Multiband EFHW Transformers: Where Ferrite and Winding Loss Appear
Multiband EFHW Transformers: Where Ferrite and Winding Loss Appear
An 80–10 metre SWR plot can show where a station finds a match. It cannot show, by itself, how much accepted power reaches the wire or where transformer heat is being produced.
The attraction of a multiband EFHW is obvious: one wire, one feedpoint and several amateur bands. The transformer has the harder assignment. It must work into a high, complex antenna impedance that changes with band, geometry, height, ground and the chosen return path. A nominal ratio and a cool enclosure at one frequency do not settle that problem.
My rule is simple: do not call a transformer efficient because the radio is happy. Define the load, reference planes, frequency, waveform, duty cycle and temperature, then measure accepted power and delivered power.
The Ideal Ratio Is Only the Starting Point
An ideal transformer with a 7:1 voltage ratio has a 49:1 impedance ratio. It would transform 2450 Ω to 50 Ω without loss. A real EFHW feedpoint is not a fixed 2450 Ω resistor, and a real transformer contains magnetizing impedance, leakage inductance, winding capacitance, conductor resistance, dielectric loss and ferrite loss.
The antenna-terminal load also needs a second current path. Depending on the installation, that path can include a deliberate counterpoise, the coax exterior, bonding conductors and capacitive coupling to the surroundings. Moving the choke or changing the coax length can therefore change the load presented to the transformer. That is an antenna-system change, not proof that the core itself improved or failed.
What Limits the Low-Frequency End
At the low-frequency edge, the primary magnetizing impedance may be too small compared with the transformed antenna load. More current then flows in the magnetizing branch. Core and winding loss can rise, while the input impedance departs from the ideal ratio.
Adding turns can increase low-frequency magnetizing inductance, but it also adds conductor length and changes leakage inductance and inter-winding capacitance. A larger or different core can change flux density and thermal headroom, but material name alone is not a power rating. Core geometry, winding distribution, enclosure temperature and the actual load remain part of the result.
What Limits the High-Frequency End
At the high-frequency edge, leakage inductance, winding capacitance, lead length and layout stop behaving like minor imperfections. They form a distributed network around the intended transformation. A compensation capacitor may improve input match over part of the range, but that does not reveal whether delivered power increased, circulating current rose or loss moved elsewhere.
Conductor loss is also frequency dependent. Skin and proximity effects change the winding's AC resistance, while voltage distribution can concentrate electric-field stress across turns and insulation. These mechanisms are design-specific; a fixed list assigning one amateur band to one ferrite mix is not a substitute for testing the completed assembly.
Heating Is a Measurement, Not a Diagnosis
A rising case temperature proves that heat is being produced and retained. It does not identify the share from ferrite, winding, connector, compensation component or enclosure. Conversely, a case that feels cool may conceal a hot spot or simply have enough thermal mass that the test ended too soon.
Record temperature against time at declared frequency, accepted power, waveform, duty cycle, ambient temperature and load. Place sensors consistently and allow the test to reach a meaningful thermal condition. For high-impedance EFHW fixtures, maintain adequate clearance and never touch or reconfigure the circuit while energized.
Small-Signal and Transmit-Power Tests Have Different Jobs
| Question | Useful evidence | What it does not prove alone |
|---|---|---|
| Where does the input match? | Calibrated complex S11 or impedance at a stated reference plane and load | Insertion loss, thermal margin or radiation efficiency |
| How much power reaches the high-impedance port? | Calibrated accepted and delivered power, or a validated two-port method with mismatch treatment | Installed antenna pattern or common-mode current |
| Does the ratio remain useful? | Complex transformation across a representative resistance-and-reactance matrix | Performance into every EFHW geometry |
| Does it survive the operating mode? | Power and thermal test with waveform, duty cycle, time, cooling and voltage/current margins | A different mode, band, enclosure or load |
| Is the feedline exterior controlled? | Installed common-mode impedance or exterior-current measurements | Transformer insertion loss from differential S-parameters |
A VNA is valuable for locating resonances, characterising mismatch and comparing units while the network is linear. Its small test signal does not reproduce transmit-power flux, heating, insulation stress or temperature drift. Operating-power tests add those questions; they do not replace calibrated small-signal data.
One Core Can Cover Several Bands—If the Claim Is Bounded
There is no physical rule saying that one transformer cannot serve several amateur bands. The defensible claim is narrower: the completed transformer met stated match, loss, temperature and stress limits across a declared set of loads and frequencies. Wider coverage normally makes parasitic behaviour, load spread and thermal qualification more demanding.
Splitting the coverage, switching networks or using different antenna systems can make the engineering easier, but none is automatically more efficient. The right choice follows the installed load matrix, operating power, duty cycle, weather protection, acceptable loss and maintenance plan.
A Practical Qualification Sequence
- Measure or model the expected complex antenna-terminal impedance on every intended band, including the defined return path and choke boundary.
- Characterise the transformer with repeatable fixtures and more than one resistive-reactive load.
- Separate input match from insertion loss and state the calibration and reference planes.
- Repeat at operating power with declared waveform, duty cycle, test duration, ambient conditions and temperature locations.
- Map coax-exterior current in the installed antenna rather than assuming the transformer ratio controls common mode.
- Publish uncertainty and operating limits with the result instead of reducing the design to one SWR number.
Engineering References
- Fair-Rite: Use of Ferrites in Broadband Transformers
- Fair-Rite: Material Data and Magnetic Properties
- Keysight: S-Parameter Design Techniques
- Keysight: S-Parameters, Mismatch and Measurement Uncertainty
Mini-FAQ
- Does low SWR prove that a multiband EFHW transformer is efficient? No. Match and insertion loss are different measurements, and dissipation can improve the apparent input match.
- Does one ferrite mix have a fixed amateur-band range? No. Material data informs the design, but core geometry, turns, winding layout, load, flux and temperature determine the finished transformer.
- Why can the low-frequency end run hotter? Insufficient magnetizing impedance can increase magnetizing current, but the actual loss also depends on load, winding and core design.
- Why can the high-frequency match change? Leakage inductance, winding capacitance, conductor loss and lead geometry become more influential as frequency rises.
- Can a compensation capacitor be useful? Yes, when it is part of a measured network. Its effects on delivered power, circulating current, bandwidth and stress still need verification.
- Can a VNA certify the transmit-power rating? No. A VNA characterises the small-signal network; power, temperature and insulation margins require separate operating-power tests.