A 49:1 EFHW Transformer Is Not Wideband Just Because It Tunes
A 49:1 EFHW Transformer Is Not Wideband Just Because It Tunes
Wideband must describe a measured function over stated loads, frequencies, power and temperature. A smooth SWR trace is not a transformer-efficiency certificate.
A nominal 49:1 impedance ratio is easy to write: an ideal 7:1 voltage ratio transforms 2450 Ω to 50 Ω. A real EFHW feedpoint is neither a fixed 2450 Ω resistor nor constant across several bands. The transformer contains magnetizing impedance, leakage inductance, winding capacitance, conductor resistance, dielectric loss and ferrite loss. Every one changes the useful bandwidth.
Ask “wideband for what?” Input match, low insertion loss, stable ratio, low phase error, controlled common-mode conversion and rated-power thermal behaviour are different bandwidths. Publish the one that was actually measured.
The Ratio Is Nominal, the Load Is Complex
The ideal relation is:
Zin = Zload / n²
For n = 7, the ideal impedance ratio is 49:1.
That calculation says nothing about loss or bandwidth. The installed antenna-terminal impedance depends on wire length, diameter, geometry, height, ground, nearby conductors and the chosen return path. On a harmonic band, current distribution and feedpoint impedance can differ sharply from the fundamental. One fixed ratio therefore cannot guarantee 50 Ω on every installation or band.
Why the Band Edges Move
At the low-frequency end, insufficient magnetizing impedance can shunt the transformed load and increase current and loss. At the high-frequency end, leakage inductance, winding capacitance, inter-turn coupling and physical lead length can alter both match and transmission. Between those edges, conductor and core loss depend on frequency, flux, winding geometry and temperature.
Ferrite “mix” is not a complete design. Core size and number, magnetic path, effective area, turns, winding distribution, insulation, enclosure and load all matter. Material permeability data measured under one small-signal condition cannot by itself certify transformer loss at another flux density or temperature.
Do not assign one magic frequency range to a ferrite mix. Manufacturer curves are inputs to a transformer design. The finished assembly still needs insertion-loss, impedance, voltage, current and thermal qualification with representative loads.
SWR Can Improve for Several Reasons
A low input SWR can result from useful power transfer, a fortunate load transformation, compensation, dissipative loss or a combination. A shunt capacitor may compensate reactance and improve match over part of the band. It can also create a narrow resonance, higher circulating current or stress.
That is why a capacitor is neither automatically a trick nor automatically a cure. Measure the complete network with and without it at the loads the antenna is expected to present. If the input match improves, verify where the accepted power goes.
What a Defensible Bandwidth Statement Contains
| Claim | Required measurement |
|---|---|
| “The input is matched from A to B.” | S11 or input impedance at stated reference impedance and representative loads |
| “Insertion loss remains below X.” | Accepted and delivered power or calibrated two-port data with mismatch and fixture treatment |
| “The ratio remains close to 49:1.” | Complex voltage/current or impedance transformation across the declared load matrix |
| “It handles the stated power.” | Waveform, duty cycle, load, frequency, ambient, cooling, temperature and voltage/current margins |
| “It controls feedline current.” | Common-mode impedance or installed exterior-current measurements—not differential S21 alone |
Use More Than One Load
A single resistor is a repeatable benchmark, not an EFHW installation. Build a safe load matrix spanning the expected resistance and reactance at the high-impedance port. At minimum, record:
- full complex S11, S21, S12 and S22 where a two-port method is appropriate;
- fixture and calibration planes, connector and lead geometry;
- accepted and delivered power rather than forward power alone;
- case or core temperature versus time at a declared waveform and duty cycle;
- repeatability between units and after thermal cycling; and
- common-mode current on the coax exterior in the intended installation.
Small-Signal and Operating-Power Tests Answer Different Questions
A VNA is excellent for locating resonances, mismatch, transmission and unit-to-unit spread while the device remains linear. It normally uses far less power than a transmitter. Ferrite permeability, core loss, winding resistance and compensation can change with flux density and temperature, so a small-signal sweep cannot certify operating-power performance.
An operating-power test needs a rated enclosure and load, calibrated power planes, remote temperature observation, a declared duty cycle and a credible uncertainty budget. Never touch or reconfigure a high-impedance EFHW fixture while energized.
The Antenna Still Needs Its Own Evidence
Transformer efficiency is not antenna radiation efficiency. Even a low-loss transformer can feed a radiator with ground loss, nearby-object loss, pattern nulls or uncontrolled coax current. Conversely, a moderate input mismatch does not prove the transformer is dissipating large power.
For an 80–10 metre claim, qualify every intended band as a separate operating condition. Record the installed antenna impedance, transformer loss and temperature, feedline current, tuner state where used and the antenna’s measured or validated pattern. If only match was measured, call it match bandwidth.
Engineering References
- Fair-Rite: Ferrites in Broadband Transformers and Material Data
- Fair-Rite: Use of Ferrites in Broadband Transformers
- Keysight: S-Parameter Techniques for Network Design
- Keysight: RF Power Transfer and Mismatch Uncertainty
Mini-FAQ
- Does 49:1 mean an EFHW feedpoint is always 2450 Ω? No. It is the ideal transformation of one load; the installed feedpoint impedance is complex and changes with band and geometry.
- Does low SWR prove low transformer loss? No. Match and insertion loss are different measurements, and dissipation can sometimes improve the apparent match.
- Is a compensation capacitor always bad? No. It can provide useful compensation, but its effects on loss, bandwidth, circulating current and voltage must be measured.
- Can a ferrite-mix chart certify transformer bandwidth? No. It informs the design; the finished winding, load, power and temperature determine performance.
- Can a VNA sweep certify high-power operation? No. It measures the small-signal network. Rated-power loss, heating and insulation margins need a separate test.
- What should “wideband” state? State the measured function, frequency range, load matrix, power, waveform, temperature, reference planes and uncertainty.