Windom and OCF Dipoles: Design the Return Path, Not Just the Ratio
Windom and OCF Dipoles: Design the Return Path, Not Just the Ratio
A historical Windom, a coax-fed off-centre-fed dipole and an offset wire with a deliberate return branch are different antenna systems. Compare their complete currents before choosing the transformer or claiming an advantage.
The criticism behind this article is still important: a box labelled “Windom” or “4:1” can hide an accidental third conductor. If the coax exterior carries an uncontrolled current, the installed antenna includes the feed line, mast, station wiring and whatever completes that return path. A comfortable SWR does not make those currents disappear.
Joeri’s position: separate impedance transformation from common-mode control. In practical amateur installations I generally prefer a 4:1 UNUN followed by a separately specified choke when the design uses an unbalanced transformed port or an intentional return branch. A genuinely balanced OCF dipole may instead suit a measured current-balun solution. The installed current path—not the antenna name—decides.
The Original Windom Is Not a Coax-Fed OCF Dipole
Loren Windom’s 1929 antenna used a single-wire feeder connected off centre on a horizontal wire. That feeder was part of the RF structure; it was not a shielded, non-radiating transmission line. The station ground and surrounding conductors participated in the return path. Calling every modern off-centre-fed wire a “Windom” therefore hides the first architectural question: is the feed conductor intended to radiate, or is it intended to carry only a bounded transmission-line mode?
The antenna commonly sold today under the Windom name is usually a coax-fed off-centre-fed dipole. It has two unequal physical arms and a coax transition at the feedpoint. Its intended antenna current can still be differential: current entering one arm returns through the other arm. The unequal arm lengths do not by themselves prove that the radiator is electrically unbalanced to its environment.
The installed system may nevertheless become unbalanced because the two arms see different heights, ground, trees, gutters, masts and buildings. The coax may leave alongside one arm, or transformer capacitance may couple one terminal more strongly to the feed line. Those asymmetries can convert wanted differential energy into current on the coax exterior.
| Architecture | Intended RF path | Boundary that must be declared |
|---|---|---|
| Historical single-wire Windom | Horizontal wire plus a single radiating feeder and station-side return system | The feeder is part of the antenna; routing and station ground affect the system |
| Coax-fed OCF dipole | Differential current in the two radiator arms; coax centre and inner shield surface carry the feed mode | A balun or choke must impede unwanted current on the shield exterior |
| Offset-fed unbalanced system | Main conductor plus a deliberate counterpoise or declared coax-exterior segment | A choke defines where the intentional return branch ends and the feed line begins |
Off-Centre Feeding Changes Impedance, Not Kirchhoff’s Law
Moving the feedpoint away from the current maximum of a half-wave dipole usually raises the feedpoint resistance at that frequency. It does not create a fixed 200 Ω source, and it does not exempt the feedpoint from equal-and-opposite differential current. The installed impedance is complex and changes with frequency, wire length, feed position, conductor diameter, height, ground and nearby objects.
That is why “33/67 plus 4:1” is not a universal electrical recipe. On a multiband wire, each band samples a different standing-current distribution. The same physical feedpoint can sit near a useful resistance on one band, near a strong reactance on another and near a current minimum or maximum on another. A nominal 4:1 transformer scales the complex impedance within its own limits; it does not remove reactance or guarantee 50 Ω.
The upper bands also develop more current regions and a more fragmented far-field pattern. A tuner can make the transmitter see a match, but it cannot restore a lobe that the wire geometry does not place toward the desired path.
Transformation and Choking Are Different Jobs
A transformer ratio addresses the wanted differential or unbalanced transfer path. A common-mode choke inserts impedance into the unwanted exterior-current loop. One measurement cannot establish both functions.
- A 4:1 voltage transformer can provide impedance transformation, but equal terminal voltages do not guarantee equal and opposite arm currents when the arm-to-environment impedances differ.
- A 4:1 current balun can transform impedance and present common-mode impedance at a balanced port. Its balance and choking remain finite and load dependent.
- A 4:1 UNUN plus a separate 1:1 choke divides transformation and common-mode control into two networks. That is useful when the transformed port is deliberately unbalanced or when the installation no longer preserves textbook balance.
None of those labels is a verdict. A well-designed current balun can be correct for a balanced OCF dipole. A poorly designed UNUN followed by a weak choke is still poor. The useful advantage of the separate architecture is that transformer transfer, choke impedance, placement and thermal stress can be specified independently.
A choke cannot delete a conductor that the design needs. If the coax exterior is the intended counterpoise, placing the choke at the transformer removes that branch. If the coax exterior is unwanted feed-line current, placing the choke far away leaves a radiating section between transformer and choke. First decide which system you are building.
The Choke Position Defines the Antenna Boundary
For a coax-fed OCF dipole intended to use only the two wire arms, the starting boundary is usually at the feedpoint: the feed line should depart away from the wire and a suitable common-mode impedance should be placed where exterior current would otherwise begin.
For an offset-fed design that deliberately uses a separate counterpoise or a length of coax exterior, the choke belongs at the far end of that return branch. The required distance is not a universal fraction of wavelength. It depends on the installed return impedance, capacitance to ground and nearby structures, choke impedance and the current distribution on every operating band.
A choke near the station protects a different boundary. It can reduce current entering equipment and wiring, but it does not undo radiation, pickup or loss on the feed-line section between the antenna and that choke.
Common-Mode Current Is Not Automatically All Loss
Current on the feed-line exterior can radiate. That does not make every watt in that current a core or copper loss. It changes the antenna geometry and can change the pattern, polarization, tuning, noise pickup and RF exposure around the station. Some of that current may contribute useful radiation in a deliberately designed branch; uncontrolled current may point energy in the wrong direction or couple it into lossy and noisy surroundings.
Core heating has a different cause. Transformer loss depends on topology, complex load, frequency, flux, winding current, voltage, parasitic capacitance, duty cycle and cooling. An installation with common-mode current does not prove that differential energy is being destroyed by “eddy currents” in a particular core, and a low SWR does not prove that the core is cool or efficient.
Do Not Promise Lower Noise, Higher Efficiency or Stable Pattern from a Label
A deliberate current boundary can improve repeatability and reduce one route by which local noise reaches the receiver. It cannot guarantee a lower noise floor, because local electric and magnetic fields can couple directly to the antenna, feed line and station wiring through multiple paths.
Likewise, an UNUN does not automatically have lower loss than a balun. Compare the actual transformer and choke assembly under the installed complex load. Include feed-line attenuation, mismatch effect, core and conductor loss, tuner loss and any deliberate or accidental return conductor.
Pattern stability is also conditional. An OCF dipole has a frequency-dependent multi-lobed pattern even with perfect common-mode suppression. Adding a deliberate coax-exterior branch creates a different three-dimensional radiator. Both can be useful; neither is proven superior by SWR, one contact or the name printed on the enclosure.
The 6:1 Question Is a Load Question
A nominal 6:1 impedance ratio corresponds to a voltage or turns ratio of approximately √6:1 in the ideal transformer model. The ratio alone does not cause saturation. Core flux, magnetising current, winding voltage and heat depend on the actual topology, frequency, complex load, source voltage, mismatch, waveform, duty cycle, material, core geometry and cooling.
That means a high-power 6:1 device cannot be accepted or rejected from the number alone. It needs a declared matchable load region, loss and temperature data, voltage/current margins and common-mode behaviour. RF.Guru’s separate technical article discusses those boundaries; this comparison makes no product claim or universal power verdict.
Compare the Complete Installed Systems
Before replacing one antenna with another, put the reference planes and intended currents on one page:
- record complex feedpoint impedance on every required band;
- identify the two wanted radiator conductors and every possible exterior return path;
- state whether the coax exterior is excluded or intentionally used;
- choose the transformation ratio from the measured load, not a traditional label;
- choose choke impedance and placement from the installed common-mode path;
- verify transformer and choke loss, temperature, voltage and current at the intended power and duty cycle;
- map current along the coax, mast, counterpoise and accessible station wiring; and
- compare field strength, pattern, received noise and repeatability with controlled A/B/A measurements.
If moving the coax or touching the station boundary changes tuning markedly, the feed line is participating. That is not automatically a failure; it is evidence that the antenna boundary differs from the drawing. Decide whether to formalise that conductor as part of the design or suppress its current.
Primary and Authoritative Sources
- Roy Lewallen, W7EL — “Baluns: What They Do and How They Do It”—the coax exterior as a third conductor, imbalance current and the different behaviour of current and voltage baluns.
- Tom Rauch, W8JI — Windom and Off-Centre-Fed Antennas—the historical single-wire Windom, modern coax-fed OCF distinctions and feed-line common-mode paths.
- Tom Rauch, W8JI — Common-Mode Current—installed conductor current, balance and current-path boundaries.
- ITU-R Report SM.2158-3—differential/common-mode definitions and imbalance-driven mode conversion.
- C. L. Ruthroff — “Some Broad-Band Transformers”—primary transmission-line-transformer circuits and their bandwidth limits.
Joeri’s Bottom Line
I am not asking anyone to throw away a working OCF dipole because somebody called it a Windom. I am asking a harder question: where does every ampere return?
If the answer includes an uncontrolled length of coax, station wiring and whatever metal happens to be nearby, the antenna is holding you back because its boundary is accidental. If the answer names the two arms, the transformer, the deliberate return branch and the choke boundary—and the measurements agree—the architecture can be engineered and repeated.
For many practical RF.Guru wire systems, a 4:1 UNUN plus a separate choke is my robust default because it does not assume that the installation remains ideally balanced. For a genuinely balanced OCF dipole, a suitable current balun can be the correct answer. The important choice is not balun versus UNUN as a tribal label; it is controlled current path versus accidental antenna.
Mini-FAQ
- Is a modern coax-fed OCF dipole the original Windom? No. The 1929 Windom used a single-wire feeder as part of the RF structure. A coax-fed OCF dipole intends a bounded feed mode and needs an explicit common-mode boundary.
- Does a 33/67 split always present 200 Ω? No. Feedpoint impedance is complex and depends on frequency, geometry, height, ground and nearby conductors.
- Does an off-centre feed automatically make the dipole unbalanced? No. The intended arm currents can still be equal and opposite. Environmental asymmetry and feed-line coupling determine common-mode conversion.
- Is a 4:1 UNUN always better than a 4:1 current balun? No. A UNUN plus separate choke is useful for an unbalanced transformed port or deliberate return branch; a measured current balun can suit a genuinely balanced OCF dipole.
- Where should the choke go? At the boundary where the intended antenna should end. That is often the feedpoint for an OCF dipole, or the far end of a deliberate counterpoise or coax-exterior branch.
- Does a 6:1 ratio inherently saturate at high power? No. Saturation, loss and temperature depend on topology, material, frequency, complex load, voltage, current, duty cycle and cooling—not on the ratio alone.