9:1 Unun for Random Wire Antennas
A 9:1 unun lets you feed a random-length, non-resonant wire directly with coax by stepping the wire's widely varying feedpoint impedance down toward 50Ω, without needing to prune the antenna to resonance on every band. It is an autotransformer, not an isolation transformer, so it does the impedance work but relies on a counterpoise and (usually) a separate common-mode choke to keep RF off the coax shield. This guide covers the trifilar winding method, why 9:1 specifically was chosen as the standard random-wire ratio, a turns calculator, and full troubleshooting.
Why "random wire" needs an averaging transform, not a precise one
A random or long wire antenna is deliberately not cut to resonance on any particular band — its feedpoint impedance swings widely across HF, from under 100Ω to several thousand ohms, depending on frequency and wire length relative to wavelength. There is no single "correct" transform ratio that matches every band perfectly. The 9:1 ratio became the de facto standard because it targets roughly 450Ω (50Ω × 9), a value that sits in the middle of the impedance range a typical 29-84 ft random wire presents across 80-10m — close enough on most bands that an antenna tuner at the shack end can clean up the remaining mismatch, rather than presenting the tuner with a wildly reactive load it cannot match at all.
How the trifilar winding creates the 9:1 ratio
A 9:1 unun is built as an autotransformer: three identical wires are wound together (trifilar) through the core for N turns, then connected in series to form one continuous winding of 3N turns. The coax center conductor connects to the start of the winding; the coax shield and counterpoise both connect to the tap point after the first N turns; the antenna wire connects to the far end, after the full 3N turns. The coax side sees N turns, the antenna side sees the full 3N turns — a 3:1 turns ratio, which squares to a 9:1 impedance ratio.
Unun vs. balun — why a counterpoise is not optional
A balun (balanced-unbalanced) isolates its two sides and forces balanced current in a balanced load. An unun (unbalanced-unbalanced) does not isolate anything — the winding is a single tapped autotransformer, and the "return" side of the antenna circuit is whatever is connected to the coax shield/tap point, which on a random wire setup is a counterpoise, not a balanced second element. Skip the counterpoise, or make it too short, and RF has nowhere good to return except the coax shield itself, which shows up as SWR that shifts when you touch the coax, "RF in the shack," and erratic tuning. See the Counterpoise Design guide for sizing and placement.
Sizing for power and choosing a core
Because the unun handles real transformation current at whatever impedance the antenna presents on a given band — sometimes far from the nominal 450Ω design target — the core needs margin the same way a balun's does; see the Core Selection Guide for the underlying flux density math. A single FT-140-43 or FT-240-43 core with 3-4 trifilar turns handles most QRP-to-100W random wire setups comfortably; step up to a stacked FT-240-43 pair for full legal limit, especially if you expect to run into bands where the wire presents a poor match.
| Core Configuration | Primary Turns | Secondary (Total) Turns | Recommended Power |
|---|---|---|---|
| FT-82-43 (single) | 4 | 12 | Up to 50W (QRP/portable) |
| FT-140-43 (single) | 3 | 9 | Up to 250W |
| FT-240-43 (single) | 3 | 9 | 250–500W |
| FT-240-43 (stacked x2) | 4 | 12 | Full legal limit (1.5kW) |
9:1 Unun Turns Calculator
Materials for a trifilar-wound 9:1 unun
A 9:1 unun wound with 3 trifilar turns on a single FT-240-43 core, mounted in a weatherproof enclosure with SO-239, antenna terminal, and counterpoise terminal.
Building a 9:1 Unun
Budget 1-2 hours including enclosure assembly and testing. Plan the counterpoise before you finish the build — a 9:1 unun without one will not perform as tested.
Prepare and wind the trifilar set
Cut three equal lengths of wire, leaving extra at both ends for connections. Lay all three side by side and wind them together through the core for the target turn count, keeping the three wires in the same order on every turn. Mark each wire's ends with a different color of tape before winding — three unmarked wire ends coming out of a toroid are very easy to mix up.
Connect the autotransformer topology
Connect wire A's end to wire B's start, and wire B's end to wire C's start — this creates one continuous winding of 3N turns. Coax center conductor connects to wire A's start. Coax shield connects to the junction between wire A's end and wire B's start (the 1N tap point) — this same junction also connects to the counterpoise terminal. The antenna wire connects to wire C's end, the far end of the full winding.
Test with a dummy load before installing
Connect a non-inductive resistor across the antenna and ground terminals matching the design target (450Ω is standard for testing a 9:1 unun) and sweep SWR at the coax input with a NanoVNA. A correctly wound unun shows SWR under about 1.5:1 to 1.8:1 across HF when loaded this way — a much higher or wildly varying reading points to a winding or connection error, not a bad antenna.
Weatherproof, install, and connect the counterpoise
Mount the tested unun in its enclosure with sealed cable entries and a drip loop below the connector. Support the enclosure's weight independently of the terminal connections. Before transmitting, connect a counterpoise of adequate length to the ground terminal — see the Counterpoise Design guide — the unun's performance in the field depends on this as much as on the winding itself.
| Symptom | Most likely cause | Diagnosis | Fix |
|---|---|---|---|
| SWR is high on all bands even with a reasonable wire length | Winding topology error, or no counterpoise connected | Test the unun alone with a 450Ω dummy resistor across the antenna/ground terminals; a correctly wound unun shows SWR under 1.8:1 into that load regardless of counterpoise | Re-check winding connections against the topology described above; if the dummy-load test passes but field SWR is still bad, add or lengthen the counterpoise |
| SWR looks fine but RF-in-the-shack symptoms persist | Counterpoise too short, or common-mode current on the coax shield with no choke in the line | Clip on a known-good choke temporarily at the shack entry; if symptoms disappear, common-mode current on the shield was the cause, not the unun's transform | Lengthen or reposition the counterpoise, and add a dedicated 1:1 current choke at the shack entry as a permanent fix |
| SWR is sensitive to touching the coax or moving the counterpoise wire | Counterpoise is acting as part of the antenna system rather than a stable reference — often means it is too short or poorly routed | Confirm counterpoise length and routing against the Counterpoise Design guide's recommendations for your bands | Reroute the counterpoise away from the feedline and other conductors, or switch to a longer or multiple-length counterpoise setup |
| Core runs hot or cracks under power | Core undersized for actual power and the impedance mismatch seen on a specific band | Recalculate flux density per the Core Selection Guide using the actual RF voltage at the unun on the problem band, not just rated transmitter power | Step up to a larger core or add a second stacked core, especially for bands where the wire presents a poor match to 450Ω |
| Tuner cannot find a match on one particular band | Wire length happens to present an impedance far outside what a 9:1 transform and tuner combination can reasonably handle on that band | Check whether the wire length is near a multiple of a half-wavelength on the problem band, which can produce an extreme, hard-to-match impedance even after the 9:1 step-down | Adjust overall wire length by a foot or two to shift away from the problem resonance, or accept that band as a weak spot for this specific wire length |
Why 9:1 instead of 4:1 or 49:1 for a random wire?
9:1 targets roughly 450Ω, which sits closer to the middle of the impedance range a typical 29-84 ft random/long wire presents across 80-10m than either 4:1 (targeting ~200Ω) or 49:1 (targeting ~2450Ω, correct for a resonant half-wave end-fed, not a random length). The goal is not a perfect match on any one band, but a starting impedance an antenna tuner can clean up on most bands.
Do I really need a counterpoise, or can I skip it?
You need one. Unlike a balun, a 9:1 unun does not isolate the antenna circuit from the coax shield — the counterpoise is the return path the transform depends on. Skipping it, or using a counterpoise that is too short, typically shows up as SWR that shifts when you touch the coax and persistent RF-in-the-shack symptoms even with an otherwise-correct build.
Can I use a 9:1 unun on a wire that happens to be resonant?
Yes — a 9:1 unun does not require a non-resonant wire, it just does not require resonance either. If the wire happens to be close to resonant on a given band, the 9:1 transform still applies; you may simply find SWR is already low on that band without much tuner assistance.
Does wire length still matter if the unun does the matching?
Yes. The 9:1 ratio only sets a fixed transform — it does not change what impedance the wire presents at a given frequency. Certain lengths (particularly those near a multiple of a half-wavelength on a given band) can still produce an impedance extreme enough that even a 9:1 step-down plus tuner cannot find a usable match, so length still deserves some planning even in a "random wire" setup.
Do I need a separate common-mode choke, or does the unun handle that too?
Plan on a separate choke. The unun's job is impedance transformation, not common-mode suppression — its single tapped winding does not present the kind of choke impedance a dedicated ferrite choke does. Most working random-wire setups use the 9:1 unun for the transform, a proper counterpoise for the return path, and a separate 1:1 current choke at the shack entry as a second line of defense against common-mode current.