From-scratch wire antenna checklist: wire, insulators, choke, feedline, and checking it with a VNA and a model
- Replies 2
- Views 28
- 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 working through the decisions in a homebrew wire antenna, using a 40 m inverted-V or flat-top dipole as the example. The same reasoning carries over to a doublet fed with ladder line and to a quarter-wave vertical. Please push back on anything here that looks wrong.
Starting length. The usual formula for a half-wave dipole is about 468 / f(MHz) feet in total length. For 7.1 MHz that is roughly 66 ft. Treat it as a cut-long starting point, not a final answer. Height above ground, nearby trees and wires, the wire's insulation, and the inverted-V angle all pull the resonant frequency around. Insulated wire typically resonates a few percent lower than bare wire of the same length. Cut each leg longer than the formula says, leave extra at the ends, and trim after measuring.
Wire. For a span of about 66 ft, 14 AWG is a common choice and 12 AWG gives more margin on long spans. Stranded copper is flexible and easy to handle. Hard-drawn solid copper is strong but kinks, and kinks become break points. Copper-clad steel (Copperweld) stretches much less under tension, so the antenna stays on frequency, though it is stiffer and has more skin-effect loss at lower frequencies. Plain soft copper can creep and lengthen over time under load, which slowly lowers the resonant frequency. Many builders use a copper-clad steel or steel-core stranded wire when the antenna will be under tension for years.
Center and ends. Most dipole failures are mechanical. At the center insulator, loop the wire through the holes and twist or clamp it back on itself so the tension is carried by the wire and the insulator, not by a solder joint. The joint then only carries current. The coax should have its own strain relief, and its weight should not hang on the connector. At the ends, use an insulator with a smooth radius and pass the rope through it without a sharp bend. UV-resistant cord such as polyester or a polyester-jacketed rope is a reasonable choice, and nylon will stretch with moisture and temperature. For long stretches of rope between the insulator and a tree, a pulley and counterweight, or a spring or bungee in the line, lets the tree move without snapping the wire.
Choke or balun. A dipole is balanced and coax is not, so without a choke the outside of the shield can carry common-mode current and radiate, pick up noise, and shift the SWR reading. A current balun at the feedpoint, or a common-mode choke made from coax turns on a ferrite core, addresses this. A typical homebrew approach is several turns of coax through a suitable ferrite mix (31 and 43 are the usual ones for HF) sized for the power level. Air-core coiled coax works too, but its useful bandwidth is narrower. A quick check is to see whether the SWR or the reading changes when you touch or move the coax. If it does, you have common-mode current and the choke needs more impedance.
Ladder line and doublets. A center-fed doublet with open-wire or window ladder line avoids most feedline loss even when the SWR on the line is high, because the line itself is low-loss. The tradeoff is that you need a tuner that can match a balanced load. Keep the ladder line away from metal and run it away from the antenna wires at right angles for as far as you can. Avoid coiling it or taping it to a mast. Ladder line is mechanically more sensitive than coax, so route it to avoid twists and sharp bends.
Verticals and radials. A quarter-wave vertical is about 234 / f(MHz) feet long (roughly 16.5 ft at 14.2 MHz). The vertical element is only half the antenna, and the radial system is the other half. A few resonant elevated radials, sloped downward, can work reasonably well. Ground-mounted radials need many more, with cut-to-length less critical. Elevated horizontal radials at a quarter-wave give a feedpoint impedance near 36 ohms, and drooping them toward 45 degrees raises it toward 50 ohms. A fiberglass pole is a convenient mast for a wire vertical because it is non-conductive and light, but it flexes in wind, so use guying and keep the wire from taking the load.
Traps. A parallel LC trap is resonant near the higher band, which isolates the outer part of the element there. On the lower band it acts like a loading inductor, which shortens the element. Traps are lossy compared with separate wires, narrow in bandwidth, and sensitive to weather and capacitor quality. Capacitors must be rated for the voltage across them. A fan dipole (separate wires for each band on one feedpoint) is a simpler alternative that avoids trap loss, at the cost of some interaction between the wires.
Measuring. A SWR meter at the transmitter reads through the coax, so feedline loss makes the antenna look better than it is. A dummy load also shows a perfect SWR, so low SWR alone says nothing about radiation. A VNA such as a NanoVNA, connected at the feedpoint or through a known length of coax, lets you see the impedance (R and X) rather than only a ratio. Resonance is where the reactance crosses zero, and the SWR minimum may sit slightly off from it. If resonance is below your target frequency, the antenna is too long. If it is above, it is too short. Trim in small steps, and expect the 2:1 SWR bandwidth on 40 m to cover only part of the band.
Modeling. EZNEC (NEC-based) and MMANA-GAL (a different method-of-moments code) are both useful for checking a design before cutting wire. Set the wire diameter, height, and ground type accurately, and use enough segments, since too few makes the results unreliable, especially near junctions and ends. The free EZNEC demo has a segment limit, so check that it suits your model. A model tells you trends: how resonance moves with height, what the pattern looks like, how the feedpoint impedance changes with an inverted-V angle. It won't replace the measurement, because real ground, nearby objects, and wire insulation are never exactly as modeled.
Which parts of this do you think are overdone, and which are missing? In particular, what would you change about the choke and the support design?
Link to comment
https://www.hamradiobase.com/forums/topic/11304-from-scratch-wire-antenna-checklist-wire-insulators-choke-feedline-and-checking-it-with-a-vna-and-a-model/Share on other sites