VNA fixture for measuring Zcm of a common mode choke – coax wound

A common online question is what sort of fixture is appropriate to measure the common mode impedance of a common mode choke.

Above is a screenshot from a Youtube video by Trx Lab, probably about 2016 vintage. I see many problems with the fixture, lets start with the resistors.

Online user yvandelaserge posted his schematic of that fixture. His schematic shows the balun core wound with coax, the pic is not real clear to me, but I will deal with the coax case as shown in the schematic as it is commonly shown online. I might add that is a simplified schematic, and does not capture non-ideal aspects of the fixture.

This has been popular with many ham authors over the years, and there is a large amount of stuff based on measurements using this type of fixture.

Are the 25Ω resistors needed or do they actually degrade the fixture?

If you lack the skills to solve this ‘by hand’… and it is complicated as I will explain later, you could resort to an appropriate circuit analysis tool.

I will use SimNEC v2p1b9, using a RUSE block to model the fixture setup, the coaxial transmission line, and a user specified Zcm component (R5). For this analysis I will use 25+j0Ω for R1-4, and 1000+j0Ω for R5 (zcm).

Note that the model is of a common mode choke wound with a single coaxial line, it is not appropriate to other configurations.

Above is the RUSE block description. Be aware that the TL element models ONLY the differential mode of the transmission line.

Above is the SimNEC model. Let’s look at the Plot calculations next.

Above is the code in the G.plt window. It calculates some currents of interest, finds the voltages at each end of the fixture and the current through the fixture, and from those calculates Zcm of the fixture with embedded choke.

Note that Zcm @ 30MHz of fixture with 1000+j0Ω choke is 1.031K+j10.13Ω, it is NOT 1000+j0Ω so the fixture distorts the impedance of the choke itself.

The observant reader will have noticed that current in R1 (ir1m) is not zero, the magnitude is almost half that of R5 (ir5m) which is the common mode choke. This configuration has excited a relatively large differential component, yet we are trying to excite ONLY common mode component. This significant differential component that is excited leads to the non-ideal behavior of the model.

Above is a similar model, but with R1-4=0.

Zcm @ 30MHz is calculated to be 1K+j75.92zΩ. zΩ are zeptoΩ, 10^-21 ohms, so Zcm @ 30MHz is 1000+j0Ω. Exactly what was modelled, and the simulated applied voltage divided by current gives the correct value.

Above is a variation on the theme where R3-4=0. It suffers a similar problem. Returning to the model discussed earlier…

A different response, but Zcm of the fixture is not that of the embedded choke.

Short connections

The pictured fixture has longish connections.

Long connections ruin broadband performance.

If you shorten the connections and can measure a change, then they were too long (and they may still be too long).


The popular fixture espoused by online experts results in an end to end common mode impedance that is different to that of the embedded choke. Naively one might opine that subtracting the equivalent series resistance of the four resistor network (25Ω) corrects the error… but it does not.

Replacing the resistors with a short circuit fixes the problem.

The fixture is favored by people who just do not understand linear circuit theory and transmission lines, or are not thorough in checking their assertions. It is probably mostly blind copying.

As mentioned earlier, there is a lot of published stuff that relied upon this fixture with or without incorrect correction.

The example fixture has excessive length connections.

An exercise for the reader

Does the same thing occur if the core is wound with twisted pair that is well represented as a uniform two wire transmission line?

Are the resistors beneficial?

Do they degrade fixture behavior?

Then, why are the used so often?