WA5SWD’s Direct Feed 50-Ohm 2 Meter J Pole
No Adjustment 5/8ths Wavelength 2-Meter J-Pole
I’ve built J-Poles before. I even have a design out there that was an improved mounting and matching version of another Amateur’s design, Jim Weir of RST. But I could never really get a good match. By that I mean one that came down to below 1.1:1 or less at the center frequency—one that worked if you simply built it as shown. I always had to fiddle with it, and still had problems getting a good match. I have built a bunch of them, and they all worked well, but not perfectly—that is, 1.1:1 SWR or less.
Rick Huebner, 
, recently pointed me at a Web Site where the author claimed to have finally de-mystified the process of matching the J-Pole. I believe the author made one totally unjustified assertion, and then proceeded with perfectly good logic and math to his workable but unlikely conclusion. Along the way he discarded as unworkable other correct and practical solutions. This inspired me to find a better way!
J-Pole Defined
So what is the classic description of a J-Pole? Well, it’s a 1/2 wl radiator fed with a 1/4 wl matching section. Since it IS a dipole, it can have no gain over a dipole, despite many claims to the contrary. It is by its nature a BALANCED antenna. Most of us want to feed our antennas with 50 
coax, which, of course, is UNBALANCED. This will present a problem for the purest, but perhaps not for the practical.
The fact that only one leg of the matching section has a load on it will unbalance the currents in the matching section, which causes incomplete cancellation of the radiation from each half. This will distort the radiation pattern, which is generally accepted to mean that the energy at Zero degrees elevation angle will be reduced. We try to help this mismatch of currents with things such as baluns. The better the current match, the more complete the matching section cancellation, and the better the low angle radiation.
I can hear you think: ”That’s nice, Ed, but how can I test how well the currents are matched? Won’t that take a lot of fancy (expensive) test equipment and a lot of time?” Ah! Good point! Most of us cannot afford the time and expense of rigorous testing. What we can do instead is called “Sanity Tests”. You say to yourself: “I can tell if the currents are well balanced by seeing if they unbalance the same way if I do something to unbalance them. The closer the behaviors match, the closer the two currents match.”
Testing
What test equipment must you have, bare minimum? Answer: an SWR Meter. The device you can use to unbalance them? Answer: your finger! (CAUTION: only at low power!) Just run your finger from the base of each conductor of the matching section to the top of the matching section (but do this only if you are using low power, because high power could be bad for you. I use the MFJ-269). The SWR will rise smoothly from base to top, and should do so in a similar fashion on each conductor. The more similar the SWR pattern, the more similar the currents.
There are coupling issues between the conductors of the matching section which can also be determined by this test. If the two sides are not tightly coupled to each other, then whatever you have built may work after a fashion, but it is no longer a proper J-Pole. You can find improper antenna mounting problems with this test. Anything that upsets the current balance will cause a deterioration of the radiation pattern.
For example, I found that a certain commercially available dual band J-Pole antenna failed this simple test miserably when mounted on a metal mast. Unplanned currents were coupled into the metal mast which totally upset the balance. I then changed it to a nonconductive mast and re-tested. With just a two-turn “choke balun” around the non-conducting mast, I achieved an acceptable balance. This gave it at least a fighting chance to radiate properly. I informed the manufacturer of this problem, and he scoffed at my information. Very strange response, since my test is so easy to apply.
In the real world, J-Poles seem to deliver excellent performance, despite the admitted reduction of low angle radiation caused by the effects of the matching section. Perhaps they are no worse than other verticals. Please consider that even a simple ground plane antenna really requires an extensive ground plane to work well at low angles of radiation. The textbook patterns are taken over an infinite ground plane, which is rather hard to mount.
I don’t have the set-up to measure antenna patterns at present, or I would certainly do so. Remember, it is the radiation at a zero degree elevation that we primarily usually use for line of sight communications at VHF and above. But a 1/2 wl vertical radiator is not the configuration that delivers the most energy on the horizon. According to the books, a 5/8 wl radiator delivers about 3 dB more. This is approximately equivalent to doubling the transmit power at both ends, since the increase works both for our transmitter and our receiver.
A 5/8ths wl radiator presents an unusual matching problem, since we now must deal with the non-resistive component caused by not being an integral multiple of an electrical 1/4 wl. This can be handled by a modified matching section, decidedly NOT 1/4 wl. As a point of interest, please be aware that for a particular antenna impedance there can be only one unique set of values of length and impedance of the matching section for a perfect 50Ω match.
Since we are starting with values from graphs and will be doing a graphical impedance transformation on a Smith Chart, we will have some built in error. Without the (expensive) test equipment to actually measure the antenna impedance ourselves, we will have to be content with a good match, not perfection. I think 1.1:1 SWR is close enough for our purposes, and that we can easily obtain.
I built several J-poles before arriving at the design here and I discovered a lot of things along the way. First, all of the J-Poles I checked are not electrically 1/2 wl. Why? Because the length of the radiator was calculated by the formula for a CENTER-FED 1/2 wl dipole. This means a length about 95% of a free space wl. was used. The correct value for an END-FED 1/2 wl is much shorter. From the work done decades ago, more like 80-83% of a free space wavelength for the sizes we work with at two meters. Thus they are all about 17-20% too long!
Second, we are not working in a vacuum. We are working in air. This also has a slight effect, but it increases the total error. This error amounts to a shift of about 400 KHz at two meters. This is not insignificant.
Third, and this may be news to some: an end fed dipole is not 5000 except for an infinitely thin dipole. This is a theoretical construct, not useful for real antennas, especially for VHF and up. What IS the impedance we will see? A lot lower, depending on the length divided by the diameter of the antenna. For most real two-meter antennas, 500-800 ohms or so will result. More on this later, as we pursue the answer.
What does all this mean? Your “1/2 wl” J-Pole is too long, but this gives it more gain because it is not longer than 5/8 wl. Being long, it is not a resistive load and is reflecting a reactive component back into your 1/4 wl matching section. Usually, this is why you can’t get a perfect match. The 1/4 wl transformer is not the right length to take care of this problem and needs to be somewhat shorter than 1/4 wl.
Data Reference
How dare I make such rash statements, in the face of such a vast array of articles, Amateur ‘experts’ and manufacturers who claim to understand J-Poles to the last dot? Because many of them are flawed and I have other PROFESSIONAL references in support of what I say! I did not take nor make up the data. I was far too young to know or care when it was taken. I have checked it, and my antennas work just as that particular data predicted.
Okay, I’ve stuck out my neck and have made what many may consider as ignorant remarks. But, here is the reference for your consideration. Go to Johnson and Jasik Antenna Engineering Handbook, Second Edition. Turn to page 4-7 and 4-8. Study Figures 4-3 and 4-4. Note that for the greatest A/D ratio shown, the maximum impedance for a A/D of 1000 is less than 1100 Ω—not 5000Ω. At two meters, we will be much nearer an A/D of 100. Note the maximum resistance for this curve is only 750Ω. These curves are worth enlarging and saving for reference. Again, please DON’T take my word for it, but check these references and try it.
I have actually measured a 48-inch (1.2m) monopole with an OD of 0.895 inches (2.27cm) [A/D of 55] with a Boonton RX Meter. The curves led me to expect a resonance at 101.46 MHz with a Resistance of 585. I measured 580 at 102.5 MHZ. This is an error of about1 %. Not too bad for a lash up in my garage with an obsolete instrument that has not been through a calibration lab in many, many years!
Later I will detail my 5/8ths J-Pole and those who are only interested in building one may prefer to skip ahead to that part. But, if interested in how to make a design of your own, please continue reading this part.
Designing
Start by designing a 1/2 wl monopole. Estimate the A/D you expect to get. Just get close, and don’t worry too much. You will have to iterate a couple of sets of calculations, but you will get the correct A/D value quickly. For two meters, 1/2 wl is about 40 inches (1m). Divide that by the diameter of the material you intend to use, such as 3/8-inch (.95cm) rod. For 3/8 rod, this gives an A/D value of 106. Use the A/D curve for 100. Now you must decide if you want to consider the antenna resonant at the length for maximum resistance or for zero reactance. The lengths are slightly different. Since we are going to use a 5/8 wl radiating section the radiator will be reactive anyway, so I chose the maximum resistance value. Multiply the 1/2 wl length in degrees from the chart by 1.25 to get the length required for a 5/8ths radiator. Go to this point and note the resistance and the reactance values.
Since we are feeding a balanced antenna, it has a “virtual ground” right up the center of the matching section. Because of this, we must multiply the resistance and reactance values by four to correctly compute our matching section impedance.
Now comes the fun. You must find the proper values for the matching section and need to find the impedance and length. To do this requires using a Smith Chart. Smith Charts are powerful tools, and I am just beginning to use them. I have a lot more to learn, so please go to the ARRL Antenna Handbook or other source and read up on them. (Editor’s Note: Coincidentally, we have a special three-part series on the use of Smith Charts by Dr. David Jefferies with Part 1 in this same issue of antenneX for October 2000. This article should be very useful to the reader in following what WA5SWD is doing here.)
Your task is to find a Matching Section Z and length that rotates the load you have back to the resistive axis and have it present a 50Ω load. I chose an iterative approach. I assumed an initial impedance of 300Ω, normalized the Z from the antenna to this value, and plotted it on the Smith Chart (don’t forget the 4X). Then, I rotated this back to the real axis. At the intercept point, I noted the new value and multiplied by my chosen Z. This gave me a resistive value, which I wanted to be 50Ω .
I noted the value of the matching section Z and the load resistance it presented. I repeated with different values until I got as close to 50 Ω as I wished. There are perhaps better ways to do this, namely the electronic Smith Chart programs, but I have not personally tried one as yet (Editor’s Note: See web links in the Dr. Jefferies article to several sites that contain these online tools.)
Having found the desired Z for the matching section we must now calculate the spacing required to give us the impedance we desire. This is not difficult, but the formula uses a hyper-cosine, so you must have a calculator that handles these.
From Reference Data For Radio Engineers, Fifth Edition, page 22-22 (D), the formula is:
Z0 = inverse of cosh (D/d) or approximately Z0 = 276 log(base10) (2D/d)
“D” is the center-to-center spacing of the conductors, and ‘d’ is the diameter of the conductors.
I have calculated a few handy Z0 vs. spacing ratios, listed below.
| D/d | ZO Ohms |
|---|---|
| 1.1 | 53 |
| 1.5 | 115 |
| 2.0 | 159 |
| 2.1 | 165 |
| 2.5 | 188 |
| 3.0 | 212 |
| 4.0 | 248 |
| 5.0 | 275 |
| 6.0 | 297 |
| 7.0 | 316 |
| 8.0 | 332 |
| 9.0 | 347 |
| 10 | 359 |
| 11 | 371 |
| 12 | 381 |
| 13 | 391 |
| 14 | 400 |
| 15 | 408 |
| 16 | 416 |
| 17 | 423 |
| 18 | 430 |
| 19 | 437 |
| 20 | 443 |
| 22 | 454 |
| 25 | 469 |
| 30 | 491 |
One might infer that we have no limits for the diameter of the elements. This would be wrong. Remember, we are making a balanced transmission line, and that they tend to leak when the spacing exceeds 0.1 wl. This is guaranteed to foul things up. For two meters, a wavelength is about 80 inches, so we cannot use a conductor spacing of greater than 8 inches. In some cases, you will be forced to start your design over with a smaller diameter conductor.
Back to Our Smith Chart
Note the point where you plotted your starting impedance, the one that is four times the values you read from the resistance and reactance graphs for A/D ratios. Draw a line from the center of the Smith Chart through this point, extending it until it crosses the outer edge of the chart. There you will note several scales. Look at the one labeled “Wavelengths Toward Load”. Note the value, which will be close to 0.2 wl. The value you read is the length of the matching section required to rotate what you have at the antenna to what you need to feed it with 50 line. Do the calculation exactly, but the length will come out to be shorter than a quarter wave, assuming you are building a 5/8 wl. radiator J-Pole. Now change those degree values to physical lengths.
One more thing: I have found that this procedure fails to account for the length of conductor from the coax to the elements. Thus your elements will be long by about 1/2 the spacing. Also remember that it is the difference in length between the two elements that forms the radiating portion. You cannot simply change one without changing the other by the same amount or all of your calculations fly out the window. This is true of every J-Pole, not just my 5/8ths wl. design.
If you start a bit long, you can measure, then cut the same small amount from each and measure again. This way you can ‘walk’ your way in to the correct values. Then duplication is easy. If you do cut the elements a bit short, you can lengthen them for testing purposes by using the outer braid from RG-8 as a conductor. Since it is a mesh, you can snug it by stretching it and it will remain in place nicely. I even formed the end over and soldered it closed to make a better simulation of the 3/8ths aluminum rod. I wanted to use 3/8 inch Aluminum rod for the finished antenna so I used 1/4 inch copper tubing for the experimental antenna, as they have the same outer diameter. This is not expensive, and is easy to work.
Okay, I have described what I have learned by building my 5/8ths wl. 2 meter J-pole. I started with information in PROFESSIONAL reference books, not just “rules of thumb” or rough guesses. I did not utilize unsubstantiated approximations as my initial starting point. This concept is much more likely to lead to a successful outcome. Remember, it only takes one error in a series of calculations to invalidate your results.
Building
And now, for those of you who made it through all of this (or skipped ahead), here is the information you need to build your own copy of this design (all commercial rights to the designed reserved).
Since this is a balanced antenna, we need an insulated supporting structure. If I had access to plastic or fiberglass, I would have used that. Instead, I chose a hardwood. This was low moisture and inexpensive, and can be sealed with epoxy to improve its ability to withstand weathering. The “F” frame is simple to build, and only has one critical set of dimensions. These are the spacing of the rods and the location of the feed point. Follow my sizes closely and you will have no need to tune this antenna. Your reward will be a two-meter antenna that gives you a SWR of less that 1.5:1 across the entire two meter band. I typically get 1.3:1 at one end, and 1.4:1 at the other. Since I am using an MFJ-269 as my test instrument, I can’t guarantee that you will get exactly the same, but you should get very close. The minimum SWR will be very close to the middle of the band, and will be about 1.1:1 or lower.
Refer to the Figures 1, 2 and 3. Build the parts of the “F” support. Be certain to stack drill the 3/8-inch holes for the elements so they align correctly. Make the PCB by shearing it to size, then drilling all of the holes except the 3/8ths holes for the elements. You will do that after you have soldered the SO239 in place, then mounted the PCB to the bottom arm of the “F”. Be certain to cut the copper and remove it per the drawing or you will be shorting the elements together.
With the SO239 soldered in place, drill through its mounting holes to permit the #6 screws to pass. Solder an insulated wire from the center of the SO239 to the 1/16-inch hole in the PCB. Notch the lower arm of the “F: to clear this wire and allow the PCB to mount flat against the lower arm. Drill pilot holes for the #6 screws, align the PCB carefully to the lower arm and install the #6 screws.
NOW drill those 3/8-inch holes, and they will line up perfectly. Make the long and short elements and thread them 3/8-24 for 1-1/2 inches. Stick the threaded end through the top arm of the “F” and run a nut on the thread. Pass the arms through the lower arm and secure the arm with another 3/8-24 nut. The bottom of the element must be flush with the bottom of the nut if you are using a standard nut.
Check the SWR. It should be less than 1.5:1 across the band, with a minimum SWR at or below 1.1:1 at about the center of the band. The 2:1 SWR points will be about 140 and 150 MHz.
When you are satisfied with your work, coat all of the wooden parts with some weatherproofing such as clear epoxy or a good varnish. Several thin coats work better than one thick one. When all is dry, recheck the SWR. You never know. Enjoy!
Originally posted on the AntennaX Online Magazine by Edward Lawrence, WA5SWD
Last Updated : 4th June 2024