Issues with high SWR on HF Kits EFHW 49:1 UnUn kit

Question

I recently tried assembling a EFHW kit, and some problems came up in testing. I believe this is the kit; I bought it through ARRL: HF Kits website. This is my first time making an UnUn and EFHW kit.

For this I have a nanoVNA running 10 segment sweeps on the UnUn with two gator clips on each end. Between the gator clips sit one 100 ohm resistor and two 1.2k ohm resistors which my Fluke tells me gives about 2420 ohms of resistance, which is close to the 2.4k ohms resistance I've seen used in YouTube vids to test this.

The results leave a lot to be desired. SWR is between 6 and 14 on a sweep with a nanoVNA when done between 7Mhz and 30Mhz. Obviously this is nowhere near what it should be if it's resonant on 40-10m bands.

I'm looking at the windings in particular. One of my two main windings crosses over the top of the other. Would that be a problem? I think I have 14 turns around the toroid here. I've got the little capacitor in there for the higher frequency bands, though it does look like some of the coating cracked in an effort by my to get everything to fit correctly. Otherwise continuity checks out with everything showing continuity to everything else. I also noticed that I had a weak solder joint on the PL-259 connector. That should be easy to fix, but I'd like to do any toroid rewinding before resoldering it.

What am I not seeing here?

Here's some photos of my work.

resistors: general setup: primary windings: nanoVNA:

Answer 1

I'm going to guess that the crossed windings are having more effect than you think. They are the primary winding, of which there are only two turns, so a slight difference can affect the impedance of the secondary winding a great deal, as the impedance transformation is proportional to the square of the winding ratio.

But before going ahead with surgery, some of the other less invasive suggestions here would be worth trying, including checking everything with a DVM - that there are short and open circuits as expected, and that the resistance seen at the ends of the alligator clips is what you expect.

Lead length might also be an issue, especially on 10m. It might be worth soldering the resistors and then affixing the ends of the 3 resistor assembly directly on the binding posts to see if it changes anything. Resistors are cheap.

73 & GL, Chris VE3NRT

Answer 2

Assuming you've calibrated your VNA, the test point impedance appears low, which might indicate a too high transformer ratio. So check for shorts between turns anywhere your primary windings cross over or are touching each other or anything else.

Answer 3

To get such a high SWR something must be badly wrong with the balun or the test setup, it's not just the spacing of turns or the length of your crocodile clips.

General hints for investigating problems like this:

Check your VNA calibration. When it goes wrong, it can give plausible but strange results.

• repeat the calibration
• confirm that the open connector gives a nice small dot on the open circuit side of the Smith chart.
• connect the balun and short-circuit the connector with a screwdriver or knife. Confirm that you see a small short circuit on the left hand side of the chart.
• consider a port extension if you need to connect a long cable after calibration. This won't change the reported SWR but will fix the impedance. See this answer for some details on how this works.
• try your 100 ohm resistor on the VNA - connect it to the PL259 with your fingers or crocodile clips, it should show a small dot 3/4 of the way to the right of the Smith chart.

It's not clear from your photo how long the wires to the load are. Ideally you would just solder the resistors to the balun, or use a few inches of wire. Regular 1-foot crocodile clips are probably acceptable for HF though you may see the SWR change slightly at 30 MHz. A long piece of wire or twin-flex will change the high impedance of the load to something much lower, within about 1/8 of a wavelength or 1 metre in the 10 m band.

Carbon film resistors like that should be absolutely fine for this sort of test.

It is not clear from your photo that the 50 ohm side of the transformer is wired correctly, can you confirm that bit?

• Both wires start on the ground screw of the connector.
• Both make two turns.
• One of them connects to the centre of the connector.
• The other continues for a further 12 turns (that part looks fine).
• Crossing them over each other at the beginning isn't ideal but shouldn't matter too much.
• The manufacturer's web site shows a neat diagram of how it works.
• Twisting them together probably isn't important, but it won't do any harm and you can't mix them up anyway.

It's not necessary to use many sweeps - the 101 or 201 points of the nanoVNA are enough. I find the calibration more difficult and confusing using the PC software. But it is the best way to get screenshots.

In my experience an SWR of 7:1 is about as bad as it gets with a mis-wired transformer. You need to be lucky to get it worse than that. So I suspect you have a dramatic wiring mistake.

Answer 4

Based on the photos you provided, it appears that the only connection to the ground/counterpoise stud is the capacitor. If you look at the diagram, there is supposed to be a wire connecting the stud and the grounding point of the SO-239 connector.

Also visible here: