When a “Common-Mode Choke Test Jig” Measures the Jig

A common-mode current choke is not a 50-ohm attenuator.

It is a series impedance inserted into an unwanted current path.

That distinction matters.

Several ham-radio VNA fixtures are sold as common-mode current choke test rigs. They promise an easy way to measure choke effectiveness with a VNA, often by looking at S21 LOGMAG in dB.

The fault is not that a VNA is being used. A VNA can be an excellent measurement instrument.

The fault is treating raw S21 LOGMAG as “common-mode choke attenuation.” That is not a valid measurand for a common-mode choke.

A fixture may create some shield current. It may even be useful for rough comparison. But the advertised S21 attenuation result is still a two-port transmission measurement through a particular fixture, with a particular return path, particular parasitic capacitances, particular calibration assumptions, and usually a 50-ohm VNA environment.

A common-mode choke in a station, on a USB cable, on an Ethernet cable, on an audio lead, or on a DC power cable is not operating between two clean 50-ohm common-mode ports.

The 50-Ohm Error

A VNA S21 measurement is a transmission coefficient. It tells how much of a stimulus applied at port 1 is received at port 2.

S-parameters are defined relative to port reference impedances. In normal RF bench work, that usually means 50 ohms.

That is perfectly valid for a 50-ohm filter, attenuator, amplifier, coaxial device, or other network intended to live in a defined 50-ohm environment.

But the common-mode path on the outside of a coax shield is not the 50-ohm coaxial transmission line.

The 50-ohm line is between the center conductor and the inside surface of the shield. The unwanted common-mode current is on the outside surface of the shield, returning through the station, antenna, mast, chassis, grounding system, other cables, stray capacitance, nearby objects, or the operator.

So a measured value such as -25 dB S21 is not the amount by which the choke will reduce common-mode current in a real installation.

It is only the transmission through that jig’s artificial path.

Once the Outside of the Cable Is the Circuit, the Bench Is Part of the Circuit

This is the part that often gets missed.

If the measurement intentionally places current on the outside of a cable, then the outside world becomes part of the RF circuit.

The fixture metal, connector shells, unused sockets, output cable routing, cable length, nearby metal, VNA chassis, stray capacitance, operator hand capacitance, and the bench itself can all affect the result.

That does not make the measurement useless.

It does mean the result is fixture-dependent.

If the measured curve changes because of switch layout, connector spacing, output cable routing, unused connectors, or nearby objects, then the trace is not a clean measurement of the choke alone. It is a measurement of the choke plus the fixture plus the environment.

The “Digital” Version Makes the Error Worse

For coaxial shield-current work, a simplified fixture may still have some value as a comparison tool. At least the physical current of interest is reasonably easy to visualize: current on the outside of the coax shield.

But extending the same idea to Ethernet, USB, audio, and DC power connectors is much harder to defend.

Common mode on those cables is not simply “current on the outside of the shield.”

Some cables are unshielded. Some have multiple conductors. Some have balanced pairs. Some have a shield plus signal conductors. Some have no shield at all.

In those cases, common-mode current is a modal current involving the conductors as a group relative to the outside world.

For Ethernet, USB, audio, and DC power cables, the connection topology defines the mode being measured. A fixture that places multiple connector types in parallel and treats the resulting S21 trace as common-mode choke attenuation is not equivalent to a defined common-mode measurement.

Professional signal-integrity and EMC work does not rely on a generic “outside of the cable” assumption. It defines the mode, the test board, the port impedances, the return path, and the relevant mixed-mode quantities.

For example, Ethernet common-mode choke characterization may involve differential insertion loss, common-mode rejection, and mode-conversion parameters. Those are not the same thing as raw single-ended S21 through a convenience jig.

The Higher the Frequency, the Less Believable Raw S21 Becomes

At lower HF, a simple S21 fixture may give a rough comparison between two chokes, provided the same fixture, calibration, cable layout, and environment are used every time.

As frequency increases, the measurement becomes increasingly dominated by fixture parasitics.

A few picofarads of stray capacitance can bypass a high-impedance choke.

That is not a harmless detail when we are trying to measure choke impedances in the 1 kΩ to 10 kΩ range.

This is exactly where the plain S21 method becomes weak. Stray capacitance, crosstalk, connector geometry, cable coupling, and fixture layout can all create a bypass path around the choke.

The result may still be repeatable, but repeatable does not automatically mean valid.

“-25 dB” Is Not a Choke Specification

A common recommendation is to look for around -25 dB or more of S21 loss.

That sounds simple, but it creates a dangerous shortcut.

-25 dB is meaningful only in the measurement circuit that produced it. It does not mean the choke will reduce common-mode current by 25 dB in the user’s antenna system, shack, USB cable, Ethernet cable, power lead, or audio cable.

A common-mode choke does not have a universal attenuation value.

It has an impedance:

ZCM = R + jX

The actual current reduction depends on the common-mode source impedance, load impedance, return path, cable length, coupling to nearby objects, and the choke’s own impedance.

In a real station, those impedances are usually unknown, installation-specific, and frequency-dependent.

Why 50-Ohm Insertion Loss Has Always Needed Caution

This is not just ham-radio nitpicking.

EMC filter measurement practice has long recognized that 50-ohm insertion-loss curves can be misleading when the real system is not 50 ohms.

A mains filter, cable choke, or conducted-emissions filter may be tested in a defined 50-ohm environment for comparison, but that does not mean the same attenuation will appear in the real installation.

That is why EMC methods define the setup. They define the source, load, mode, impedance, ground reference, coupling method, and decoupling method.

For conducted RF immunity work, coupling/decoupling networks are used to inject common-mode disturbances into defined cable types while providing a defined common-mode impedance. That is a very different idea from interpreting raw S21 through an improvised fixture as universal choke performance.

Why Y21 Is the Better VNA Method

The Y21 method still uses VNA data, but it does not pretend that S21 LOGMAG is the answer.

Instead, it converts measured two-port S-parameters into an admittance matrix and extracts the series impedance of the device under test.

In the K6JCA/W1QG formulation:

ZDUT = -1 / Y21

Or, in terms of S-parameters:

ZDUT = Z0 × ((S11 + S22 + S11S22 - S12S21 + 1) / (2S21))

That is not the same as the simpler S21-only approximation:

ZDUT(S21) = Z0 × (2 / S21 - 2)

The difference matters.

The simple S21 method assumes a cleaner series model than the fixture often provides. The Y21 method is more defensible because it extracts the series impedance from the two-port network and can account for shunt parasitic paths that the simple S21 interpretation cannot separate.

The Corrected Interpretation

A two-port VNA common-mode choke jig can be useful as a rough comparison tool, especially at lower HF, when the same fixture, cable layout, calibration, and environment are used every time.

But raw S21 LOGMAG is not a valid measurement of common-mode choke attenuation.

It is a fixture-dependent transmission number.

The error becomes worse as frequency increases, because fixture capacitance, crosstalk, connector geometry, cable routing, unused connectors, and environmental coupling become part of the result.

The error becomes worse again when the same method is extended from coax to Ethernet, USB, audio, and DC power connectors, because common mode on those cables is not simply “current on the outside of the shield.”

For serious choke characterization, report:

ZCM = R + jX

Preferably extract it with the Y21 method or measure it in a defined EMC-style common-mode setup.

Do not sell or interpret raw S21 LOGMAG as universal common-mode choke attenuation.

The Bottom Line

The issue is not whether a small VNA fixture can produce an interesting graph. It can.

The issue is whether that graph is being interpreted as “common-mode choke attenuation” without defining the measurand.

A choke is not judged by a magic dB number. A choke inserts impedance into a current path.

The physics still asks the same questions:

• Where does the current flow?
• Where does it return?
• What impedance does the choke insert in that path?
• What part of that impedance is resistive?
• What part is reactive?
• Where are the fixture parasitics?
• Does the setup resemble the real installation closely enough to matter?

Until those questions are answered, an S21 attenuation trace is only a trace.

It is not, by itself, proof of common-mode choke performance.

Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.

Questions or experiences to share? Feel free to contact RF.Guru for practical RF and antenna measurement support.