it must be balanced for operation over both elements.

Omega match is similar to the gamma match but uses a shunt capacitor from the parallel matching line to the dipole at the feed end. This is done to null the inductive reactance of this type of system.

Hairpin and Beta matches . This form of matching must be fed with a balanced line such as a 1:1 balun to coax. It consists of a tuned length of wire or tube which is "U" shaped and connected across the dipole ends, with the 1:1 balun connected to those same terminals.

Baluns - Because antennas and their matching systems can be balanced or unbalanced structures, and transmission lines can also be either, a method or connecting balanced things to unbalanced things is necessary. Common ways to achieve this is with a quarter wave open balun, (pawsey stub), a coaxial sleeve balun, or a 1/2 wave coax balun (giving a 4:1 transformation as well) - the last 2 are my favourites.

VSWR meter

Pe2Erwin@hotmail.com published a VSWR project that looked so good and useful that I thought I would have a go. I scrounged up some surface mount resistors and caps and preceeded to lay them out the best way I thought. Unfortunately I lost the small pad on the "rf in" side, but this shouldn't be too much of a problem as the components and pin should all solder together properly and nearly form the same shape as the original pad. The above photo shows the "N" connector pins ground down to about 2 mm, the 10 pF caps and diode yet to be placed.

I deviated form the original article slightly because I thought that by making the "critical" components in the bridge's layout as small as possible I might increase it's useable frequency. (and hopefully cater for the 5Ghz wlan cards too). The board I used is double sided with the pads sliced out with a trimming knife. ( keeping in mind that 2.8 to 3mm widths yield 50 impeadance with 1.5mm thick PCB (k=0.52)). The PCB is soldered to the N connector, (top and bottom), to form a good ground plane, and the BNC output "DC" connector is also soldered to the N connector. The photo below shows all the components in place with the other N connector also soldered to the ground plane. 1/8 metal thread and spacer provide more mechanical strength.

I had a bit of trouble finding a suitable diode for this application as my PCB layout tied me to an axial type, so, in short I had to use something a bit lower than what I had hoped for, a 1N5711 (about 2pf junction cap). the layout is a bit hard to see because it is fairly small, but the "N" connector pins are seperated by the length of a SMD 100 ohm resistor (2 in parallel), and I tried to balance the other pads length and ground plane seperation in the bridge so it would exhibit a simular characteristic and hopefully null out.

The photo's below show the completed SWR "head", (and the underneath view), more info will follow when I test it . . . .

Re-Hashing an old parabolic dish for Wi-Fi.

I saw a wire grid parabolic dish (cut down version) at a recycling centre and couldn't resist it. ( For the overseas people - this is a place in Australia where the local councils allow us to buy our rubbish back, with no guarantees and sometimes at relatively low costs). At that time I hadn't a use for it, but, the Amateur Radio part of me said you need this, and 3 years later so I do. (photo below shows reclaimed unit with radome apart)

The dish was probably part of an old satellite TV system, don't really know. There are probably many of these still stuck on roofs around the country after the demise of a couple of companies offering this TV service, so if you see one grab it, it may be useful for Wlan use. There is not much really to talk about , the modification is very simple. I pulled the radome apart, found it to be using the same concepts as the 10 dB trough, a "dipole with balun and reflector", so I rebent the balun to the same dimensions as it, (the 10dB trough described and pictured above), and cut the dipole and reflector to the same dimensions as well, and then re-glued it all up again. The hardest part of the modification is to get the radome apart without destroying it, (some are severely glued, and some may be brittle). The dipole must be set back up at the dish focus, and if you note the position of the original before you pull it apart, thats not too tricky either. (My dish has an adjustable radome height). Providing you use the same diplole "fixing pegs" as the original there is no need for any height adjustment. Unfortunately for me there were no "fixing pegs" 22mm away from the dipole for the reflector, so I used some spare pegs to create a mound of melted plastic for the reflector to sit in, that worked fine. (The reflector must be mounted parallel to the dipole and directly opposite and symmetrically , at the same "height").

The photos above show the original dipole and reflector, on the right is the "adjusted" version, reflector is held in place by melted plastic that wraps over the element. If you have the testgear to "tune" it, there will no doubt be a slight improvement when it is tuned up on the new frequency. Otherwise, a lick of white paint, fingers crossed and it's ready for use, below is the finished unit. Another node nearby is finalising his antennas and performance reports should be soon.

"The guts of a 2.4 Ghz rubber duck"

Fairly easy to pull apart an make or fix, some are 2.2dBi others 3 dBi, depending on the sales pitch. Collinear versions are available too at about 5 dBi.

Some terms -
dBi - is the gain over an isotropic source( theoretical "point" source)
dBd - Gain over a dipole antenna (0dBd = 2.15 dBi)
dBq - gain over a 1/4 wave whip
SWR - standing wave ratio
VSWR - voltage SWR
Lobe - shape of the field pattern from the antenna, shows directivity and field strength areas.
Polarization - is the orientation of the electric field that gives maximum radiation
(receiving a horizontally polarized signal from a vertically polarized antenna is very inefficient, and has a 3dB loss in signal strength)
Gain - the attribute of an antenna to "focus" the incoming energy into one direction. (quantified as so many times better than a dipole or isotropic radiator)
Capture or aperture area - the antenna's ability to recover a signal, this relates to cross-sectional area (wavelength) but ALSO to gain
(dipole = (wavelength) squared / 8, and is 1.64 times an isotropic radiator)

Baluns

Baluns are a method of matching an balanced line to a unbalanced line - hence the name BAL - UN. They quite often serve a dual role in that they also provide a way to match impedances. There are many types of baluns and many are frequency specific which may only work at low frequencies and others specific to microwave frequencies.

Coaxial cable can be easily pressed into service as a balun because of the nature of signals which traverse them. The cable slows the propagation of the signal through it by some amount which relates to the physical qualities of the coax. (like sound through water, or sound through air). This then gives rise to the fact that at certain points along the transmission line length there will be current maximums and current minimums ( and same for voltage), and that these points relate to the "electrical length" of the signal within the cable. This length for the signal to complete one cycle can be calculated out knowing the signals "frequency" and the coaxial cables "velocity factor". Using the "electrical length" of coax, the coax can then be used to apply some interesting characteristics to the "circuit" under consideration.

A half wave section with a short at one end will appear as a very low impedance (like series L and C at resonance)
A half wave section "open" at the other end will appear as a very high impedance (like parallel L and C at resonance)
A quarter wave section with a short at one end will appear as a very high impedance (like parallel L and C at resonance)
A quarter wave section "open" at the other end will appear as a very low impedance (like series L and C at resonance)
And interestingly enough a shorted line longer than 1/4 wave but shorter that 1/2 wave will exhibit a capacitive component.
An "open" line longer than 1/4 wave but shorter that 1/2 wave will exhibit an inductive component.
An "open" line shorter than 1/4 wave will exhibit a capacitive component.
An "shorted" line shorter than 1/4 wave will exhibit an inductive component.

(and the reactive components magnitude can also be varied by varying the coax line length )

This concept also applies to any conductor, a piece of wire or even to some wet string, and these principals underlie other matching systems such as "T" and gamma matches, hairpin stubs, omega matching and series section matching, also to the correct "phasing" of antenna arrays, or in some filters.

The 1/4 wave stub balun.

The voltages at the antenna terminals are different, (or the device we are matching too), one is directly coupled to the source, while the other is weakly coupled. (ie - unbalanced). This imbalance of voltages produces currents and fields flowing through both conductors which only partly cancels causing a large net current flow through the outer shield. The introduction of a shorted 1/4 wave stub (diagram below) provides a high impedance path for the current which then cannot travel down across the shield. The balun is constructed by using a conductor 1/4 wave length long which shorts out the transmission line at the feed point and to 1/4 wave length below the feed point onto the shield. Usually another section of transmission line is cut and used as the 1/4 wave stub section, but a wire conductor, or part of a chassic ground plane etc, will suffice, so long as it is electrically 1/4 wave long. The separation of the stub from the transmission line should be enough that the PVC covering of the 1/4 wave transmission line represents a minor proportion of the dielectric between them, the bulk of which is "air", and so the electrical length should be worked out with a velocity factor of 0.95. The stub should be a "parallel" line section to the transmission line.

The " sleeve" or “bazooka� balun.

This balun (diagram above) operates by incorporating an additional sleeve over the outer surface of a coaxial cable and doesn't provide any impedance transformation. (it's 1:1) The sleeve is a short circuited 1/4 wavelength away from the feed point. A "second" coaxial transmission line is then created with the new sleeve being the “braid� of the new line and the original coaxial outer becomes the “inner� of the new line. It works by presenting a very high impedance to the signal (as it is a shorted 1/4 wave section in air), which doesn't allow the unbalanced voltage at the feed point to develop any current to flow over the surface of the sleeve and outer transmission line sheath. The sleeve performs better than the 1/4 wave transformer as it's radiation losses are lower. The diameter of the sleeve to the coax line should be about 2 to 4 times for it to work satisfactorily, and its electrical length should be calculated at 0.95 as the principal dielectric is air. Tuning can be done by making the sleeve slightly longer so it can be trimmed back to achieve resonance, or mechanically provide a "sliding electrical" surface which will allow the sleeve to move around the 1/4 wave node point , and solder into position when satisfied. The equation below can be used to design an "air" dielectric transmission line of a specific impedance (for the collar), if you wish.

(Zo (for our use is 50ohms) = 138 LOG D/d , {where D = outer dia, d = inner dia, (coax diameter), and dielectric is air}

Bead Baluns

As stated somewhere in here, 10 turns of transmission line wound around a 6 inch circle will also provide a choke to limit unwanted transmission line current. This can be further enhanced with beads or toroids which greatly increase the reactance and resistance of the conductor. (see "N1HFX" article using toroids). Because ferrites are designed to operate at specific frequencies, correct selection is important, and generally ferrites become useless above VHF frequencies too. Ferrite baluns are extensively used with wire primary and secondaries to provide impedance matching, and sometimes for circuit isolation, (as in "hot" chassic TV sets de-coupling to the antenna system - I still get a laugh remembering a mate fixing the bosses TV with an earthed CRO lead, the expression on the bosses and his face when he blew the crap out of it . . . Hi Paul.)

RF Connectors

SMA Coaxial Connectors:

The SMA coaxial connectors are available in many configurations, printed circuit mount, panel mount, and flange mounting. Typical applications are telecommunications, data communications cellular networks, wireless communications, gps, gsm, umts and measurement and test equipment. The impedance is 50 Ohm. The outer body of the connectors are made from high-grade brass, gold or nickel plated versions are available. The inner (center) contacts are made in brass or beryllium copper and are gold plated. Several plating options are available. SMA coaxial connectors applicable standards: EC169-15, CECC 22110, MILL-3902

BNC Coaxial Connectors:

BNC connectors are the most prevalent in computer networks, audio, video, data processing and telecommunications equipment because of their size and relatively low installed cost. CONEC BNC Connector family is available with characteristics impedance of 50 Ohm and 75 Ohm. The 50 Ohm version has been designed for the frequency range 0 - 4 GHz 50 Ohm and the 75 Ohm version for 0 - 1 GHz. The termination dimensions meet IEC 169-8, MIL-C-39012, VG 95200 and CEEC 22120 requirements. Cable terminations are available in crimp, clamp, solder and jacket quick twist configurations. The two-stud bayonet lock coupling provides ease of connecting and disconnecting and is ideally suited for applications such as test equipment where this feature is notably desired.

SMP Coaxial Connectors:

The SMP series of coaxial connectors is designed for superior performance with semi-rigid cable. Performance with semi-rigid cable is from 0-18GHz and if used with an adapter extends the frequency up to 40GHz. The SMP coaxial connector is designed utilizing a snap-on mating system, allowing fast connect and disconnect, plus the miniature size allows for high density printed circuit board mounting. Standard configurations include cable connectors, straight and right angle types, suitable for semi-rigid cable .047 and .085, full detent, limited detent, including mating shrouds for bulkhead or flange mounting.

TNC Coaxial Connectors:

The TNC series of coaxial connectors are similar to the BNC type connectors. The exception being for their mating threaded coupling which is designed to provide low reflection from DC to 11 GHz under extreme environmental conditions such as shock and vibration. Cables can be terminated in crimp, clamp, twist on and solder. The 7/16"-28 thread coupling provides positive mating. TNC Coaxial connectors are designed for use in medical equipment, test and measurement, telecom and other types of communication systems TNC coaxial connectors are available in 50 Ohm 0-11 GHz and 75 Ohm impedance from 0-1.0 Ghz.

SMC Coaxial Connectors:

The SMC series of coaxial connectors features the following, 50 Ohm impedance 0-10Ghz and 75 Ohm impedance 0-2 Ghz operating frequencies. Adding features are the threaded coupling that provides secure mating in applications where vibration or shock are evident. SMC coaxial connectors are small in size, provide high performance at a reasonable cost. The series is ideal for design where high density packaging is required.

SMB Coaxial Connectors:

The SMB series of coaxial connectors features snap-on coupling mating of plug and socket connectors. This feature of quick connect and unmating is well suited for applications in test equipment, mobile telecommunication, hand held instruments and GPS systems. Due to the small size, the fast and easy unmating make the SMB coaxial connectors the choice where packaging density and ease of connecting and disconnecting a prerequisite. The SMB series is available in right-angle and straight PCB configuration. Semi-rigid or flexible cable can be terminated. Crimp, clamp and solder attachment are possible.

N Coaxial Connectors:

Type N coaxial connectors have been an industry standard series for many years. This series of coaxial connectors is designed to operate in 50 Ohm impedance from 0-11 Ghz and in 75 Ohm impedance from 0-1.5 Ghz. The type N coaxial connectors have been made and are widely in use in low broad band VSWR applications. The type N coaxial connectors is impedance matched to work with 50 Ohm coaxial cables. Crimp, clamp and solder types are available. The threaded coupling provides secure mating in high vibration applications. Type N coaxial connectors are suitable for broadcast, audio, video and aircraft industries. These coaxial connectors also find use in microwave applications as couplers, amplifiers, attenuators and filters.

Coaxial Connector Adapters:

CONEC RF/coaxial connector adapters are for use between series of coaxial connectors and are manufactured with the same high quality as the individual coaxial connectors. The adapters make it easy to convert from one series to another without having to make changes to the termination. Change over from plugs to jacks in either cable to cable or panel mount is achieved quick and easy.

Additional info :

The "Bayonet Neil-Concelman" or "Bayonet Navy Connector" or "Baby Neil Connector", depending on the information source. Karl W. Concelman is believed to have created the "C" connector. The BNC was designed for military use and has gained wide acceptance in video and RF applications to 2 GHz. The BNC uses a slotted outer conductor and some plastic dielectric on each gender connector. This dielectric causes increasing losses at higher frequencies. Above 4 GHz, the slots may radiate signals, so the connector is usable, but not necessarily mechanically stable up to about 10 GHz. Both 50 ohm and 75 ohm versions are available.

MCX connectors are sub miniature coaxial connectors with very good electrical performance used to address the rapid implementation of the U.S. digital cellular PCN infrastructure, Global Positioning Systems (GPS) and instrumentation and Wireless LAN Systems.

Mini UHF connectors are a miniature version of the original UHF connector and feature a threaded coupling mechanism for reliable mating. The mini UHF connector is designed for use in cellular mobile telephone systems where size, weight and cost are critical. Featuring crimp cable termination for low installation costs, these connectors provide excellent RF performance in applications through 2.5 GHz.

MMCX connectors provide a more robust interface for greater durability, this series is ideal for high volume wireless SMT or PCMCIA applications in cellular base stations, cellular phones and personal communicators, global positioning systems (GPS) and wireless LAN (WLAN) applications.

The SMA (Subminiature A) connector was designed by Bendix Scintilla Corporation and Omni-Spectra Corporation as the OSM connector, and is one of the most commonly used RF/microwave connectors. It is intended for use on semi-rigid cables and in components which are connected infrequently. It takes the cable dielectric directly to the interface without air gaps. A standard SMA connector is designed for interconnects to 12.4 GHz. Fortunately, a good SMA is usable to 18 GHz in most cables, and if well constructed with greater loss and lower return loss to 24 GHz.

The SMB (Subminiature B) connector is a snap-mount connector rated at 4 GHz, but usable to 10 GHz. The Mil standard is MIL-STD-348. Typical insertion loss is as much as 0.3 dB at rated frequencies, making this connector a poor choice in critical low noise microwave RF interconnects, but quite acceptable in other signal delivery applications. SMBs can be used up to 10 GHz without moding.

The SMP (Subminiature P) connector is rated to 40 GHz, and depending on detent type, it can withstand from 100 to 1,000 interconnect cycles.

TNC connectors are of miniature size like the BNC connector but feature a threaded coupling nut for applications requiring performance through 11 GHz. Chosen for their durability and reliability, TNC connectors are widely used in the cellular/mobile communication industry for equipment cabling and antenna interfaces.

Type F connectors have screw type coupling with a frequency range up to 1.5GHz used for CATV, TV, antennas and other similar applications.

The Type N 50 ohm connector was designed in the 1940s for military systems operating below 5 GHz. One resource identifies the origin of the name as meaning "Navy". Several other sources attribute it to Mr. Paul Neil, an RF engineer at Bell Labs. The Type N uses an internal gasket to seal out the environment, and is hand tightened. There is an air gap between center and outer conductor. In the 1960s, improvements pushed performance to 12 GHz and later, mode-free, to 18 GHz. Hewlett Packard, Kings, Amphenol, and others offer some products with slotless Type N outer conductors for improved performance to 18 GHz. Type N connectors follow the military standard MIL-C-39012. Even the best specialized type-N connectors will begin to mode around 20 GHz, producing unpredictable results if used at that frequency or higher. A 75 ohm version, with a reduced center pin is available and in wide use by the cable-TV industry.

UHF connectors are economical and all purpose connectors. They are designed with a non-constant impedance for use at comparatively low voltage and low frequency applications including all C.B. communications systems such as: Public address systems, CCTV, civil defense, landing systems, ground control apparatus, ship to shore communications and mobile radio equipment hookups between antenna and transmitter or receiver.

1.6/5.6 connectors are well-suited for both radio frequency analog and digital signals in modern telecommunications equipment. Due to their mechanical sturdiness and coupling mechanisms, these connectors are recommended in equipment where there is a particular need for resistance against environmental and mechanical stress.

7/16 connectors are designed for use in medium to high power communication systems. These connectors perform exceptionally well in multichannel cellular systems where power levels approximate 100 watts per channel. Designed for both flexible as well as corrugated cables, these connectors are used in a variety of cellular base station and broadcast communication applications.)

Other types of Antenna

Yagi - These are directional antenna and usually high gain, that consist of a driven element and an in-line parasitic array, they were first conceived by "Uda" and constructed by "Yagi". There is many variations employing different element types, and are very successfully employed in combinations of arrays, yielding very high gains

Loop Yagi - a yagi which operates at microwave frequencies. A high degree of accuracy is required in building these, and care taken in handling these. The parasitic elements consist of "loops", and because of the operating frequency, many elements can be employed in a small package to yield high gains..

Quads - A quad is another example of a driven element with an in-line parasitic array, but with the elements arranged as perpendicular squares of 1/4 a wavelength long to each side. This results in a more compact antenna because they don't usually employ as many directors (parasitic elements) as a yagi, and these are also successfully arranged in arrays as well.

Log Periodic - So called because it's structural format causes it's impedance and radiation characteristics to repeat periodically as the logarithm of frequency. It is a broadband antenna consisting of multiple elements attached to two booms with each boom supplying one side of elements. A variation to this is the "log periodic Yagi" which consists of a cross coupled log periodic section and the traditional "yagi" parasitic section to yield a broad bandpass antenna and sharp attenuation on both sides.

Corner Reflector - An antenna using a "V" shaped reflector to capture and reflect signals back to a dipole at it's focus. Because there are optimal wavelength dimensions required in it's construction it is usually only used only above VHF because of physical size. There are many variations to this antenna too.

Trough Reflector - To reduce the size of "corner reflectors", the vertex is cut off and replaced by a plane reflector. This is known as a trough reflector. It has the good gain of a corner reflector but without the size, and both theses types can be stacked in arrays, or employ collinear dipoles to increase the reflector width and gain.

Axial mode Helix - Provides wide band performance and high gain by having a spiralling element with a circumference of around 1 wavelength. The radiation pattern is circularly polarized with the helix having a right or left handed sense. (helix wound left of right thread) The helix operates above a plane reflector, and a typical 7 turn helix will have a gain of about 12dBi over a 2:1 frequency range..

Waveguide antennas - A waveguide antenna can be considered as a low loss transmission line with a coaxial probe and an aperture to capture and radiate signals. It could have an open end like the "pringle" antenna, or fully enclosed with slotted windows in an array pattern allowing emission of phased radiated energy. Very little energy is lost providing this type with high directivity and high gains. There are many variations. (Omnidirectional propagation is possible with slots in opposite sides)

Horn - A horn is usually fed by a waveguide. The horn is a broadband device and gains up to 25 dB are achieved without it getting too big. (optimal wavelength dimensions like the corner reflector). It is used for microwave frequencies and highly directional.

Parabolic Dish - By virtue of the parabola's shape, all the signal is captured by the dish's surface area and focused to a single point. Very high gains can be achieved, and the bigger the dish , the higher the gain. Different frequency operation is accommodated by the design of the dipole and feed at the focus.

Discone - This is an omnidirectional antenna (all the above are unidirectional). It has a very wide bandwidth, 10:1. It consists of a disc mounted above a cone but very practical versions have been built from a "skeleton" of elements that approximate this. It is not widely used for Amateur radio use because of it's ability to radiate any transmitter harmonics equally as well. It is very tolerant of bad weather conditions and even operates well covered in ice.

Types of Coax

Some Links:

Brismesh - This is an organisation set up to further wireless networks within and around Brisbane Australia. (many good links)

Wi Fi Antennas - The DXZONE provides some andtennas and info.

A cheap Wi-Fi Collinear antenna. - Another "wire" collinear using 3/4 wave sections over 1/2 wave.

2.4 Ghz Collinear using brass sections - Is a very good article covering everything in detail, from Captain Kaboom and Brian Oblivion

EliteGeek.org - Yep the rodent is on air. . . . , many interesting up to date write ups on the PC computer industry

TEARAS site - Lots of info for 2.4ghz ATV.

Collinear - a good article from QST.net, on how to build a UHF / CB collinear.

Poor Mans Wi Fi Dish - a good article using a Dick Smith USB wireless adaptor and a cheap chinese wok (the dish 15dB and 3-5 Km).

FRARS - an interesting group with a lot of info and projects on 2.4Ghz.

Waveguide antennas - Explanations and theories on waveguide ants, and how to build some. - very good

2.4 Ghz collinear - A nice looking design. (more links stepping back into their main site)

Wireless.org - Has many good articles on their site on 2.4 Ghz antennas (helical too) as well as useful links for other wireless wannabes

Cubical Quad Antenna - A interesting 2.4Ghz antenna project.

Back to projects page