Skip to content

Your cart is empty

Continue shopping

Have an account?

Log in to check out faster.

Your cart

Loading...

Estimated total

€0,00 EUR

Tax included and shipping and discounts calculated at checkout

NEW - CM/DM Filter for Analog Hotspot

  • New
  • Swag
  • HotSpot
  • Repeater
    • Build Your Own Repeater
    • ON0ORA
  • BalUn/UnUn
    • Balun/LineIsolator/Choke
    • Unun/Transformers
    • Lightning & Surge Protection
    • AC/DC Choke/LineIsolator
    • Grounding
    • Anti-Corrosion
  • Filters
    • VHF-UHF Filter
    • Line Filters
  • Antenna
    • HF Active RX Antenna
    • HF End Fed Wire Antenna
    • HF Verticals - V-Dipoles
    • HF Rigid Loops
    • HF Doublets - Inverted Vs
    • HF Stealth POTA/SOTA Antennas
    • UHF Antenna
    • VHF Antenna
    • Dualband VHF-UHF
    • Grounding
    • Masts
    • Guy Ropes & Accessories
    • GPS Antenna
    • Mobile Antenna
    • Handheld Antenna
    • ISM Antenna 433/868
    • Antenna Tools
    • Anti-Corrosion Lubricants
    • Dummy Load
  • Coax
    • Coaxial Seal
    • Coax Connectors
    • Panel Mount Connectors
    • Coax Adaptors
    • Coax Tools
    • Coax Cable
    • Coax Surge protection
    • Jumper - Patch cable
  • 19"
  • 13.8 V
    • DC-DC
    • AC-DC
    • Powerpole
    • 13.8 V Cable
  • PA
    • VHF Power Amplifiers
    • UHF Power Amplifiers
  • Parts
    • Ferrite
    • Pi
    • Routers
    • Enclosures
  • PCB
  • SDR
  • APRS
  • KB
    • Why we started RF.Guru
    • Mission Statement
    • Product Whitepapers
    • Knowledge Base
    • Transmit Antennas
    • Baluns and Ununs
    • Receive Antennas & Arrays
    • Technical Deep Dives
    • Debunking Myths
    • Transmission lines
    • Radio Interference
    • Grounding and safety
    • Ham Radio 101
    • Calculators
    • Ham Florida Man
    • Errata & Modern Context
    • The Scientists Who Built RF
    • %λΦ#@!Ω
  • ON6URE
    • on the road ...
    • collaborations ...
    • on4aow ...
    • on4pra ...
Log in

Country/region

  • Belgium EUR €
  • Germany EUR €
  • Italy EUR €
  • Sweden EUR €
  • Australia EUR €
  • Austria EUR €
  • Belgium EUR €
  • Bulgaria EUR €
  • Canada EUR €
  • Croatia EUR €
  • Czechia EUR €
  • Denmark EUR €
  • Estonia EUR €
  • Finland EUR €
  • France EUR €
  • Germany EUR €
  • Greece EUR €
  • Hungary EUR €
  • Ireland EUR €
  • Italy EUR €
  • Latvia EUR €
  • Lithuania EUR €
  • Luxembourg EUR €
  • Netherlands EUR €
  • New Zealand EUR €
  • Norway EUR €
  • Poland EUR €
  • Portugal EUR €
  • Romania EUR €
  • Slovakia EUR €
  • Slovenia EUR €
  • Spain EUR €
  • Sweden EUR €
  • Switzerland EUR €
  • United Kingdom EUR €
  • United States USD $
  • YouTube
RF.Guru Logo
  • New
  • Swag
  • HotSpot
  • Repeater
    • Build Your Own Repeater
    • ON0ORA
  • BalUn/UnUn
    • Balun/LineIsolator/Choke
    • Unun/Transformers
    • Lightning & Surge Protection
    • AC/DC Choke/LineIsolator
    • Grounding
    • Anti-Corrosion
  • Filters
    • VHF-UHF Filter
    • Line Filters
  • Antenna
    • HF Active RX Antenna
    • HF End Fed Wire Antenna
    • HF Verticals - V-Dipoles
    • HF Rigid Loops
    • HF Doublets - Inverted Vs
    • HF Stealth POTA/SOTA Antennas
    • UHF Antenna
    • VHF Antenna
    • Dualband VHF-UHF
    • Grounding
    • Masts
    • Guy Ropes & Accessories
    • GPS Antenna
    • Mobile Antenna
    • Handheld Antenna
    • ISM Antenna 433/868
    • Antenna Tools
    • Anti-Corrosion Lubricants
    • Dummy Load
  • Coax
    • Coaxial Seal
    • Coax Connectors
    • Panel Mount Connectors
    • Coax Adaptors
    • Coax Tools
    • Coax Cable
    • Coax Surge protection
    • Jumper - Patch cable
  • 19"
  • 13.8 V
    • DC-DC
    • AC-DC
    • Powerpole
    • 13.8 V Cable
  • PA
    • VHF Power Amplifiers
    • UHF Power Amplifiers
  • Parts
    • Ferrite
    • Pi
    • Routers
    • Enclosures
  • PCB
  • SDR
  • APRS
  • KB
    • Why we started RF.Guru
    • Mission Statement
    • Product Whitepapers
    • Knowledge Base
    • Transmit Antennas
    • Baluns and Ununs
    • Receive Antennas & Arrays
    • Technical Deep Dives
    • Debunking Myths
    • Transmission lines
    • Radio Interference
    • Grounding and safety
    • Ham Radio 101
    • Calculators
    • Ham Florida Man
    • Errata & Modern Context
    • The Scientists Who Built RF
    • %λΦ#@!Ω
  • ON6URE
    • on the road ...
    • collaborations ...
    • on4aow ...
    • on4pra ...
Log in Cart

Broadband HF Transformers: Topology, Flux and Honest Measurement

An RF.Guru transformer engineering guide

Broadband HF Transformers: Topology, Flux and Honest Measurement

A ratio is only the beginning. Bandwidth and power depend on which transformer you built, which impedances terminate it, how flux and line modes divide the work, and where you measured.

ON6UREHF transformersTransmission linesFerriteVNA measurementTechnically reviewed
Related reading:
Ferrite mixes on HF: chokes vs broadband transformers Why back-to-back EFHW measurements keep fooling people Why the Y21 method is the only ham measurement that actually works

“Broadband transformer” is not one circuit. A conventional flux-coupled transformer, an autotransformer, a Guanella transmission-line transformer and a Ruthroff-style voltage transformer can share the same nominal impedance ratio while presenting different flux, isolation, common-mode, parasitic and termination behavior.

Engineering boundary: a transformer claim is meaningful only when the topology, schematic, core part and material, winding geometry, conductor and insulation, complex source and load impedances versus frequency, waveform, duty cycle, temperature limit, fixture and measurement planes are declared. The methods below qualify a design; they are not a universal winding recipe, bandwidth promise or watt rating.

First Separate the Transformer Families

Family How the wanted signal transfers Questions that decide performance
Conventional isolated transformer Voltage applied to one winding creates mutual core flux that induces voltage in another winding. Turns, magnetizing impedance, leakage, winding capacitance, flux density, loss and insulation.
Autotransformer or tapped voltage transformer A shared winding combines conductive transfer and transformer action. Tap ratio, common connection, core flux, leakage, current distribution and the absence of galvanic isolation.
Transmission-line transformer (TLT) Power propagates in the line mode of one or more controlled transmission lines; their ends are interconnected in series, parallel or phase-reversing combinations. Exact Guanella/Ruthroff or other topology, line impedance and length, common-mode choking, port references and parasitic modes.
Hybrid Conventional and transmission-line sections deliberately share the transformation. Which section sets low-frequency flux, high-frequency phase, isolation and loss.

The names are not interchangeable. A two-conductor winding may look bifilar yet operate mainly as a conventional transformer in one connection and as a transmission line in another. Conversely, a coax wound through ferrite carries the wanted differential or line-mode signal inside the coax while the ferrite impedes unwanted common-mode current on the outside. The schematic, current modes and port definition—not the appearance—identify the device.

DC isolation is also topology-specific. Separate conventional windings can provide galvanic separation. Autotransformers and many transmission-line-transformer connections are DC continuous. A “balun” label only says that balanced and unbalanced ports are involved; it does not prove safety isolation, a particular impedance ratio or good common-mode rejection.

Write the Ratio With Ports and Direction

For an ideal conventional transformer, define a = N1/N2. With voltage and current reference directions chosen consistently:

V1/V2 = a

I1/I2 = 1/a

Zin,1 = a2ZL,2

The last equation transforms the complete complex impedance in the ideal model: both resistance and reactance are multiplied by a2. It does not cancel reactance or make a frequency-dependent antenna load constant. A nominal 49:1 transformer connected to 2450 + jX Ω ideally presents 50 + jX/49 Ω in one direction—not exactly 50 Ω unless X = 0, and not after real leakage, capacitance and loss are included.

Declared impedance ratio, high:low Ideal conventional turns ratio, high:low What must still be stated
4:1 2:1 Port direction, isolated or tapped, balanced/unbalanced state and nominal termination.
9:1 3:1 Whether 9:1 describes full windings, a center-tapped section or a TLT interconnection.
49:1 7:1 Frequency-dependent load, compensation, voltage stress and actual topology.
70:1 8.37:1 That this square root is an ideal conventional-transformer value, not a winding instruction.
120:1 10.95:1 Feasibility, parasitics, insulation, load range and measured band—not a generic multi-section promise.

A bare “4:1” is ambiguous because some documents write input:output, others high:low, and some use turns rather than impedance. A defensible specification says, for example, “50 Ω unbalanced to 200 Ω differential, 4:1 impedance, low:high.” On a balanced port, differential impedance is measured between the two conductors; it is not automatically the sum of two equal impedances to chassis unless the circuit is symmetric and the common-mode condition is defined.

Transmission-Line Ratios Come From Connections

A TLT does not obtain every impedance ratio by squaring a visible turn count. Its voltage and current ratios arise from how line sections are combined. In an ideal 4:1 Guanella transformation from 50 Ω on the low-impedance side to 200 Ω on the high-impedance side, two equal lines are paralleled at the 50 Ω port and placed in series at the 200 Ω port. Each line then operates at about 100 Ω:

Zline = 2Rlow = Rhigh/2 = 100 Ω

√(RlowRhigh) = √(50 × 200) = 100 Ω

The geometric-mean result works for that ideal 4:1 series/parallel case. It is not a universal TLT rule. A different number of lines, a Ruthroff phase-addition connection, unequal terminations, complex impedances, finite electrical length or additional compensation changes the required modal impedances and voltages. Solve the actual network before choosing a twisted-pair, coax or PCB geometry.

In a Guanella current-balun structure, the wanted equal-and-opposite line currents ideally cancel their net core excitation, while the core presents impedance to common-mode current. That does not make the core irrelevant or “flux-free.” Imperfect balance, line-to-core capacitance, common-mode voltage, finite choking impedance and asymmetric loading can still create core loss and heating. Ruthroff and conventional voltage-transformer sections can impose more direct transformer flux.

The Low-Frequency Limit Is an Admittance and Flux Problem

In the conventional equivalent circuit, magnetizing impedance shunts the ideally transformed load. If it is approximated as an inductance Lm, then Xm = 2πfLm. The familiar target Xm ≥ 4R means magnetizing current is no more than about one quarter of the resistive load current in that simplified model:

Lm ≥ mR/(2πfmin), where m is the chosen reactance margin

For R = 12.5 Ω, fmin = 2 MHz and m = 4: Lm ≥ 3.98 µH

That calculation is dimensionally correct, but 4R is a starting heuristic—not a universal pass limit. Choose m from allowed return loss, phase error, source current and loss. Ferrite permeability is complex and frequency-, temperature- and drive-dependent, so the real magnetizing branch is lossy. A small-signal AL value or LCR-meter reading at one test frequency does not prove large-signal impedance or low loss across HF.

For a TLT, the low end may instead be set by insufficient common-mode or choking impedance, by the residual magnetizing branch of a hybrid section, or by both. Evaluate the impedance in the current mode the core is supposed to control. A material sold primarily for suppression is not automatically the lowest-loss transformer material at the required flux and frequency.

Volts per Turn Set Flux; Current and Loss Set Heat

Faraday’s law provides the robust boundary for any winding that sees transformer voltage:

ΔB = (1/(NAe)) ∫v(t)dt

For a sine wave: Vrms = 4.44 fNAeBpeak

Use the effective or minimum core area specified by the manufacturer, consistent SI units, and the actual voltage waveform across that winding. Crest factor, AM or SSB envelope, key-down duty cycle, DC imbalance and switching asymmetry matter. “Watts” alone cannot determine flux because the same power produces different voltage and current at different impedances.

For example, 100 W into matched 50 Ω is 70.7 V RMS and 1.414 A RMS. The same ideal power at 200 Ω is 141.4 V and 0.707 A; at 12.5 Ω it is 35.4 V and 2.828 A. A high-impedance winding faces more electric-field and insulation stress, while the low-impedance path faces more conductor and joint heating. Mismatch can increase one or both beyond the matched values.

A ferrite data sheet value for initial permeability or flux density is not automatically a safe RF operating limit. Manufacturer curves state their specimen, frequency, temperature and excitation conditions. Core loss can become unacceptable far below the material’s saturation flux, especially over long duty cycles. Establish a conservative loss/temperature design point from material data, then verify the assembled transformer at the actual waveform and cooling condition.

The High-Frequency Limit Is a Distributed Network

Leakage inductance and winding capacitance matter, but a broadband transformer rarely has one clean self-resonance. It is a multi-conductor distributed network containing:

  • leakage inductance and imperfect mutual coupling;
  • turn-to-turn, winding-to-winding, winding-to-core and winding-to-chassis capacitance;
  • transmission-line delay and characteristic-impedance error;
  • frequency-dependent complex core permeability;
  • skin and proximity effects in conductors;
  • dielectric loss, connector and solder-joint inductance; and
  • common-mode and differential-mode resonances that need not occur together.

The estimate 1/(2π√(LleakCstray)) can identify one possible lumped resonance, but it is not a complete upper-frequency formula. Adding turns usually raises magnetizing inductance and line length, but its effects on leakage and capacitance depend on spacing, layering and connection. Separating windings can reduce capacitance while increasing leakage; closer multifilar winding often does the reverse.

Winding Geometry Is a Three-Way Trade

Bifilar and trifilar winding can reduce leakage and make line geometry repeatable. Tight spacing also increases interconductor capacitance and electric-field stress. At HF, copper loss includes skin and proximity effects, while dielectric heating depends on loss tangent, field distribution and frequency. Tight coupling does not inherently “increase loss”; it changes several loss and parasitic terms at once.

Construction Potential advantage What prevents a generic ranking
Twisted pair or multifilar wire Compact, adjustable coupling and practical custom line impedance. Twist pitch, insulation thickness/loss, conductor size, turn pressure and repeatability.
Coax or twinax Defined conductor relationship and shielding or balanced geometry. The available impedance may be wrong; bend, shield-current path, dielectric, voltage and heat limits still apply.
Planar or etched winding Repeatable geometry and integration with the fixture or amplifier. Copper thickness, laminate loss, field crowding, via/termination inductance and thermal path.
Tube, strap, braid or one-turn path Low DC resistance and potentially high current capacity. Large loop field, current crowding, leakage, winding-to-core capacitance, joints and insulation.

No family is inherently the highest-power or widest-band option. Compare the actual line impedance, conductor loss, core excitation, voltage spacing, thermal resistance and modal response. A tightly twisted pair on a core is not automatically a known Z0; measure or model the pair with its real insulation, pitch, length and surroundings.

Compensation Must Correct a Measured Model

A few picofarads can flatten one build and destabilize another. A shunt capacitor may resonate leakage inductance, alter port impedance or move a pole and zero, but it can also raise circulating current, port voltage and sensitivity to component tolerance. Series, shunt and distributed compensation are different networks.

Do not copy a capacitor merely because the nominal ratio and core look similar. Fit a physically credible equivalent or electromagnetic model to de-embedded complex measurements, choose the response target, check tolerance and temperature, and re-verify voltage, current, loss and out-of-band behavior at power.

There Is No Universal Watt Rating

The first limit reached can be core loss, conductor heating, dielectric loss, insulation breakdown, corona or discharge at a termination, ferrite saturation, connector temperature, solder fatigue or load-mismatch voltage. The result changes with frequency, waveform, average and peak power, duty cycle, ambient temperature, enclosure, airflow, installation and SWR.

A defensible rating therefore states:

  • frequency band and exact source/load impedances;
  • waveform, average power, PEP and duty cycle;
  • maximum permitted port mismatch or load set;
  • ambient, enclosure, airflow and test duration;
  • temperature measurement locations and limits;
  • core, conductor, insulation, capacitor and connector limits; and
  • pass criteria for loss, return loss, balance, distortion and permanent drift.

Enamelled wire has manufacturer- and standard-specific thermal and breakdown properties. A DC hipot between windings does not by itself prove turn-to-turn RF endurance, creepage, clearance, partial-discharge margin or safe separation at the operating frequency. Galvanic isolation is not automatically certified safety isolation. Where the transformer protects users or crosses an energy boundary, apply the relevant equipment safety standard and insulation system—not an amateur rule of thumb.

Measure at the Impedances the Transformer Was Designed For

A 50 Ω VNA at both ports does not directly characterize a 50-to-200 Ω transformer under its intended termination. Calibrate to the closest practical planes, characterize and de-embed fixture sections, and then embed matching networks or renormalize the network to the declared port impedances. Record the reference impedances and reference planes with every plot.

For a balanced port, use enough receiver ports to preserve both conductors and convert the measured multiport data to mixed-mode quantities. Report differential insertion or transducer gain, input and output match, amplitude balance, phase balance and common-mode conversion. A two-terminal shortcut cannot reveal every imbalance or common-mode path.

Insertion loss is not simply “minus S21.” It is the additional loss relative to an ideal transformer of the same ratio under the specified source and load conditions. A raw 50 Ω S21 can include intended impedance transformation, mismatch and fixture loss. Separate reflected power from dissipative loss before assigning heat to the core.

Why back-to-back can mislead

Two nominally identical transformers connected back-to-back return to the analyzer’s impedance and can be a useful production comparison. Dividing total loss in dB by two is valid only when the units behave identically, see their intended terminations, do not couple thermally or electromagnetically, and the fixture and mismatch contributions are known. One S21 trace cannot uniquely assign unequal loss or reveal each unit’s individual return loss and balance.

Small-signal and high-power tests answer different questions

  1. Small signal: de-embedded complex S, Z or Y parameters over a band; magnetizing impedance; balance; common-mode conversion; resonances and repeatability.
  2. Moderate drive: compression, harmonic/intermodulation change and load sensitivity.
  3. Rated drive: temperature versus time at each critical frequency, loss, waveform, enclosure and mismatch corner.
  4. After stress: repeat the small-signal sweep and insulation checks to detect permanent drift or damage.

No single parameter method is universally “the only one that works.” S-parameters are natural for travelling-wave RF systems; Z and Y matrices can expose magnetizing and transfer behavior; mixed-mode parameters expose balance. The right representation must be derived from a calibrated network measurement with the transformer’s actual port topology and terminations.

A Reviewable Transformer Specification

Define before winding Verify after winding
Exact schematic, polarity, port names, balance state and DC paths Continuity, polarity, isolation and multiport mode conversion
Complex source/load versus frequency and allowed mismatch De-embedded input/output match and transducer response at declared references
Conventional, Guanella, Ruthroff, autotransformer or hybrid operating mode Magnetizing/common-mode impedance and distributed resonances
Core manufacturer, part, material curves, area and thermal boundary Flux calculation, temperature rise, compression/distortion and drift
Conductor, line impedance, insulation, spacing and compensation Loss, balance, electric stress and component tolerance
Waveform, PEP, average power, duty, ambient, airflow and SWR Steady-state and transient thermal performance at the worst credible corners

That specification turns “4:1 broadband, 1 kW” from a slogan into a reproducible engineering claim. Without it, the apparent ratio can be correct at one frequency while loss, balance, insulation or temperature fails elsewhere.

Primary Engineering Sources

  • Mini-Circuits AN20-001, How RF Transformers Work and How They Are Measured: ideal ratio definitions, terminating impedances, insertion-loss references and transformer measurement methods.
  • Mini-Circuits AN20-002, Application Note on Transformers: conventional and transmission-line configurations, low/high-frequency boundaries, balance and multiport characterization.
  • Mini-Circuits transformer-configuration lexicon: current manufacturer taxonomy for conventional, autotransformer, transmission-line, Guanella and Marchand implementations.
  • Fair-Rite 17th-edition catalogue and design formulas: core geometry, AL, flux-density calculation, material conditions and complex magnetic behavior.
  • Fair-Rite 43 Material data sheet: an example showing that permeability, flux, loss, temperature and impedance claims have explicit specimen and test conditions.
  • Keysight de-embedding and embedding application note: reference-plane definition, fixture removal and S-parameter network de-embedding.
  • Coilcraft MS520RFA manufacturer data: an example of a transformer specification that states winding ratio, small-signal inductance test conditions, insertion loss, bandwidth, current and isolation voltage separately.
  • IEC 60317-0-1:2013+AMD1:2019+AMD2:2026: current general requirements and breakdown/continuity boundaries for enamelled round copper winding wire.
  • IEC 62368-1:2023, corrected 2025-08: equipment-level energy-source, insulation, clearance, creepage, electric-strength and temperature safeguards where applicable.

Follow the Current Path, Not the Folklore

Explore more RF.Guru technical deep dives on transmission lines, common-mode current, baluns, chokes and antenna measurement—and subscribe for new engineering articles and laboratory notes.

Join the notification list →

Mini-FAQ

  • Does a 4:1 impedance ratio always mean a 2:1 turns ratio? Only for the ideal conventional-transformer definition with explicit ports. A transmission-line transformer obtains its ratio from line connections, and its visible turns around a core do not provide the same rule.
  • Will a 49:1 transformer turn every end-fed antenna load into 50 Ω? No. It transforms the load’s resistance and reactance according to its real frequency-dependent network. The antenna, return path, transformer loss and parasitics still determine the input impedance.
  • Is magnetizing reactance equal to four times resistance a universal design rule? No. It is a starting heuristic that permits about one-quarter reactive magnetizing current in a simplified model. Required return loss, phase, flux, loss and source current may demand another margin.
  • Does a transmission-line transformer avoid core flux and heating? No. Wanted power mainly travels in line mode, but finite common-mode voltage, imbalance, choking impedance, hybrid transformer action and parasitic coupling can excite and heat the core.
  • Can I divide a back-to-back loss measurement by two? Only if the two transformers are effectively identical, correctly terminated and uncoupled, with fixture and mismatch effects known. One combined trace cannot prove each unit’s individual loss, match or balance.
  • How is a broadband HF transformer power rating established? By stating frequency, impedance, waveform, average and peak power, duty cycle, SWR, ambient and cooling, then verifying loss, temperature, insulation, distortion and post-stress drift at the worst credible conditions.

Questions, antenna-factor records or height trials to share? Contact RF.Guru.

Joeri Van Dooren, ON6URE — RF engineer, antenna designer and founder of RF.Guru, specialising in practical HF/VHF receiving systems and RF components.

Subscribe here to receive updates on our latest product launches

  • YouTube
Payment methods
  • Bancontact
  • iDEAL Wero
  • Klarna
  • Maestro
  • Mastercard
  • MobilePay
  • PayPal
  • Visa
© 2026, RF Guru Powered by Shopify
  • Refund policy
  • Privacy policy
  • Terms of service
  • Contact information
  • News
  • Guru's Lab
  • Press
  • DXpeditions
  • Fairs & Exhibitions
  • Order Withdrawal
  • Choosing a selection results in a full page refresh.
  • Opens in a new window.
Purchase options
Select a purchase option to pre order this product
Countdown header
Countdown message


DAYS
:
HRS
:
MINS
:
SECS