From-scratch wire antenna build checklist: materials, feedpoint choke, supports, and checking the result with an analyzer and a model
- Replies 2
- Views 48
- Created
- Last Reply
Top Posters In This Topic
-
Mod_Tom 1 post
-
Mod_Chrissy 1 post
This thread is meant as a working checklist for building HF wire antennas from raw materials, plus a place to argue about the tradeoffs. The example is a center-fed dipole for 40 m with a possible upgrade to multiband, but most of it applies to inverted-Vs, fan dipoles, doublets and quarter-wave verticals too.
Wire. Hard-drawn copper or copper-clad steel in 14 AWG is the usual compromise for a permanent span. It is strong and stretches little, and the resistive loss is negligible at dipole lengths. Stranded insulated 14 to 18 AWG is easier to handle and coil, but the insulation lowers the resonant frequency, so expect to cut it a few percent shorter than the bare-wire figure. For a first cut, a half-wave dipole is roughly 468 / f(MHz) feet. At 7.1 MHz that is about 66 ft overall. Cut long, then trim. Height, nearby trees, sloping ends and wire type all move the resonance, so the formula is a starting point, not a final length.
Insulators and supports. Anything that survives UV and does not absorb water works: commercial ceramic or molded dogbones, cut HDPE or acrylic, even PVC offcuts. Mechanical strength matters more than electrical rating at HF power levels. For end supports, UV-stabilized polyester rope is the usual choice, and cotton or cheap polypropylene will rot or fray. Trees move, so a pulley with a counterweight or a length of bungee at one end keeps the wire from being pulled apart in wind. At the center, the coax weight should never hang on the connector. Use a center insulator with a strain-relief hole or a short tie-off loop so the pull goes to the wire and rope, not the SO-239.
Feedpoint choke. Feeding a balanced dipole directly with unbalanced coax lets current flow on the outside of the shield. That distorts the pattern, can put RF in the shack, and makes SWR readings depend on how the coax is routed. A 1:1 current balun, really a common-mode choke, at the feedpoint prevents it. Two common builds are a stack of ferrite beads over the coax, or a number of turns of coax wound through a large toroid or on a former. Mix choice matters. Type 31 is generally favored for the lower HF bands, and type 43 does better higher up. The goal is a common-mode impedance of at least a few kilohms across the band you care about, not a particular number of turns. If you have an analyzer, you can measure it with a simple test jig. For a doublet fed with ladder line, the 4:1 balun at the tuner end is a separate question from choking the feedline.
Ladder line. 450 Ω window line into a balanced tuner (or an unbalanced tuner with a good balun) lets one doublet cover many bands with relatively low feedline loss, even when the antenna is far from resonance. Keep it clear of metal gutters and downspouts, avoid tight bends, and bring it away from the antenna at a right angle for a few feet. Certain feedline lengths on certain bands produce very high voltage or very high impedance at the tuner, so changing the line length by a few feet can fix an otherwise unmatchable band.
Fiberglass poles. Telescoping fiberglass masts are an easy way to hold the center of an inverted V or to support a vertical wire. Collapse risk is at the top sections, which are the thinnest, so do not load them with a heavy balun and a lot of coax. Clamp or tape the sections where they overlap, never use one near power lines, and remember that it holds a wire antenna, not a guyless tower.
Verticals and radials. A quarter-wave vertical is about 234 / f(MHz) feet, so roughly 33 ft on 40 m. On the ground, performance depends heavily on the radial field. Many shorter radials usually beat a few long ones, and a field of a few dozen is common practice. Elevated radials are a different tradeoff. Four resonant radials can do well, and sloping them downward raises the feedpoint impedance toward 50 Ω. A vertical still needs a feedline choke, since the radials are not the same as a perfect ground.
Tuned versus trap antennas. Trap dipoles use parallel LC circuits to isolate outer sections at the higher band, which makes the antenna shorter but adds loss and narrows bandwidth. The inductor Q, capacitor voltage rating and weatherproofing all matter. A fan dipole, with separate wires for each band on a shared feedpoint, is easier to build and tune, though the elements interact. Tune from the highest band down, because trimming a lower-band element has less effect on the higher bands than the other way around.
Measuring. An antenna analyzer at the feedpoint, or at the end of a known length of coax, is better than an SWR meter at the transmitter. Look at resistance and reactance, not just SWR. The frequency where reactance crosses zero is the resonant point, and the SWR minimum can sit slightly off it. Coax loss can make a poor antenna look good, since the loss hides reflected power. And SWR is not efficiency. A dummy load reads 1:1.
Modeling. EZNEC uses a NEC engine and handles real-ground effects more credibly than MMANA-GAL, which is MININEC-based, free, and very convenient for quick shape and length work. The free EZNEC demo has a segment limit, which constrains complex models. Common mistakes include too few segments, mismatched segment lengths at junctions, leaving the ground at perfect when it should be real, and forgetting to set wire loss. A model will not tell you your absolute SWR with insulated wire and a nearby fence. It is best used to compare options, like height changes, sloper angle, or radial count, and then you confirm on the real antenna.
Where do people here draw the line between building it once and trimming, versus modeling first? And what is your preferred choke for a feedpoint on the lower bands?
Link to comment
https://www.hamradiobase.com/forums/topic/11232-from-scratch-wire-antenna-build-checklist-materials-feedpoint-choke-supports-and-checking-the-result-with-an-analyzer-and-a-model/Share on other sites