Back to main page

(Description last updated 2009-03-31)

This vertical can be used on all the HF bands, efficiency is good on 40m and up, less on 80m and of course very low on 160m.

Status: Built and erected in May 2008, currently in use.

Recently, when moving into a rented upper story of a duplex, I was faced with the challenge of putting up an antenna that would not offend my new landlord too much. While it has been demonstrated that elevated vertically polarized antennas perform worse than their horizontally polarized counterparts (see ARRL Antenna Handbook, 19th ed, chapter: "The effects of Ground"), one's choices may be limited by practical circumstance. Also, lately I have been able to compare the antenna described in this article to a full size horizontal 1/2 wave 20m dipole, elevated 9m (30'). On average I do not notice much of a difference between the two antennas. Sometimes one is better, sometimes the other, based on the polarization and direction/elevation angle of the incoming signal.

Back to the subject at hand. LB Cebik, W4RNL (SK) and others have demonstrated that loaded verticals in a Double T configuration can be considerably shortened without significant loss of performance. Modeling with EZNEC confirms this.

As my only choice of mounting an outdoors antenna was to attach it to a balcony banister, my best choice seemed to be a self contained antenna that wouldn't need a counterpoise, this seemed to be worth exploring further.

A market search revealed two antennas being sold which were based on this exact approach. Both antenna makers sold their antennas with a multiband option, utilizing a reasonably efficient switchable hairpin match. However, when studying these antennas, a few thoughts struck me:

• • • Firstly the very high price, I would surely be able to build myself something similar for a fraction of the price these antennas were being sold at.

    • Secondly, why center feed the antenna and have an unsightly coax hang down from the middle of it. One manufacturer tries to explain this at length, concluding that this gives some extra dBs. However the logic appears to be seriously flawed. Offset feeding the antenna should yield exactly the same results as a symmetrical feed, if the antenna is fed and matched to properly (balanced + impedance match). A double T antenna was modeled in EZNEC, with both centre feed and offset feed, confirming that they were indeed exactly the same. Moreover, while the centre feed would yield an impedance considerably lower than 50 ohm (the lower the smaller the antenna is), an offset feed would yield a higher impedance, closer to 50 ohm, an extra added bonus.

    • Thirdly, while walking through the local hardware store and taking inventory of the aluminum tubing they had available, I realised that a double-T vertical antenna construction might be impractical as it was rather top-heavy, why not make the whole antenna assymmetrical, that it would be better to trim down the wind load of the top section. The following design shows the end result, guided by the material I found in the hardware store. Again, despite some of the writeup referred to above, it has no effect on its effectiveness whether the antenna is center fed and symmetrical or "bottom" fed and assymmetrical in its dimensions. This can easily be demonstrated by playing with the dimensions in EZNEC or other antenna modeling software. What matters is the antennas overall electrical dimensions, the characteristics of the ground (reflecting the radiowaves, thus introducing losses) and the elevation of the antenna and its current maximum point above the ground.

(click on picture to enlarge)

The design was based on a couple of basic criteria. Resonance at or slightly above 20m, to be compensated by a hairpin match if/as necessary, and elevation of antenna not to exceed 4m. Dimensions of the antenna were decided otherwise by the length of the material purchased (Aluminum tubes came in lengths of 8 feet and aluminum rods in lengths of 4 feet). This design was modeled in EZNEC, which calculated the resonance frequency to be 16.8 MHz with the material and dimensions selected. This suited me fine. With these dimensions EZNEC predicted no performance loss of consequence on 20m when compared with a full size dipole with its current maximum (center) at the same elevation above ground, and quite acceptable gain on 40m as well. Using EZNEC again, I designed a hairpin match for 20 and 40m:

(click on picture to enlarge)

Here are the EZNEC predicted radiation patterns for 20m and 40m:

(click on picture to enlarge)

(click on picture to enlarge)

To try to make a long story approach its end, the antenna was assembled for a total cost of probably less than $100, spray painted black for less visibility to the neighbours and erected after dark :) The antenna has now been up for 10 months, or since late May 2008.

(click on picture to enlarge)

(click on picture to enlarge)

On this image, the Balun used is faintly visible, RG-58 wound through a Ferrite Toroid. Hairpin match was not used, see below.

To put an end to this rather too long story: The antenna was erected. Its resonant frequency was measured to be quite a bit lower than expected, or just below 20m. (Hairpin match depicted above not useable :) This is most likely due to the capacitive coupling in the aluminum angle bracketing method used to hold the vertical and the horizontal elements as well as the base of the antenna together. Because of capacitive coupling into the base, the balcony banister probably participates in the antenna to some degree. The design could obviously be improved, but I very much doubt that this has a significantly detrimental effect on the performance of the antenna. A few centimeters were cut off of each end of the horizontal section for best resonance at 14 MHz. Due to my inherently lazy nature, I ended up mounting an autotuner below the antenna, rather than building a band switchable matching circuit. With the autotuner, the antenna can be tuned from 160 through 6m, efficiency is of course degraded at 80 and 160m, but the antenna appears to be quite usable on 80m.

An SWR measurement on 20m through an inactive autotuner and 10m of RG58 is shown here:

(click on picture to enlarge)

This measurement was taken today (2009-03-21) with the autotuner in the circuit, but turned off. Note that I couldn't find the original measurement from 10 months ago, taken without the autotuner in the circuit. Although inactive, the autotuner appears to worsen the SWR noticeably in the 20m band where it is not needed. If my memory isn't too far off, the original measurement indicated an SWR of about 1.2 to 1.4 across the 20m band. In any case, as can be seen, the bandwidth of the antenna is very high. This is not due to any lossy elements, but due to the top and bottom loading incidentally having the makings of a decent emulation of a wide band conical antenna (disc cone, duo cone). Not a bad bonus.

Back to main page