How to match low impedance of NVIS Dipole?

Question

I want to match a 50-ohm coax feed wire to a 40-meter NVIS half-wave dipole. NVIS antenna examples I've found are usually based on a regular dipole (72 ohms in free air). However the impedance drops by a factor of four (to 12 ohms) when it is lowered to 7 feet (https://www.w9xt.com/page_radio_gadgets_nvis_antenna.html). (Note: the author's numbers seem not to correlate with his factor, but at any rate the impedance drops very low).

Just for reference, a folded dipole in free air is between 200 and 300 ohms (https://www.w8ji.com/folded_dipole.htm). Therefore, its impedance would drop by a factor of four to 50-75 ohms, making it closely match 50 ohm coax.

My idea was for a regular dipole (not a folded dipole) and use a 4:1 balun to raise the antenna impedance by a factor of four from 12.5 ohms to 50) to match the coax and transmitter impedance.

Problem: Maybe some baluns are not designed to be reversed because the side designed to be high impedance balanced antenna would then be on the unbalanced coax, and vice versa with the low impedance side. Furthermore, transferring power the side driving the low impedance antenna might draw more current than the balun wire was designed for.

I have not found answers in any Radio Amateur's Handbook (I have six ranging from 1956 through 2012), or on the Internet.

So what is the proper way to match 50-ohm unbalanced coax to 12.5 ohm balanced dipole?

Answer 1

Many baluns will work just fine in either direction, though there isn't just one kind of "4:1 balun".

This kind is wound on two cores, and works as a common-mode choke:

^{simulate this circuit – Schematic created using CircuitLab}

A common-mode choke works in either direction, so it matters not which end is balanced and which is unbalanced (or if both ends are unbalanced, or both balanced, for that matter.)

However this kind is wound on a single core and relies on ground potential being at the midpoint of the balanced terminals:

^{simulate this circuit}

This can work OK given a well-balanced load on the right, and an unbalanced load on the left. But flipped the other way (with the balanced load on the left) it will try driving each terminal of the balanced load to different voltages relative to ground, so you'll get quite a lot of common-mode current.

Answer 2

From personal experience, a good 1:4 balun (50 to 200 Ohm) doesn't necessary work as a good 4:1 balun (50 Ohm to 12.5 Ohm). You can wind just a 4:1 transformer and use it with a 1:1 balun though.

Another simple way to match any impedance is to use an LC-network. There are a lot of online calculators. Personally I particularly like this one. Using this approach I matched my 40m delta loop antenna to all bands from 10m to 80m using a separate LC-network for each band.

Answer 3

Take three 1/4 wave sections of 75 ohm coaxial cable (let's say RG11) and solder together. You now have a 1/4 wave section of 25 ohm. Doing the math (25 * 25) / 12.5 = 50 ohm. Now use any 50 ohm coxial cable length to the radio. Ps: don't forget to take into account the 75 ohm cable speed factor to calculate the 1/4 wave section. Good luck. Luiz PY4ACP

Answer 4

Where did you hear that a 40m center-fed dipole λ/4 high has a feedpoint impedance of only 12.5Ω? It will be 75Ω or more. A center-fed λ/2 dipole is only 50Ω at one height.

Here are some graphs. As you can see, you don't need a 4:1 balun.

Ignore the top graph in this first image.

---

The graph below is based on theoretical values

---

75Ω coax will be a better match, and the worst case VSWR mismatch to your rig will be only 1.5:1.

Answer 5

How to match a 12.5 ohm balanced antenna to a 50 ohm unbalanced line... I think that has been well answered but one point worth mentioning, and a relevant anecdote...

There are 2 books by Jerry Sevick directly addressing how to do this: - Sevick's Transmission Line Transformers, Theory and Practice, 5th Edition (2014), by Mack and Sevick - Understanding, Building and Using Baluns and Ununs (2002), by Sevick

The first book is very theoretical, the second --which I have not yet read-- is said to be more hands-on. Both are worth reading.

I happen to have a 12.5 ohm 40m balanced dipole (short dipole, not NVIS). A lot of work went into designing and building, including the 4:1 balun (wound using 25 ohm coax; not easy to find). When I finished and took measurements, I found the impedance was closer to 50 than 12.5 ohms. It works better without the 4:1 balun! Of course it certainly has a strong 1:1 common mode choke.

My point is that you might consider building the antenna first. Measure the actual impedance, then build the balun. You might find that a matching network may work better for you, as mentioned already by @alexanderalekseev-r2auk

Answer 6

A 4:1 Guanella Balun would resolve the issue. All you need do is move the earth connection you are worried about to the other side. That flips it to a 1:4. Easy.

However, the idea that you should lower your dipole to 7ft to create an NVIS antenna is fundamentally wrong. All you will do is kill the efficiency as the ground soaks up your radiated power. A 40m dipole at 10m off the ground is already omnidirectional with plenty of vertical radiation and you really don't want to get any lower. The dipole only starts to show a classic dipole pattern at about 20m high. If you want a balun, then choose a 1:1 Guanella.

Answer 7

Today I made the antenna and included some data, for what it's worth, although this isn't really an answer because the other answers helped me to get to this.

It is a dipole with 33 feet per leg (untrimmed) of insulated #14 stranded wire, with a ceramic insulator and about 10 feet of polyester rope at each end. The height is 10 feet. There are three 8-foot metal T-posts with 3 feet of 1" PVC pipe clamped to the top for RF isolation. I laid 70 feet of extension cord on the ground below the antenna to act as one reflector (more are recommended). The center post holds a plastic box with a current choke made of 13 turns bifilar on an FT140-61 torroid. The feed cable is 66 feet long.

I tested it for SWR in the 40-meter band at 10 watts and was able to get SWR of 4:1 at 7.280 MHz, 2:1 at 7.150, and 1.3 at 7.000 MHz. That indicates to me that the resonance is slightly low so it needs to be trimmed. So it looks like it will tune with good SWR as soon as I can trim it. I believe the height of 10 feet is approximately 0.073 wavelength, so the graph of impedance vs. height indicates that it might be about 12 ohms. I'm concerned that my SWR readings are not right if that's the case. More experimenting is in store.