Building HF wire antennas from scratch: wire, insulators, chokes, ladder line, radials, and checking the build in EZNEC or MMANA
- Replies 2
- Views 29
- Created
- Last Reply
A better way to browse. Learn more.
A full-screen app on your home screen with push notifications, badges and more.
This thread is for the practical details of a homebrew antenna: which materials hold up, where a build tends to go wrong, and how to check the result with an analyzer and a model. Take the examples as starting points, because every site changes the numbers.
Start with the dipole, since most other wire antennas are variations on it. The usual formula for a half-wave dipole is 468 divided by the frequency in MHz, which gives feet. For 7.1 MHz that is about 65.9 ft overall. Treat it as a first cut, because height, nearby trees, wire type and insulation all shift the real resonance. Cut a little long and trim. On 40 m, shortening each leg by one inch moves resonance up roughly 15 to 20 kHz. Insulated stranded wire usually resonates a few percent lower than bare wire, so allow for that before cutting.
For wire, 14 AWG hard-drawn copper is a common choice for long spans because it resists stretching. Copperweld (steel core, copper cladding) is stronger still and suits long unsupported runs, though it is stiffer and can kink if you coil it tightly. Thinner wire such as 18 to 20 AWG is fine for portable or short-span use, but it stretches, and the antenna slowly detunes as it sags. Larger-diameter wire gives slightly more bandwidth, but the difference is small next to the effect of height and surroundings.
The center insulator and end insulators carry the mechanical load, so build them for strain relief first and electrical performance second. The coax or feedline should not hang from the solder joints. Use a UV-stable center block with a hole for a rope or a strain loop, and add a drip loop so water follows the cable away from the connector. Seal the coax end, since water wicks into braid. At the ends, small dogbone-style insulators are enough for the low tension of a sagging dipole. If the antenna is hung on a tree, leave slack or use a pulley and counterweight, because trees move and will pull a rigid wire apart.
A balun or choke matters more than most beginners expect. A dipole fed with coax is balanced at the feedpoint, but the shield can carry common-mode current and become part of the radiating system. That shows up as RF in the shack, SWR that changes when you touch or move the coax, and a distorted pattern. A 1:1 current balun at the feedpoint is the standard fix. Alternatives are a stack of ferrite beads on the coax or a coiled-coax choke. Mix 31 is usually recommended for the lower HF bands and mix 43 for higher HF, but the best check is to measure the choking impedance rather than trust a rule of thumb. For a coiled-coax choke, remember it is narrowband, since its self-resonance sets the useful range.
Ladder line gives you a flexible multiband option. A center-fed dipole on 450-ohm window line, driven through a balanced tuner or a tuner with a good 1:1 current balun, works on several bands without traps. The line itself has low loss even at high SWR, and the losses mostly come from the tuner and any balun. Keep the line away from metal, avoid tight bends, and bring it away from the antenna at right angles for a few feet before routing it down. If it goes near gutters or the mast, the line becomes unbalanced and RF finds its way into the shack.
Fiberglass poles work well as a support or a vertical mast. Avoid carbon-fiber fishing poles for anything near the antenna wire, because carbon fiber is conductive and detunes things in unpredictable ways. A telescoping 10 m (33 ft) glass pole can hold a vertical wire, or act as a center support for an inverted-V. The poles bend, so allow for wind loading, and use a ground sleeve or a tilt-over base if you plan to raise and lower it often.
Traps let one antenna cover several bands, but they have costs. A trap is a parallel LC circuit resonant at the band edge, so on that band it blocks the outer section, and on lower bands it acts as a loading inductor. Traps add loss, narrow the bandwidth, and shorten the antenna. They are also a weak point for water and heat. Fan dipoles and linked dipoles avoid traps. They are simpler and easy to model, but they need more wire and interact with each other. Check the trap resonance with an analyzer or a grid-dip meter before you cut the antenna to length.
For verticals, a quarter-wave is about 234 divided by the frequency in MHz, in feet. On 14.2 MHz that is roughly 16.5 ft. The radial system is what makes or breaks it. A ground-mounted vertical needs many radials, often 16 or more, because the ground carries return current and its loss takes power. Elevated radials need fewer, typically two to four resonant ones, and sloping them downward raises the feedpoint impedance toward 50 ohms. Many short radials generally beat a few long ones. Add a choke at the feedpoint as with a dipole.
On measuring: an SWR reading is not the same as efficiency. A dummy load has 1:1 SWR and radiates nothing. Measure at the antenna end where you can, because a long, lossy coax run makes the SWR at the radio look better than the SWR at the feedpoint. An antenna analyzer gives resonance and impedance directly, which is more useful than a single SWR number. Low analyzer output makes it sensitive to strong nearby broadcast signals, so expect an erratic reading if you live near a transmitter.
For modeling, EZNEC and MMANA-GAL are both useful, but they do not agree on everything. EZNEC uses NEC, with options for real ground and detailed wire-loss settings. MMANA-GAL uses a MININEC-based engine and is quick for first sketches. In either one, segment the wire carefully, at minimum around ten segments per half-wavelength for a simple antenna, and avoid sharp changes in segment length at junctions. Set ground type and height correctly, because feedpoint impedance depends strongly on both. Check the model's average gain test; a value far from 0 dB means the geometry or segmentation is wrong. A model gets you close, not exact, so use it to choose the starting length and then trim the real antenna.
Which parts of your own builds have caused the most trouble: support structure, choke performance, radial layout, or getting the model to match the analyzer?
Link to comment
https://www.hamradiobase.com/forums/topic/11462-building-hf-wire-antennas-from-scratch-wire-insulators-chokes-ladder-line-radials-and-checking-the-build-in-eznec-or-mmana/Share on other sites