Broadbanding 80m Dipoles, Monopoles, and Loops

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by AC6LA, Oct 12, 2018.

ad: L-HROutlet
ad: l-rl
ad: MessiPaoloni-1
ad: L-MFJ
ad: Left-2
ad: Left-3
ad: Subscribe
  1. AC6LA

    AC6LA Ham Member QRZ Page

    Here's a new take on an old idea.

    Frank Witt, AI1H, wrote several QST and Antenna Compendium articles a few decades ago describing the use of "transmission line resonators" to create a broader SWR curve. The technique is also described in Chapter 9 of recent editions of The ARRL Antenna Book. See references below.

    A transmission line resonator (TLR) usually consists of an open stub, a shorted stub, and a linking section of transmission line between the stubs. One of the more well-known implementations is the 80m "DX Special" dipole (or inverted vee). The outside of the coax braid serves as part of the radiating antenna wire while the inside of the braid along with the center conductor serves as the TLR matching network.


    That's a very elegant design but there is no need to resort to that level of mechanical complexity. A more general solution is to use standard #12 or #14 wire for the entire antenna, connect an open stub in parallel at (or near) the feedpoint, add the linking section of transmission line, connect a shorted stub in parallel, and then continue the transmission line run to the station. See below left. Another approach is to add a half-wave of transmission line at the antenna feedpoint, connect the open stub (probably now at ground level), insert the link between the stubs, connect the shorted stub, and then complete the run to the station. See below right.


    In The ARRL Antenna Compendium Volume 4 Witt describes a fairly complex set of formulas that can be used to determine the lengths of the stubs and linking section. However, the formulas are not exact and he concludes by saying that some manual adjustments of the calculated lengths may be necessary.

    The advent of EZNEC v5, which has the ability to include realistic (lossy) transmission lines in a model, coupled with the AutoEZ optimizer, which can adjust multiple variables simultaneously to achieve a desired "target" result, offers either an alternative or an adjunct to the Witt analytical approach. The following example will focus on an 80m inverted vee but the same technique can be used for a vertical, inverted L, T, short vertical with base loading, or loop, on 80m or 160m or any other band where a broad SWR response is wanted.

    From the zip file below, sample model "AI1H TLR 80m Dipole.weq" is used for flat dipoles, inverted vees, or dipoles with a "catenary sag" in the center (by setting the angle between the arms to greater than 180 degrees). Here are the variables; dimensions are in feet. In this example a half-wave of transmission line has been added (variable K) between the antenna feedpoint and the open stub of the TLR.


    And here is the AutoEZ Transmission Lines table showing the lengths of the various sections as set by variables K through O.


    After making changes to the variables for your antenna's center height and included angle, along with the wire diameter and optional wire insulation properties on the Wires tab, the parameters (Zo, VF, Loss) of the transmission line type you intend to use in the Transmission Lines table on the Insr Objs tab, and the ground characteristics on the Calculate tab, you can proceed directly to the Optimize tab and click Start. The optimizer will adjust variables B, L, M, and N.

    Hint: You need not let the optimizer run to completion. When the Best-Worst delta is consistently less than 0.05% you can click the Stop button. But don't jump the gun. Many times the optimizer will seem to "plateau" for several iterations and then take off again with improvements. And keep in mind that even if you let the optimizer terminate normally the final values will be somewhat dependent on the initial start values.

    Note that the resulting SWR curve will depend on the desired bandwidth. A wider bandwidth will result in slightly higher overall SWR values. Here is a comparison showing the response with bandwidth choices of 3.5 to 3.8 MHz, 3.5 to 3.9 MHz, and 3.5 to 4.0 MHz, along with the SWR response when no transmission line resonator is included.


    There is no free lunch. Adding the TLR also adds a small amount of additional loss, just like adding a conventional lumped-element matching network adds additional loss. You can see that loss by comparing the calculated gain with and without the TLR included, making sure that the feedpoint-to-source length of transmission line is the same in both cases. That is, calculate a sweep with the TLR included (Q = True), take a snapshot of the "Slice Max Gain" on the Custom tab, then set Q = False and set variable O to the length of the now-excluded TLR link section. Do another calculation and compare the loss curves. The difference in gain is equal to the loss in the TLR.

    Here is a gain comparison assuming the desired bandwidth is 3.5 to 3.9 MHz.


    It is also interesting to see the impedance transformations that take place when using a transmission line resonator matching system, starting at the feedpoint and working backward toward the station. In the following Smith chart, increasing frequency is shown as clockwise rotation on each colored arc. As with the previous illustrations this example TLR has been optimized for 3.5 to 3.9 MHz so the final (most clockwise) four dots in each arc, representing frequencies above 3.9 MHz, can be ignored.


    As mentioned previously, Witt developed a set of equations to calculate the lengths of the open stub, the linking section, and the shorted stub. The scratch pad area of the Variables tab includes the Excel formulas necessary to do the Witt analytical calculations. Note that if the half-wave of transmission line between the feedpoint and the TLR had not been used, as would probably be the case when broadbanding a vertical with a ground-level feedpoint, the Witt calculations would be much closer to the final optimized result.


    You can use the Witt calculator to compute starting values for the optimizer if you like. You can also use it to get an idea of the stub and link lengths for the various permutations of a transmission line resonator matching network. In the basic TLR the sum of the lengths of the open stub, shorted stub, and link section will be approximately one-quarter wavelength at the bandwidth center frequency, taking velocity factor into account. Other variations will have the total TLR length (stubs plus link) close to two-quarters or three-quarters wavelength. The scratch pad area includes a brief description of the alternatives.


    You can try optimizing the alternatives and then using the Custom tab Snapshot button to compare SWR curves and/or gain differences. Depending on the geometry of the antenna itself (height and included angle in this example) some alternatives may not be viable.

    You can also compare SWR and/or gain differences when you intentionally alter the optimized lengths by a small amount, say 0.25 or 0.5 ft (3 or 6 inches), to get a feel for how precise you must be when cutting the coax lengths and/or the overall antenna length. And of course everything depends on the model of the initial antenna being accurate in the first place, so the best approach would be to use an antenna analyzer to measure the impedance values (R and ±jX) at the point where you intend to insert the TLR, then adjust the model (with Q = False) to produce those same impedances.


    The zip file below contains AutoEZ format (*.weq) models for several example configurations:

    AI1H TLR 80m Dipole.weq (for dipole, inverted vee, and catenary sag shapes)
    AI1H TLR 80m Monopole.weq (for vertical, vertical with cap hat, inverted L, or T shapes)
    AI1H TLR 80m Short Vert with Base Loading.weq
    AI1H TLR 80m Delta Loop.weq


    Frank Witt, AI1H, "The Coaxial Resonator Match and the Broadband Dipole", QST, April 1989.
    ___, "The Coaxial Resonator Match", The ARRL Antenna Compendium Volume 2, 1989.
    ___, "Broadband Matching with the Transmission Line Resonator", The ARRL Antenna Compendium Volume 4, 1995.
    ___, "Optimum Lossy Broadband Matching Networks for Resonant Antennas", QEX, Sep/Oct 2008 (reprinted from RF Design, April/July 1990).
    ___, Chapter 9, "Broadband Antenna Matching", The ARRL Antenna Book, 19th and later editions.

    Dan, AC6LA

    Attached Files:

    AI3V, KK4OBI and KD6RF like this.
  2. WA7PRC

    WA7PRC Ham Member QRZ Page

    According to the free design program on (link), my 4-wire cage inverted vee for 80m is supposed to have a 2:1 VSWR bandwidth of 318 kHz:

    Not being sure it'd play as designed, I made it a few feet long and found this VSWR curve:

    I rarely operate in the top 100 kHz of 80m so, I left it alone. With the apex at 70' AGL, it seems to play VERY well. With about 85' of coax feeding it, I surmised the ATU in my rig might be able to tune it on 160m as a vertical w/ top hat. So, I removed the choke at the apex, added a SPDT vacuum relay at the bottom of the feedline, and drove it against a bunch of radials:
    As you can see, the relay coil (12VDC) is powered via a bias tee (from the shack PS) thru the feedline. I've been able to work the other coast using only 100W. IMO, not bad.

    Bryan WA7PRC
    N2EY, KA2RRK and K1TGX like this.
  3. G3YRO

    G3YRO Ham Member QRZ Page

    Is it really worth all that trouble?

    With decent coax, the actual loss at the edges of the band on a normal 80m dipole I would have thought are pretty negligible !

    Roger G3YRO
    Last edited: Oct 12, 2018
    KK4YPQ likes this.
  4. N3DT

    N3DT Ham Member QRZ Page

    Couple years ago, I was playing with inserting sections of RG11 into the feed line at the shack end to lower the SWR to my transceiver. My method was to make a trace of the antenna/feedline with my AA600 then using that data, go into the 'add cable' and using RG11 I could easily add lengths by just clicking the up arrow on that tab. I found a piece of roughly 50' RG11 would broadband my 80M dipole across the band under 2:1. Tried it and sure enough it worked just like calculated. I mentioned it a year or so ago and was reminded that this same method was used in a QST article that I had read and forgotten. There also is a formula for feed line length and RG11 length, it depends on how long your 50Ω feed line is and the characteristics of the antenna. But using the AA600 and it's software, it was done in about 2 minutes without calculating anything. Easy enough to do with you RigExpert owners. Yes there is some loss inserted, but at least you can use the antenna where you couldn't before and who's complaining about less than a dB?
  5. AC6LA

    AC6LA Ham Member QRZ Page

    Yep, lots of ways to skin this cat, although I suspect a "cage" of wires would be a bit unwieldy if one wanted to broadband something like an 80m loop.

    Roger makes a good point. Using his region's 80m band limits of 3.5 to 3.8 MHz, with an inverted vee at 60 ft cut to resonate at 3.65 MHz, the loss in 100 ft of RG-213 is ~0.6 dB at the band edges and ~0.4 dB at band center. Blue trace below. But the fly in the ointment is the SWR at the station end of the coax. You only get about 160 kHz with the SWR below 2:1 (red trace) where many transmitters start to reduce power.


    That's a pretty slick way to use the RigExpert software. What you ended up with sounds like the feed system that is documented here:

    Frank Witt, AI1H, "A Simple Broadband Dipole for 80 Meters", QST, Sept 1993. (link)

    And modeled here (fourth section):

    As shown in that modeling study, one possible glitch is the height of the dipole/vee. For example, the Witt QST article says his antenna was an inverted vee with a 140 degree included angle and an apex height of 60 ft, fed with ~one wavelength of 50 ohm coax and ~qtr wavelength of 75 ohm coax. That's shown in blue below. But put the center at 30 ft (which makes the ends at about 9 ft) and you get the red curve.


    The Witt QST article also discusses all kinds of variations, like using two paralleled half-wave sections of 50 ohm coax instead of one full-wave of 50 ohm, and using a half-wave of 50 ohm instead of a full wave as was done in the modeling study linked above.

    You guys inspired me to do some additional research for other published broadband solutions. Here are two by N6LF:

    Rudy Severns, N6LF, "An Improved Double Extended Zepp", The ARRL Antenna Compendium Volume 4, 1995. (link)

    Here's the SWR response for an 80 meter version of Rudy's DEZ:


    Of course, for a flat version you'll need more than an American football field worth of length between supports. :)

    Rudy Severns, N6LF, "A Wideband 80-Meter Dipole", QST, July 1995. (link)

    Haven't modeled that one yet, looks promising. This is a folded dipole with an "open sleeve" (aka "coupled resonator") additional wire in the center. That should get Dave's attention!

    And another "coupled resonator" variant by W7ZZ:

    Douglas Smith, W7ZZ, "A Truly Broadband, Efficient Low-Band Dipole", The ARRL Antenna Book, 23rd edition, supplement. (link)

    This one can easily be modeled using the "CRdipoles.weq" model available in the second section here:

    The zip below now contains two additional models:

    Simple 80m Broadband Dipole.weq
    N6LF DEZepp.weq

    Dan, AC6LA

    Attached Files:

    N2EY likes this.
  6. G3YRO

    G3YRO Ham Member QRZ Page

    Not if you have a Valve PA or Amplifier !

    To be honest, I don't even have a problem with my 160m coax-fed dipole.

    Roger G3YRO
  7. KD6RF

    KD6RF XML Subscriber QRZ Page

    tnx Dan. Always appreciate the interesting info you provide.
  8. W5DXP

    W5DXP Ham Member QRZ Page

    One more solution:

    N2EY and AI3V like this.
  9. W4OP

    W4OP Ham Member QRZ Page

    Hi Dan,
    While the W7ZZ article is interesting, all of that work was already completed in 2012 by Ed Shortridge, W4JOQ and published in 2014 in his 1st of 3 antenna books- Super-Dipole Antennas CH13. I guess it shows that inquiring minds think alike.

    Dale W4OP
  10. N3DT

    N3DT Ham Member QRZ Page

    I just checked the antenna this morning and here's the 2 plots from the AntScope adding a section or RG11, 38' exactly was the best compromise. Can't complain about that. Takes all of about 5 minutes except making up the cable.
    80M.jpg 80M w rg11.jpg

Share This Page