W8JI’s “Properly Testing Baluns” Revisited
By ON6URE (RF.Guru)
A reader recently sent me one of the most-circulated pages in amateur radio antenna land: W8JI – “Properly Testing Baluns”.
If you’ve been around HF long enough, you already know why that article became “famous.” W8JI has helped a lot of hams avoid bad assumptions and bad hardware. And yes, I’m going to say this clearly up front: there’s a lot in W8JI’s balun test article that I agree with and still recommend people to read it.
But I’m also going to say the other thing that needs saying: some of the measurement culture around baluns and common-mode chokes is outdated, incomplete, or simply measuring the wrong thing. In particular, many “balun/choke measurements” fail because they do not correctly account for parasitics and test-fixture effects — and they do not use mode-correct methods such as the Y21 approach.
Why I’m revisiting W8JI’s article
I’ve written a lot at RF.Guru about baluns, ununs, common-mode chokes, and measurement myths: Baluns and Ununs – Technical Articles.
Some topics overlap heavily with W8JI’s page — and in those areas, you’ll see us landing on similar conclusions: many commercial “baluns” are marketing first and engineering second, and SWR alone is a terrible judge of function.
Where I diverge is mainly on how we should measure and what results we should trust in 2026. This isn’t a takedown. It’s a revisit — with nuance — because W8JI is a respected ham and his work is still widely referenced. My goal is simple: keep what’s correct, fix what needs updating, and help you measure what actually matters.
What W8JI gets right (and I agree)
SWR is not a balun test.
W8JI is absolutely correct to warn that low SWR does not prove you have a functioning balun or a functioning choke. SWR can look “fine” while common-mode current is still riding the outside of your coax and making your feedline part of the antenna (and part of your RFI problem). That message is still not widely understood — and it’s one of the best reasons his article remains worth reading.
Balun function is about system behavior, not a “pretty match.”
A choke/balun’s job (in the real world) is to reduce unwanted common-mode current. If your test setup doesn’t actually excite common-mode, you can end up “proving” the wrong thing. W8JI’s skepticism toward oversimplified tests is healthy and justified.
Many commercial baluns deserve criticism.
I’ll say it in my own blunt RF.Guru style: a lot of commercial baluns are built to hit a price point and a marketing claim, not a real engineering target. W8JI calling out bad baluns is not controversial — it’s necessary.
Power handling and heating matter.
If you transmit real power, core loss and temperature rise become the reality check. Small-signal VNA measurements do not automatically predict what happens under mismatch, continuous duty, or high reactive currents. The “no smoke = good” mentality is wrong, and W8JI is right to push back on that.
Where W8JI’s article needs a modern measurement update
This is the important part — and it’s where I’m going to be direct.
W8JI’s page discusses measuring choke impedance using lab gear and a low-capacitance fixture, and it shows results and setups that many hams have copied for years.
The problem is not the intent. The problem is that in many choke measurements, parasitics dominate — and you can build a “low-capacitance fixture” and still end up with data that looks scientific but is not actually the choke’s real common-mode impedance.
Here’s the core issue in plain language:
- Common-mode chokes are often high impedance devices (kΩ range) when they are doing their job.
- A few pF of stray capacitance in a fixture or cable is not “tiny” at HF.
- At 10 MHz, 3 pF is about 5.3 kΩ of capacitive reactance. At 30 MHz, it’s about 1.8 kΩ.
- So the test setup can easily become the limiting impedance — and your measurement becomes “fixture + DUT” instead of DUT.
This is why I do not accept many classic one-port or casual fixture plots as authoritative. Not because the authors were dishonest — but because the method does not robustly remove the fixture’s parasitic shunt admittance from the result.
The missing piece: measuring common-mode impedance the way it should be measured
At RF.Guru, we strongly favor the modern ham-accessible method that actually solves the fixture-parasitic problem: the Y21 method.
The key idea is simple: instead of hoping parasitics are “small,” you model them and mathematically separate them.
In practical terms: you do a 2-port VNA measurement, convert to Y-parameters, and extract the choke’s series common-mode impedance cleanly (with parasitic shunt admittances handled instead of ignored).
If you want the deep technical explanation and how-to, start here:
- Measuring common-mode chokes with the Y21 method
- Why the Y21 method is the only ham measurement that actually works
And if you want the uncomfortable truth about how often VNA choke tests are accidentally measuring the wrong mode (differential instead of common-mode), read this: Debunking common myths in common-mode choke measurements with a VNA.
What I think happened
This is where I want to be fair to W8JI.
The conceptual warnings in his article aged well: SWR isn’t proof, many baluns are bad, and the system matters. What aged less well is how the broader ham community treats choke measurements:
- People copy an old fixture, run an S11 “impedance” plot, see a peak, and call it “the choke’s impedance.”
- They ignore (or don’t realize) that a few pF of shunt capacitance can create fake resonances and fake roll-offs.
- They end up comparing chokes using measurement artifacts, not real ZCM.
This is exactly why I push the Y21 approach so hard. It is not trendy — it is necessary if you want results that mean what people think they mean.
Current balun vs voltage balun without the religion
Balun discussions can turn into “teams” and ideology. I’m not interested in that. I’m interested in what happens on your feedline.
- A voltage balun can transform impedance and enforce equal voltages, but it does not automatically stop common-mode current.
- A current balun / common-mode choke is specifically there to raise the impedance of the unwanted common-mode path, so the outside of your coax stops acting like part of the radiator (or part of your noise pickup).
If you want the short “baluns explained like an engineer, not a brochure,” start here: Baluns in a Nutshell.
How much choking impedance is enough?
W8JI makes a valid point: the “required” choke impedance depends on the system. Geometry, routing, frequency, and environment all matter.
Where I add nuance is that hams still need practical targets to build and buy intelligently.
In many real coax-fed HF situations, a choke that maintains several kΩ of common-mode impedance across the bands you care about is a very reasonable engineering target. If you want the practical guidance version (and the math behind the usual dB claims), read:
The takeaway is simple: impedance is a property of the choke; “dB” is a property of the choke inside a specific system.
How I recommend testing today
Use quick functional sanity checks.
W8JI’s basic functional checks and “does it behave like a current balun?” style tests can be useful as a first filter. They catch obvious junk fast.
Measure common-mode impedance with a method that survives parasitics.
If you want a ZCM(f) curve you can defend, use Y21 with a 2-port VNA and a fixture designed for repeatability and mode correctness.
Validate under real power and real installation conditions.
Small-signal bench plots are not the whole story. Confirm temperature rise under realistic duty, and confirm that common-mode current on the coax shield is actually reduced where it matters in your station.
My bottom line for readers who found W8JI’s article
If you’ve never read W8JI’s page, read it — especially for the conceptual warnings: Properly Testing Baluns (W8JI).
Then read it again with this modern filter in mind:
- What W8JI says about SWR and balun function is still correct.
- What many people think they learned from classic choke plots may be contaminated by fixture parasitics.
- If you want to measure chokes properly today, use a mode-correct method like Y21.
- And always remember: the antenna system is the final judge, not a screenshot of a VNA trace.
Mini-FAQ
- Is SWR a valid balun/choke test? No. SWR can look perfect while common-mode current still turns your coax into part of the antenna (and part of your noise/RFI path).
- Are one-port (S11) “choke impedance” plots useless? They can be useful for quick comparisons and spotting obvious resonances, but they often become fixture-dominated once impedances reach the kΩ range.
- Why do fixtures create fake peaks and roll-offs? A few pF of shunt capacitance can be only a few kΩ at HF, so your measurement becomes “fixture + DUT” unless you explicitly separate the parasitics.
- What’s a reasonable HF choke target? “Several kΩ” across the band(s) you care about is a practical, defensible target for many coax-fed installations — with the reminder that installation geometry always matters.
- What does the Y21 method change? It treats the setup as a network and mathematically extracts the choke’s series common-mode impedance instead of hoping stray capacitance is negligible.
Interested in more technical content? Subscribe to RF.Guru updates for deep-dive RF articles and lab notes.
Questions or experiences to share? Contact the RF.Guru team.